Kristjanson, L., Lobb, E., Aoun, S., Monterosso, L.

Prepared by the WA Centre for Cancer & Palliative Care, Edith Cowan University, Pearson Street, Churchlands, Western Australia 6018. Phone: (08) 9273 8728

This publication was funded by the Australian Government Department of Health and Ageing Commonwealth of Australia 2006 ISBN No: This work is copyright. Apart from any use as permitted under the Copyright Act 1968, no part may be reproduced by any process without prior written permission from the Commonwealth available from the Australian Government Department of Communications, Information Technology and the Arts. Requests and inquiries concerning reproduction and rights should be addressed to the Commonwealth Copyright Administration, Intellectual Property Branch, Australian Government Department of Communications, Information Technology and the Arts, GPO Box 2154, Canberra ACT 2601 or posted at http://www.dcita.gov.au/cca. The opinions expressed in this document are those of the authors and not necessarily those of the Australian Government. This document is designed to provide information to assist policy and program development in government and nongovernment organisations.

TABLE OF CONTENTS ABBREVIATIONS ......................................................................................5 EXECUTIVE SUMMARY...........................................................................6 CHAPTER 1: INTRODUCTION & METHODOLOGY ..........................9
Aim..................................................................................................................... 10 Definition of terms............................................................................................ 10
Bereavement ...................................................................................................................................................10 Grief .................................................................................................................................................................10 Complicated grief...........................................................................................................................................10

Literature search strategy................................................................................ 11

Inclusion criteria.............................................................................................................................................12 Literature search process and results ..........................................................................................................12 Dimensions of evidence................................................................................................................................15 Quantitative evidence ....................................................................................................................................16 Qualitative evidence.......................................................................................................................................17 Methodological limitations of the review...................................................................................................19

CHAPTER 2: TERMINOLOGY, THEORIES, AND DIAGNOSTIC CRITERIA IN COMPLICATED GRIEF.................................................20

Normal Grief ....................................................................................................... 20 Complicated Grief................................................................................................ 20 Theories that influence bereavement and grief research ............................. 22
Grief work perspective..................................................................................................................................22 Attachment Theory........................................................................................................................................23 Meaning-making or meaning reconstruction.............................................................................................24 Cognitive Stress Theory ................................................................................................................................24 Dual Process Model.......................................................................................................................................24

Diagnostic criteria for complicated grief ........................................................ 25

Horowitz criteria ...........................................................................................................................................25 Prigersons criteria..........................................................................................................................................26 The inclusion of diagnostic criteria for complicated grief in DSM V ...................................................28

Summary ........................................................................................................... 30

CHAPTER 3: MEASURES IN COMPLICATED GRIEF ....................... 31

Texas Revised Inventory of Grief (TRIG) ................................................................................................31 Hogan Grief Reaction Checklist (HGRC) .................................................................................................32 Grief Evaluation Measure (GEM) ..............................................................................................................33 Core Bereavement Item (CBI).....................................................................................................................34 Inventory of Complicated Grief-Revised (ICG-R) ..................................................................................34 Revised Grief Experience Inventory (REGI)............................................................................................34 Bereavement Risk Index (BRI) ....................................................................................................................35 Grief Experience Questionnaire (GEQ)....................................................................................................35 Perinatal Grief Scale (PGS) ..........................................................................................................................35

Summary ........................................................................................................... 36

CHAPTER 4: COMPLICATED GRIEF AS A CONSTRUCT DISTINCT FROM ANXIETY, DEPRESSION, AND PTSD .....................................37
Complicated Grief and other mental disorders subsequent to bereavement.......................................37

Summary ........................................................................................................... 40

CHAPTER 5: VIOLENT AND TRAUMATIC DEATH .........................42

Complicated grief and suicide ideation.......................................................................................................43 Predictors of psychological distress in traumatic death ...........................................................................43 Predictors of complicated grief in traumatic death ..................................................................................45

Psychological outcomes of traumatic death ..............................................................................................45

Summary ........................................................................................................... 48

CHAPTER 6: RISK FACTORS FOR COMPLICATED GRIEF .............50

Predictors of risk for complicated grief......................................................................................................50 Practitioners views on risk factors that may predict complicated grief................................................52

Summary ........................................................................................................... 53

CHAPTER 7: OUTCOMES OF BEREAVEMENT AND THE RELATIONSHIP TO COMPLICATED GRIEF.....................................55

CG and Bereavement Outcomes.................................................................................................................55 Complicated grief as a risk factor for adverse health outcomes.............................................................59

Summary ........................................................................................................... 61

CHAPTER 8: COMPLICATED GRIEF IN SPECIFIC POPULATIONS62

Bereaved children and adolescents..............................................................................................................62 Summary..........................................................................................................................................................64 Bereaved parents ............................................................................................................................................65 Spouses ............................................................................................................................................................67 Palliative Care .................................................................................................................................................72 HIV/AIDS......................................................................................................................................................76 Euthanasia .......................................................................................................................................................77 Older adults.....................................................................................................................................................79 Mental Illness..................................................................................................................................................81 Cultural groups ...............................................................................................................................................83 Indigenous populations.................................................................................................................................84

Summary ........................................................................................................... 87

CHAPTER 9: GRIEF INTERVENTIONS ..............................................89

1. 2. 3. 4. Pharmacotherapy.....................................................................................................................................89 Support groups or counselling ..............................................................................................................90 Psychotherapy-based interventions ......................................................................................................90 Other interventions ..................................................................................................................................93

Summary ........................................................................................................... 93

CHAPTER 10: SUMMARY AND RECOMMENDATIONS ...................96 ACKNOWLEDGEMENTS ..................................................................... 108 REFERENCES........................................................................................ 109

TABLES Table 1.1: Table 1.2: Table 1.3: Table 1.4: Table 1.5: Table 1.6: Table 3.1: Table 3.2: Search Categories Search Results NHMRC Dimensions of Evidence Quantitative Levels of Evidence Original Studies: Quality Criteria Qualitative Levels of Evidence Horowitz et al. (1997) Criteria for Complicated Grief Disorder. Prigersons Criteria for Complicated Grief Proposed for DSM-V * 11 13 15 16 16 17 26 27

APPENDICES A-G ABBREVIATIONS Term Complicated Grief Major Depressive Disorder Posttraumatic Stress Disorder Inventory of Complicated Grief Abbreviation CG MDD PTSD ICG

EXECUTIVE SUMMARY In February 2005, the Western Australian Centre for Cancer & Palliative was asked to undertake a systematic review of the literature on complicated grief (CG). This review identified 2,262 abstracts as being potentially relevant. Of these, 705 were retrieved and assessed. Finally, 88 studies were reviewed. The following criteria were used to identify material that would be included for analysis in the current project: evidence based; published in an English language, peer-review journal between 1990 and 2005; and originating in a country with comparative health system and social and cultural similarities to Australia. Considerable diversity in the use of adjectives to describe variations from normal grief and the conceptualisations of complicated grief were noted in the literature. Studies demonstrated methodological difficulties such as high attrition, demographic differences between cases and controls, variations in methods of measurement of complicated grief and related outcomes, differences in length of time since death and recruitment techniques contributing to sample biases. Notwithstanding these limitations, some helpful findings were retrieved. The term Complicated Grief (CG) adopted in this review is grief that involves the presentation of certain grief-related symptoms at a time beyond which is considered adaptive. These symptoms include: (a) separation distress, such as longing and searching for the deceased, loneliness, preoccupation with thoughts of the deceased; and (b) symptoms of traumatic distress, such as feelings of disbelief, mistrust, anger, shock, detachment from others, and experiencing somatic symptoms of the deceased. The instruments that have been developed and tested to measure grief responses demonstrate good estimates of reliability and validity. The extent to which the instruments are able to predict complicated grief responses has not been well documented given the cross-sectional nature of the study designs. This systematic literature review confirms that a small percentage of the population (approximately 10% 20%) experience complicated grief, and that these individuals appear to be at greatest risk for adverse health effects. Risk factors specific to complicated grief suggest that insecure attachment styles play a crucial role. Other identified risk factors include childhood abuse and serious neglect, childhood separation anxiety, close kinship relationship to the deceased, marital closeness, support and dependency. Additional research is needed to clarify the nature of the association between complicated grief and adverse health outcomes and to identify the specific psychological and biological pathways through which CG is expressed in poor health. There is insufficient information on the clinical symptoms, clinical needs, and risk factors associated with unexpected and traumatic death.

Further research to examine situational factors (e.g., place of death, time from diagnosis to death), personal factors (e.g., gender, personality traits), and interpersonal factors (e.g., perceived lack of social support, poor coping skills) in the Australian context is needed. Our review provides evidence that survivors of suicide have an increased risk of complicated grief. This supports the notion that the unique features of traumatic death, when present in suicide or in any other traumatic loss account for much of the variance in bereavement outcome in comparison to natural causes of death. Studies relating to circumstances surrounding the death provide some evidence that complicated grief is an independent risk factor for suicidal ideation. There were a number of limitations in these studies, and the authors call for longitudinal data to determine whether CG and depression are preludes to suicidal ideation. Further investigation of this phenomenon is warranted. Further research is needed to identify the elements of a palliative approach to care that may be instrumental in achieving positive family bereavement outcomes. As well, further research is needed to better understand the needs of older adults who are not in a spousal relationship. The bereavement needs and grief risks for individuals that have never married, are divorced, have experienced the death of an adult child, friend, sibling or other relative are notably absent in the empirical literature and should become a future research priority. No intervention studies have been undertaken with children or adolescents to address CG. The only studies on complicated grief identified for children and adolescents focused on children exposed to trauma. No studies were identified in this review that specifically addressed complicated grief in Indigenous populations. The bulk of the research material related to Indigenous peoples identified focussed on intergenerational grief, historical grief, or grief associated with the stolen generation. Given the exposure of Aboriginal people to more perceived high-risk bereavements due to the high rate of premature mortality and the types of losses (accident, violent or illness) and the closeness and connectedness of Indigenous communities, this gap in knowledge is of particular concern. Further research specifically related to CG in this population, as well as new approaches or interventions are required to address the needs of this culturally disadvantaged population.

Most of the evaluated studies adopted tertiary preventive interventions for complicated grief. Studies of psycho-dynamically oriented treatments and behavioural/cognitive treatments indicate some proven effectiveness and hold promise for complicated grief. Additional research is needed to demonstrate the efficacy of pharmacotherapy for the reduction of symptoms of grief and randomised placebo-controlled trials are needed before more definitive conclusions can be drawn about their efficacy. Much more remains to be learned about the multiple sources of resilience and other protective factors. And future

research needs to examine links between assessment, intervention and outcomes that are targeted to welldefined patient populations at well-defined phases of bereavement framed within a public health agenda.

CHAPTER 1: INTRODUCTION & METHODOLOGY

The aim of this Report was to provide an overview and analysis of current bereavement interventions to inform future planning and work in complicated grief following bereavement within the National Palliative Care Strategy and the National Suicide Prevention Strategy. The Report also considered the following research questions in relation to: death in a palliative care context, suicide, and traumatic and unexpected death. The definitions of uncomplicated and/or non-traumatic grief versus complicated grief and its effects on the individual and the community (psychological, physical and social); How those at risk of complicated grief following bereavement will be identified, including whether appropriate and validated screening and/or assessment tools exist, and whether the risk of complicated bereavement is different for those where the death is expected versus unexpected; What interventions and the evaluations of those interventions currently exist including the client outcomes they identify; What are the pathways and relationships between assessment, intervention and outcomes for people identified with complicated grief; What is the best available evidence to guide bereavement support for special populations (eg, children, adolescents, Indigenous, individuals from non-dominant cultural groups); and What gaps exist in the above areas and what recommendations are needed for possible future research areas.

The Report begins by outlining the methodology used for the systematic review of the literature (Chapter 1). In Chapter 2 we examine the terminology, theories and diagnostic criteria for complicated grief, its relationship to normal grief and the theories and models that underpin it. Current opinion on the inclusion of complicated grief into the DSM-V is highlighted. Chapter 3 discusses the measures in complicated grief and demonstrates measures that are specific for certain populations. In Chapter 4 complicated grief is discussed as a construct distinct from anxiety, depression and PTSD. Chapter 5 reviews studies on violent and traumatic death. In Chapter 6, risk factors for complicated grief are discussed. Chapter 7 describes the outcomes of bereavement and the relationship to complicated grief. Chapter 8 discusses complicated grief within particular populations including children and adolescents, parents, spouses/partners, older adults and Indigenous populations and within the areas of HIV/AIDS, palliative care and euthanasia. Chapter 9 examines interventions in complicated grief. The Report concludes with an overall summary and recommendations for future research, policy and procedures (Chapter 10).

Aim The aim of this review was to obtain an overview and analysis of current bereavement interventions to inform future planning and work in complicated grief following bereavement within the National Palliative Care Strategy and The National Suicide Prevention Strategy. The review also considered the research questions in relation to: death in a palliative care context, suicide, and traumatic and unexpected death. As a first step to understanding this review a definition of core terms was needed. Definition of terms Bereavement Bereavement in the context of this review refers to the death of a loved one and in its broadest terms encompasses the entire experience of family members and friends in the anticipation, death, and subsequent adjustment to living following the death of a loved one [1, p. 554]. Bereavement includes the internal psychological processes and adaptation of family members, and expressions and experiences of grief. It also encompasses changes in external circumstances such as alterations in relationships and living arrangements [1]. Grief Grief is a normal reaction to loss and refers to the distress resulting from bereavement. Grief is multidimensional with physical, behavioural and meaning/spiritual components and is characterised by a complex set of cognitive, emotional and social adjustments that follow the death of a loved one. Although individuals vary in the type of grief they experience, the intensity of their grief, its duration and the ways in which they express their grief [1], most grieving people show similar patterns of intense distress, anxiety, yearning, sadness and pre-occupation and these symptoms gradually settle over time. The majority of the population appears to cope effectively with bereavement-related distress and most people do not experience adverse bereavement-related health effects [2, 3]. Complicated grief Complicated grief occurs when integration of the death does not take place. People who suffer from complicated grief experience a sense of persistent and disturbing disbelief regarding the death and resistance to accepting the painful reality. Intense yearning and longing for the deceased continues, along with frequent pangs of intense, painful emotions. Thoughts of the loved one remain preoccupying often including distressing intrusive thoughts related to the death, and there is avoidance of a range of situations and activities that serve as a reminder of the painful loss. Interest and engagement in ongoing life is limited or absent [4, p. 253]. It is estimated that between 10 and 20% of bereaved people experience complicated grief [5-7].

10

Literature search strategy To maximise the success rate of identifying all relevant work that has been conducted on complicated grief a combination of search terms was developed. These search terms were grouped into the three major search categories relevant to this project; Bereavement, Complicated, and Intervention (refer to Table 1.1). These search categories were then linked together to generate literature that contained at least one of the terms from each category. If this approach was not successful in generating information (eg, as was the case for the Indigenous databases) the search was widened by dropping the Intervention category. If this strategy was still not successful, the Complicated category was also dropped from the search resulting in a search that identified all information related to bereavement. Table 1.1: Search Categories Search Category 1. Bereavement bereave*, grief, grieving, mourn*

2. Complicated 3. Intervention complicated, absent, Search intervention*, postTerms abnormal, distorted, morbid, intervention, postmaladaptive, atypical, intervention, treatment*, intensified, prolonged, therapy, unresolved, neurotic, pharmacotherapy, dysfunctional, chronic, psychotherapy, counsel*, delayed, inhibited, cognitive-behaviour*, pathological psychodynamic, drug* Note: Boolean logic was used for the searches. Boolean Operators (and, or) were used to link the search terms. Or was used to link the search terms within each search category, whilst And was used to link the search categories together. Words were truncated with an asterisk (*) to allow for multiple endings for these words (bereave* finds the words bereavement or bereaved). During the literature review process the following specialist databases and resources were searched: MEDLINE; PsycInfo; CINAHL; EMBASE; APAIS; DRUG; AIATISIS bibliography; Current Contents; Science Citation Index; Cochrane Collaboration/Evidence Based Medicine; Database of Abstracts of Reviews of Effects (DARE); PsychBOOK; Dissertation Abstracts International; Caresearch; Australian Government Department of Health and Ageing Website; and other additional websites. The database SIGLE was not able to be searched as it is no longer available. Leading researchers in the field from the UK, Canada and the USA were contacted by letter with a list of the inclusion criteria for the review, and asked for information regarding any additional published or in press papers (see Appendix A). All Psychology Department Heads and Heads of Palliative Care Research Units in Australia were contacted by letter to locate eligible unpublished or ongoing studies (see Appendix A).

11

Inclusion criteria The following criteria were used to identify material that would be included for analysis in the current project: Evidence based; Published in a peer-review journal;* Published book chapters, government and non-government reports, therapeutic guidelines, standards of care and other guidelines were not evaluated. However, they are included in an additional resources reference list in the appendix; Published between 1990-2005; Seminal work pre-1990 are recorded in a reference list in the appendix; Published in English language; Originating in a country with comparative health system to Australia; Originating in a country with social or cultural similarities to Australia; In examining cultural aspects of complicated grief articles from USA, Canada, the Netherlands, Ireland, Pakistan and Israel are included.

*Given the relevance to the topic, the special edition of Omega: The Journal of Death & Dying (In Press) on complicated grief is included with special permission from the editors. Literature search process and results The abstracts identified in the searches were read to identify materials suitable for retrieval. Articles that appeared to discuss complicated grief, and met the inclusion criteria, were then selected and retrieved. Due to the inconsistent use of adjectives to describe complicated grief and the various conceptualisations of complicated grief, if the abstract did not contain enough information to ascertain whether or not the article was relevant, the article was retrieved. A 50% check for inclusion of the abstracts was conducted by a second reviewer for quality assurance. The search of the databases resulted in a total of 2262 references that were identified as potentially relevant to the Review (refer to Table 1.2). After reviewing the abstracts, 705 references were selected for retrieval.

12

Table 1.2:

Search Results Limits Searching Abstracts, 1990-2005, English All Fields Searching Topic Searching Abstracts Search Categories Bereavement + Complicated + Intervention Bereavement + Complicated Bereavement + Complicated + Intervention Bereavement Results 1229

Database Medline, CINAHL, & PsychINFO DRUG, APAIS, AIATISIS Current Content & Science Citation Index Cochrane & DARE PsychBOOK Dissertation Abstracts International EMBASE SIGLE CareSearch

6 303 98 13 84 275

Bereavement + Complicated 1990-2005 Bereavement + Complicated + Intervention 1990-2005, Bereavement + English Complicated + Intervention No longer available 1990-2005 Bereavement

254

In press publications were received from the invited editor of OMEGA the Journal of Death & Dying, Professor Colin Murray-Parkes for the special edition on Complicated Grief due for publication in 2006. The editors, Dr. Ken Doka and Professor Murray-Parkes gave permission for the use of this material prior to publication. Additionally in press manuscripts were received from key authors such as Neimeyer, Doka, Prigerson, Strobe, Shut and others.

Websites, such as the Australian Department of Health and Aging website, were hand searched for material about complicated grief. Additional material was also identified via specific searches for key authors and cross-checking the bibliographies of articles that met the inclusion criteria. In total 929 articles, books and additional resources were retrieved for review. After analysis 88 studies were considered to have met the inclusion criteria and were included for final review. Once full text versions of the articles were received, they were included in the review if they discussed CG and met the inclusion criteria. The majority of the studies excluded at this stage did not discuss complicated grief. Fifty percent of excluded materials were checked by a second reviewer to confirm exclusion. Articles that met the inclusion criteria and presented original research about complicated grief were evaluated and data from these articles were extracted into evidence tables (see Appendix B). Eighty percent of the included material was checked by a second reviewer to determine if it met the inclusion criteria and 10% of these studies were cross-checked by a second reviewer to confirm the evidence ratings for the papers.

13

Non-empirical reports that discussed complicated grief were not evaluated. Instead they were included in the reference list Articles by population (see Appendix C). Seminal work that was published before 1990 was included in the Seminal work reference list (see Appendix D). Articles that include case studies are listed in the Case Studies reference list (see Appendix E). Additional resources were included in the Additional resources reference list (see Appendix F) these included; dissertations, conference papers, website material, government reports, and guidelines. Relevant theses were identified through searching the Dissertation Abstracts International database. Thirty-seven theses that discussed complicated grief were identified and the summary available online was downloaded. Additional searches were undertaken to identify any published work arising from the theses. Conference papers were identified through CareSearch, The New York Academy of Medicine Library's Grey Literature Report, and the other various databases that where searched. However, it was difficult to locate conference papers because libraries do not usually hold conference proceedings. Conference papers that were located tended to be from recent conferences that still had information available online. In addition to database searches for published articles, websites were searched for additional resources. The majority of information on the internet about complicated grief comes from the United States and mainly consists of brief explanations of CG. Various bereavement resource kits were identified in areas such as suicide, palliative care, and Indigenous Australians. However, complicated grief was only very briefly discussed in a small number of cases. Resource packs aimed at the consumer, such as the Australian Governments suicide and sudden death information and support pack [8] outline the various emotions that may be felt during bereavement. These documents emphasise that although it is normal to experience a range of emotions, if they continue for an extended period of time and interfere with a persons life, additional support may be needed. The consumer packs also provide contact details of various support agencies, as well as additional references for useful websites and books. No resource kits aimed specifically at complicated grief were identified. Finally, guidelines for practitioners were identified that outline risk factors and ways of identifying and assessing people at risk (e.g., Bereavement counselling: Guidelines for practitioners by Dianne and Mal McKissock) [9].

14

Dimensions of evidence The aim of this literature review was to find the highest quality evidence to answer the questions. In accordance with the National Health and Medical Research Council [10] criteria, the following dimensions of evidence were reviewed for each of the included studies (Table 1.3). It is important to recognise that the value of a piece of evidence is determined by all of these dimensions, not only the level of evidence. Table 1.3: NHMRC Dimensions of Evidence Reviewers Definition The study design used, as an indication of the degree to which bias has been eliminated by the design alone. The levels reflect the effectiveness of the study design to answer the research question. The methods used to minimise bias within an individual study. An indication of the precision of the estimate of effect reflecting the degree of certainty about the existence of a true effect, as opposed to an effect due to chance. Determines the magnitude of effect and whether this is of clinical importance. The considers the relevance of the study to the specific research question and the context in which the information is likely to be applied, with regard to a) the nature of the intervention b) the nature of the population and c) the definition of outcomes.

Dimension Strength of the evidence Level (see Table 1.4)

Quality Statistical precision Size of effect Relevance of evidence

15

Quantitative evidence The levels of evidence defined by the NHMRC [10] were used to categorise the study design of the individual studies. The hierarchy of evidence is summarised in Table 1.4. Table 1.4: Levels I II III-1 III-2 Quantitative Levels of Evidence Criteria Systematic review of all relevant randomised controlled trials (RCTs) At least one properly designed RCT Well-designed pseudo-RCT Comparative studies with concurrent controls and allocation not randomised, case-control studies or interrupted time series with a control group III-3 Comparative studies with historical control, two or more single-arm studies, or interrupted time series without a parallel control group IV Case series, either post-test or pre-test and post-test

The highest level of evidence available is a systematic review of randomised controlled trials because they are considered the study type least subject to bias. Individual randomised controlled trials also represent good evidence. However, comparative observation studies such as cohort and case control studies or non-comparative case series are often more readily available. Even within the levels of evidence stated above, there is considerable variability in the quality of evidence. In accordance with NHMRC guidelines, it was necessary to consider the quality of each of the included studies. The characteristics and quality of each included study were assessed using a number of quality criteria as shown in Table 1.5, with studies rated as good, fair or poor quality. Table 1.5: Quality criteria (A) Has selection bias (including allocation bias) been minimised? (B) Have adequate adjustments been made for residual confounding? (C) Was follow-up for final outcomes adequate? (D) Has measurement or misclassification bias been minimised? Original Studies: Quality Criteria

16

Qualitative evidence For qualitative studies, the level of evidence and the quality of evidence were combined to create a single category that deals with the intent of the study and the methodological appropriateness of the study. Qualitative studies are usually descriptive with the aim of providing a context for peoples experience and behaviours through analysis that is detailed, rich and integrative. Examples of qualitative studies include observational or case study methods that explore comparisons within a group to describe and explain a particular phenomenon (e.g., comparative case studies with multiple communities). According to the Campbell Collaboration [11], incorporating relevant qualitative studies in a systematic review is beneficial because it can: (a) (b) (c) Contribute to the development of a more robust intervention by helping to define an intervention more precisely; Assist in the choice of outcome measures and assist in the development of valid research questions; and Help to understand heterogeneous results from studies of effect.

However, the inclusion of evidence from qualitative studies, while resolving some of the short-comings of quantitative studies also raises some concerns, such as the potential for biases in the methodology that may invalidate the conclusions [12]. To overcome this problem, qualitative evidence was reviewed and examined using criteria to measure the quality of these studies. No appropriately validated tool existed that could measure the quality of qualitative studies. Therefore, the Cochrane Collaboration [12] and Campbell Collaboration [11] guidelines have been substantially modified to provide an appropriate evaluation tool (see Table 1.6; for a complete example of the tool see Appendix G). Table 1.6: Questions 1. Aim of the study: Was the aim clear? 2. Paradigm: Was the paradigm appropriate for the aim? Quality of evidence: 3. Methodology: Was the methodology appropriate for the paradigm? 4. Methods: Were the methods used appropriate for the methodology? Qualitative Levels of Evidence Yes = 1 / No = 0

17

Questions 5. Checking methods: Did checking methods establish rigor? 6. Sample: Did the sampling strategy address the aim? 7. Data analysis: Was the data analysis appropriately rigorous? 8. Findings: Were the findings clearly stated and relevant to the aim?

Yes = 1 / No = 0

Level of evidence score:

(Sum scores. Score range from 0 to 8 with 8 being highest level)

Additional - (Do not add additional scores to previous totals).

Strength of evidence (choose only 1 score): 4. Very high 3. High 2. Low 1. Very low

Explanation

Strength score:
(Score range 1 to 4 with 4 being highest level of strength) A single category was created to determine the level of evidence to measure the quality for the qualitative studies (designated as qualitative evidence or QE). Quality was assessed using eight questions (See Table 1.6). Each question in this category required the reviewer to answer yes or no with yes scored as 1 and no scored as 0. The score range for the level of evidence was 0 to 8 with 8 being the highest level of evidence, and, therefore, the best quality. The reviewer then considered the theoretical rigour (strength) of the study (e.g. did the study have soundly constructed arguments and analysis that followed on from each other and were supported with evidence from other sources?). Strength had a score range of 1 (very low no strength) to 4 (very high very strong). Studies that scored a 1 or 2 for relevance were not included in the review, even if they had a high level of evidence and/or strength.

18

Methodological limitations of the review All types of studies are subject to bias, with systematic reviews being subject to the same biases possible in the original studies that are included, as well as biases specifically related to the systematic review process. Reporting biases are a particular problem related to systematic reviews and include publication bias, timelag bias, multiple publication bias, language bias, and outcome reporting bias [13]. Other biases can result if the methodology to be used in a review is not defined before the review commences. Detailed knowledge of studies performed in the area of interest may influence the eligibility criteria for inclusion of studies in the review and may therefore result in biased results. For example, studies with more positive results may be preferentially included in a review, thus biasing the results and overestimating treatment effects. We endeavoured to minimise these biases by contacting key authors for information about current studies that may be either in press or published recently. In addition, the use of broad terms in our searches, cross-referencing and searches by authors name have produced a comprehensive and systematic review. Searches were limited to articles published in English. English language journals are predominantly published in first world countries and this may subsequently limit some exposure to bereavement issues. Studies may not be listed because of a time delay between an article being published, and appearing on the database, the journal not being cited on the database, or the database not providing an abstract. Results from a study presented at the Australian and New Zealand Society of Palliative Medicine Meeting, Townsville, Australia in September 2002 reported approximately 30% of the palliative care literature was missed on the electronic databases [14]. Different search engines use different key words and different search strategies to identify articles and these differences may have limited the capture of appropriate articles.

19

CHAPTER 2: TERMINOLOGY, THEORIES, AND DIAGNOSTIC CRITERIA IN COMPLICATED GRIEF Normal Grief Shear and Shair [4] in a recent review, give a succinct outline of the difference between normal grief and CG. They describe normal grief as the state that occurs when people are deeply saddened by the death of an attachment figure during a period of weeks or months of acute grief [4, p. 253]. They acknowledge the individuality of grief and that responses vary. However, the person who typifies normal grief experiences an intense yearning, intrusive thoughts and images, and/or a range of dysphoric emotions and that these symptoms do not persist [4, p. 253]. The initial reaction subsides, interest and engagement in daily activities is renewed and the loss is integrated into the bereaved persons on-going life [4, p. 253]. As this integration occurs painful feelings lessen and thoughts of the loved one cease to dominate the mind of the bereaved [4, p. 253]. For a minority of people, a normal grief adjustment does not occur. It is estimated that between 10 and 20% of people find coping painful and difficult [5-7]. Shear and Shair [4, p. 253] propose that for this small percentage of people, integration of the loss does not occur and acute grief is prolonged in the form of CG. People who suffer from CG experience a sense of persistent and disturbing disbelief regarding the death [4, p. 253]. There are feelings of anger, bitterness, and resistance to accepting the painful reality. Intense yearning and longing for the deceased continue, along with frequent pangs of intense, painful emotions [4, p. 253]. Thoughts of the loved one remain preoccupying often including distressing intrusive thoughts related to the death, and there is avoidance of a range of situations and activities that serve as a reminder of the painful loss. Interest and engagement in ongoing life is limited or absent [4, p. 253]. This type of response has been described using different terminology; however, for the purpose of this review the term complicated grief has been adopted. Complicated Grief This systematic literature review on complicated grief indicates that the majority of researchers in the field agree that complications of grief do exist. However, the terminology, definitions and criteria used to describe complicated grief have not been consistent [15-17]. The diagnostic term for complications that arise from grief has been variably defined over the past 20 years, with a multitude of adjectives used to describe variations from normal grief. These adjectives include absent, abnormal, complicated, distorted, morbid, maladaptive, atypical, intensified and prolonged, unresolved, neurotic, dysfunctional, chronic, delayed, and inhibited. Further refinement of this terminology was undertaken and more consistency appeared in the literature around 1993 with the use of terminology such as delayed or absent grief, inhibited or distorted grief and chronic grief [18, 19].

20

Historically, researchers have argued that complicated grief is an expression of a major depressive disorder or an anxiety-based disorder, that has been triggered by the death [20, 21]. More recently, researchers have concluded that grief symptoms only partially overlap with symptoms of depression and other DSM categories such as anxiety and post traumatic stress disorder and that although there may be some expected shared variance, complicated grief reactions do display sufficient unique variance to warrant separate consideration [21-26]. Within the last decade several studies have attempted to establish a definition of complicated grief that extends beyond clinical descriptions and that allows for empirical validation [19, 22, 27-29]. Most current researchers have attempted to identify the symptoms of complicated grief by the taxonomy provided by Prigerson and Jacobs (2001) that follows the format of existing disorders in the DSM [28]. Their rationale is that if the requirements for a distinct psychiatric illness are met, then complicated grief should be considered as a separate diagnosis. (See summary of current opinion in Chapter 3). Prigerson and Jacobs (2001) suggested that symptoms of complicated grief fall into two categories: (a) symptoms of separation distress, such as longing and searching for the deceased, loneliness, preoccupation with thoughts of the deceased and (b) symptoms of traumatic distress, such as feelings of disbelief, mistrust, anger, shock, detachment from others, and experiencing somatic symptoms of the deceased. This schema allowed bereavement experts to identify a class of symptoms for a disorder of grief and the Inventory of Complicated Grief-Revised (Prigerson & Jacobs 2001) has been developed to measure these sets of symptoms. [28] Prigerson and colleagues have subsequently revised their 2001 criteria as outlined in Table 3.2 in this chapter [15]. However, for the purposes of this review, we examined studies against the criteria identified by the consensus meeting with clinical and scientific experts in bereavement, mood and anxiety disorders, and psychiatric nosology in 2001 and upon which the Inventory of Complicated Grief-Revised was developed. These symptoms are proposed by Prigerson to be indicative of pathology and that the issue is not whether the symptoms sort themselves into seemingly pathological versus seemingly normal symptom clusters, but that the set of CG symptoms identified is persistent (beyond six months post-death) and severe (marked intensity or frequency, such as several times daily) and predict many negative outcomes distinguishing them from normal grief symptoms [15]. The term Complicated Grief (CG) adopted in this review is grief that involves the presentation of certain grief-related symptoms at a time beyond which is considered adaptive. These symptoms include: (a) separation distress, such as longing and searching for the deceased, loneliness, preoccupation with

21

thoughts of the deceased; and (b) symptoms of traumatic distress, such as feelings of disbelief, mistrust, anger, shock, detachment from others, and experiencing somatic symptoms of the deceased. Prigerson previously used the term "Traumatic Grief" because it was felt that their term captured the essence of the underlying forms of symptomatic distress, conceptualised in the original version of the Inventory of Complicated Grief developed in 1995. After the 9/11 terrorist attacks the need to distinguish a grief disorder from Post Traumatic Stress Disorder became apparent (Prigerson, personal communication, 2006). As a result, they reverted to their original term of Complicated Grief in an attempt to minimise confusion between this reaction to loss and the psychological reaction following exposure to traumatic events such as the 9/11 attacks (i.e., Post-traumatic Stress Disorder). Prigerson perceived a basic distinction between Complicated Grief, which was rooted in interpersonal attachment issues, and Post Traumatic Stress Disorder, which was grounded in a sense of impending dangerous events feared likely to harm one-self or others. The decision to revert back to the term Complicated Grief was made to clarify the distinction between these two disorders (Prigerson, personal communication, 2006). Differences of opinion about CG appear to focus on the specifics of the diagnostic criteria and their categorisation, determination of the boundaries between normality and pathology, concerns about social coercion and issues of stigmatisation [30]. Theories that influence bereavement and grief research Five overarching theories were identified that shape bereavement and grief responses: the Grief Work Perspective, Attachment Theory, Meaning-making or meaning reconstruction, Cognitive Stress Theory, and the Dual Process Model. Grief work perspective The grief work perspective has dominated thinking about bereavement and grief and is based on Freudian theory [31]. The grief work hypothesis states that it is necessary to bring the reality of the death into awareness to avoid complications in the course of grief [32]. Some theorists and researchers have suggested that the absence of empirical evidence in support of the grief work perspective has led to questioning of its effectiveness [33, 34]. For example, in 1989 Wortman and Silver found no empirical support for the five dominant ideas around grief work in the professional and lay literature at that time such as; (i) distress or depression is inevitable; (ii) distress is necessary, and failure to experience distress is indicative of pathology; (iii) the importance of working through loss; (iv) the expectation of recovery and (vi) reaching a state of resolution. [35] In a study to examine the efficacy of grief work, Stroebe and Stroebe [36] measured five different types of behaviours associated with confrontational grief work or its avoidance. They concluded that there simply has been very little empirical evidence that working through grief is a more effective process of

22

coming to terms with loss than not working through it [32, p. 885], and that more precise specification of the nature of grief work and additional research into specific elements of the processes involved are needed before the hypothesis can be completely disregarded. Whereas the traditional grief work perspective has emphasised the necessity of breaking the attachment bond with the deceased loved one, more recently, researchers have highlighted the important role played by maintaining a continued sense of attachment with the deceased [37]. They conclude that maintaining continuing psychological and emotional bonds with the deceased is not necessarily an indication of problematic grieving [32, 38]. The wide range of grief patterns demonstrated in the study by Bonanno and colleagues [39] points to a need to re-evaluate common notions about what constitutes a normal response to a major loss. Views about normal grieving are not only prevalent among researchers and health providers, but are also held by people and the bereaved themselves. Because they are unaware of the striking variability in response to loss, potential supporters are often critical or judgemental of bereaved individuals who show too little or too much grief. Also the bereaved themselves may become concerned that their reaction to the loss is abnormal and this may add to their distress. Studies suggest that many of the assumptions that have guided interventions with the bereaved may need to be re-evaluated. It is widely assumed that absent grief is indicative of under acknowledged problems related to the loss, that individuals must work through the loss, and that bereavement is one of the most stressful life events that most people will encounter. Bonannos studies add further weight to previous studies that question the grief work theory, suggesting that more research into how people cope with loss is needed. Attachment Theory Attachment theory [40-42] provides a framework for understanding the effects of bereavement in terms of the disruption of affectional bonds and in terms of individual differences in response to the death of a significant other. Much of the core phenomenology of CG arises from the sundering of a securityenhancing attachment bond with the deceased, making attachment theory a highly relevant conceptual context within which to interpret the separation distress that follows intimate loss. However, it has been argued that some forms of insecure attachment, such as those involving avoidance or dismissal of intimacy based on defensive exclusion of vulnerable feelings of rejection, might actually mitigate against the pursuit of a continuing bond with the deceased [43]. They may also mitigate against the core yearning and longing symptomatology suggested in the criteria for CG. For these and other reasons, the interface between attachment histories and styles on the one hand, and complicated versus adaptive forms of grieving on the other, deserves further exploration[44].

23

Meaning-making or meaning reconstruction A further theoretical framework focuses on struggles with meaning reconstruction in the aftermath of bereavement [45, 46]. Meaning-making or meaning reconstruction emphasises responses to bereavement from the perspective of an individual striving to make sense of troubling events and which is often expressed in the organisation of experiences into narrative form [47, p. 499]. Meaning is sometimes framed in terms of the individuals interpretation, beliefs and self-statements. Individual consciousness represents one site for construction of meaning, which also resides and arises in language, cultural practices, spiritual traditions, and inter-personal conversations, all of which interact to shape the meaning of mourning for a given individual or group [48] p.248. Cognitive Stress Theory Cognitive stress theory (e.g. Folkman) [49] has also been influential in recent research on bereavement and grief, especially in terms of recognition of the role played by positive emotions in adaptive response to bereavement. Traditionally, bereavement theorists have assumed that recovery from loss is based on the concentrated review and expression of the negative emotions brought about by grieving. This process, considered part of the work of mourning, is thought to foster acceptance of the finality of the death and aid in the necessary severing of attachment to the lost relationship. The social-functional perspective on grief and emotion has shifted attention away from an emphasis on the expression of negative emotions and hypothesises that recovery following the death of a loved one is made more likely when grief-related distress is minimised and positive emotion is activated or facilitated [50, p. 493]. A sizeable minority of people do not experience (or do not report experiencing) distress following loss [38, 50]. Some researchers speculate that this need not indicate absent grief or a delayed grief reaction [38] both of which have been considered problematic from the grief work perspective. It has been suggested that bereavement in some circumstances may represent the end of a difficult situation e.g., a stressful care giving situation or painful terminal illness, or even the end of an abusive relationship [27]. Dual Process Model The dual process model of coping distinguishes two types of stressor, namely, loss-oriented (focusing on the deceased and death events; confronting and dwelling on loss) and restoration-oriented (dealing with secondary stressors, such as coping with finances, learning to run a household) [51]. It proposes that bereaved individuals oscillate between these two types of coping, that is, between efforts to resolve the loss experience itself and efforts to master or adapt to challenges associated with the changes in life circumstances resulting from bereavement. The extent to which bereaved individuals will engage in either loss-oriented or restoration-oriented processes depends on various factors, such as personality or cultural expectations and practices [32].

24

The theories that underpin CG provide some useful conceptual parameters within which to examine treatment approaches. Examination of competing theories is not unhealthy; however, this multiplicity of theories, together with definitional inconsistencies has created uncertainty for health care providers and services that endeavour to make sense of the findings related to CG. It is evident that future research must make explicit the theoretical framework and definitions guiding the study to allow development of empirical evidence that can permit sound comparisons. Diagnostic criteria for complicated grief In recent years, studies have been undertaken to provide the empirical data that would establish CG as a distinct clinical entity. Thus CG would be a unified syndrome distinct from bereavement-related depression and anxiety and from normal reactions to bereavement. The debate in these studies centres around the extent to which CG represents a truly unique pathological entity, when contrasted with depressive or anxiety disorder, post-traumatic stress disorder (PTSD), and uncomplicated grief [52]. Key researchers active in debating and establishing diagnostic criteria for CG include Horowitz and his colleagues at the University of California and Holly Prigerson and her colleagues at Yale University. Each of these groups has proposed a different set of diagnostic criteria (it should be noted that in earlier publications on Prigersons criteria she has used the term traumatic grief). Following a study by Hogan and colleagues at Loyola University in Chicago where they undertook to empirically test the CG disorder criteria [53], a special edition of Omega: Journal of Death and Dying was commissioned. The ensuing contributions from leading researchers in the field could be considered the most recent in the debate and are included in this review with permission from the editor Dr. Ken Doka and invited guest-editor Professor Colin Murray Parkes. Contributors were asked to respond to three questions: Is there a type of grief that can justifiably be regarded as a mental disorder; If so, how should the disorder be classified in relation to other disorders? and finally, what criteria for diagnosis are best supported by systematic research? [54]. Horowitz criteria In establishing their criteria, Horowitz and colleagues [22] followed the method of the Structured Clinical Interview used for DSM-III-R along with self-report rating scales in subjects studied 6 and 14 months after bereavement (Level III-3). They identified 30 questions relating to possible symptoms of CG. The data were analysed using sophisticated methods of latent class and signal detection techniques in order to produce a model set of criteria for CG Disorder. The criteria that resulted are shown in Table 3.1.

25

Table 3.1: Horowitz et al. [22] Criteria for Complicated Grief Disorder. A. Event Criterion/ Prolonged Response Criterion Bereavement (Loss of spouse, other relative or intimate partner) at least 14 months ago (to avoid anniversary). B. Signs and Symptoms Criteria In the last month any three of the following, with a severity that interferes with daily functioning: Intrusive Symptoms 1.) Unbidden memories or intrusive fantasies related to the lost relationship. 2.) Strong spells or pangs of severe emotion related to the lost relationship. 3.) Distressing strong yearnings or wishes that the deceased were there. Signs of Avoidance and Failure to Adapt 4.) Feeling of being far too much alone or personally empty. 5.) Excessively staying away from people, places or activities that remind the subject of the deceased. 6.) Unusual levels of sleep interference. 7.) Loss of interest in work, social, caretaking, or recreational activities to a maladaptive degree. In a sample of 70 self-selected bereaved subjects, 41% met these criteria for CG 14 months after the loss. Thirty-one percent met criteria for Major Depressive Disorder (MDD) with a concordance of both diagnoses in only 9%. Despite this relatively low concordance, a previous history of depression or anxiety disorder was associated with a significantly increased risk of CG [22]. Prigersons criteria Prigersons group first developed their Inventory of Complicated Grief (ICG) in 1995 [27] and have subsequently demonstrated its specificity, reliability, validity and ability to predict a variety of measures of physical and mental health in a series of studies [23, 24, 27, 55]. This has been the tool on which consensus criteria were developed in 2001 [28]. In that paper they describe the process through which the criteria were established [30]. They started by holding a consensus conference to review the evidence and develop a preliminary set of criteria. The group agreed that, for the time being, bereavement by death should be an essential criterion as should symptoms of separation distress which they see as at the core of the diagnosis. Because these criteria do not distinguish complicated from uncomplicated grief, they added three further requirements: symptoms of traumatisation, impairment of functioning and, that the condition has continued for at least two months from the time of onset (which, in the case of delayed reactions, is not from the time of death).

26

Prigerson and colleagues received a grant to test the 2001 criteria on a community-based sample of bereaved and the latest criteria and assessment are the result of that research (Prigerson, personal communication). This full set of new criteria is reproduced with permission in Table 3.2. and appears in the special edition of Omega [15]. Table 3.2: Prigersons Criteria for Complicated Grief Proposed for DSM-V *

Criterion A: Chronic and persistent yearning, pining, longing for the deceased, reflecting a need for connection with deceased that cannot be satisfied by others. Daily, intrusive distressing and disruptive heartache. 1. Yearning/longing/heartache - `Do you feel yourself yearning and longing for the person who is gone?

Criterion B. The person should have four of the following eight remaining symptoms at least several times a day or to a degree intense enough to be distressing and disruptive: 1. Trouble accepting the death `Do you have trouble accepting the loss of ___? 2. Inability to trust others `To what extent has it been hard for you to trust others since the loss of ___? 3. Excessive bitterness or anger related to the death - `Do you feel angry about the loss of ___? 4. Uneasy about moving on `Sometimes people who lose a loved one feel uneasy about moving on with their life. To what extent do you feel that moving on (for example, making new friends, pursuing new interests) would be difficult for you? 5. Numbness/Detachment - `Do you feel emotionally numb or have trouble feeling connected with others since ____ died? 6. Feeling life is empty or meaningless without deceased `To what extent do you feel that life is empty or meaningless without ___? 7. Bleak future `Do you feel that the future holds no meaning or prospect for fulfilment without ____? 8. Agitated Do you feel on edge or jumpy since ____ died? Criterion C. The above symptom disturbance causes marked and persistent dysfunction in social, occupational, or other important domains.

Complicated Grief Diagnosis = Criteria A, B, and C are met. Permission for reproduction given Prigerson 2006 [15]

27

In comparing Horowitzs criteria [22] and Prigersons 2001 criteria [15] , Goodkin [52] identifies the following differences. Horowitz did not make separation distress an essential criterion[22], Prigerson did [15]; Horowitz insisted on the lapse of 14 months from bereavement to diagnosis, Prigerson required only 2 months (from the onset of symptoms) [15]; Only Horowitz included interference with sleep and avoidance of reminders [22]; Only Prigerson included loneliness, emotional blunting, identification symptoms, disbelief, shattered world view and anger [15]; Although both included impairment of functioning, only Prigerson made this an essential criterion (Criterion D) [15].

There is agreement that based upon the research to date, the criteria set proposed by Prigerson et al. appears to have advantages over that proposed by Horowitz et al. [22] as evidenced by higher estimation of internal consistency and in construct validity (related to its focus upon separation distress and impaired social and occupational function) [52]. Parkes [56] agrees that Prigersons criteria best meet the psycho-metric requirements, and that Horowitz criteria fails to clearly differentiate CG from other possible consequences of bereavement and that their criteria places undue emphasis on traumatic avoidance [44, 52]. The inclusion of diagnostic criteria for complicated grief in DSM V The diagnosis of complicated grief is a separate issue to its inclusion in the proposed DSM-V and although it is not the brief of this review to make recommendations about the inclusion of CG in the DSM V, the current debate in Omega (in press) provides an opportunity to summarise the views of key researchers. The DSM is a non-theoretical categorisation system with an emphasis on phenomenology, etiology, and course as defining features of mental disorders. It offers guidance to mental health professionals with regard to what is pathological and what is normal [57]. The existing DSM-IV-TR has recognised that grief symptoms may warrant clinical attention; however, they do not acknowledge CGs unique set of symptoms [57]. The DSM classifies bereavement as a normal stressor, but more severe pathology is classified in existing diagnostic categories (eg., Major Depressive Disorder) [34] . Prigerson and colleagues call for CG to be established as a unique diagnosis. However, Strobe cautions that the inclusion of CG in DSM would have far reaching impact as it is a leading guide for practitioners

28

[58, p. 58] and there is a need to look at appropriate care for the bereaved. Some people would benefit from receiving treatment, but may not be able to obtain access to it because CG does not have a diagnosis category. However, there is a need to ensure that appropriate guidelines are developed so that therapy is appropriate [58, pp. 58-59]. Other concerns are that DSM IV-TR is culturally bound and that the diagnosis category of bereavement needs to be expanded to deal with Indigenous peoples experience of broad-based cultural losses [59, p. 52]. One argument is that the relatively small subset of people who experience CG are adequately captured by existing diagnosis categories in the DSM as they [appear] to experience symptoms similar to individuals suffering from depression and anxiety disorders, and to some extent trauma reactions [34]. Horowitz suggests that complicated grief disorder (preferred term) should be included in a separate category of Stress Response Syndromes or in a separate category of its own with diagnostic criteria [60]. Goodkin and colleagues suggest a compromise position that incorporates CG into DSM-V, but relegates it to Appendix B (disorders proposed for further study) due to the lack of clarity surrounding its diagnostic criteria. Prigerson [15] agrees with Goodkin [52] and Stroebe and Shut [44] when they note that the Horowitz criteria [22] places undue emphasis on traumatic avoidance. She recommends that the focus of the criteria for CG is on the relationship and the meaning behind (and in front) of the loss of the important relationship the survivor has lost [15]. Therefore, she recommends that CG neither be grouped among mood nor anxiety disorders (including PTSD and similar stress response syndromes), not be event based, but rather be placed separately within a new category of Attachment Disorders [30]. Stroebe and Schut [44] summarise the current different views of researchers active in debating and establishing a diagnostic categorisation of CG in the DSM V: 1. CG should be incorporated within the DSM classification systems diagnostic category PTSD. 2. Two separate categories are needed, PTSD (for traumatic bereavement) and CG (for nontraumatic bereavement). 3. A new category of traumatic grief (specifically for disordered grieving following a traumatic bereavement) should be developed. 4. A new category of CG covering non-traumatic and traumatic bereavement experiences is called for.

29

5. Complicated grief is an entity separate from trauma, following non-traumatic bereavement; CG alone should be the focus and concern in developing a new category. Parkes [56] expresses the view that the concept of CG is now so well-supported that it deserves to be recognised as a specific disorder rather than be assigned to any of the suggested related categories and Prigerson has recently contacted all scientists, researchers and clinicians in the field of bereavement to critically appraise and form consensus on her current 2006 criteria for inclusion in the DSM-V (e-mail correspondence, 2006). Summary There is evidence to confirm that complicated grief does occur in small proportions within the bereaved population. There is recent consensus regarding criteria for diagnosing complicated grief. Future discussions will determine if CG should be included in the DSM-V. Issues identified with diagnosis of individuals with CG include concerns about misuse of the term, distinctions between normal and complicated grief, fears regarding stigmatisation and health insurance funding issues associated with potential DSM-V classification of CG. Use of the term as described by Prigerson and colleagues reflects current best evidence, addresses concerns related to definitional error and would assist in progressing research and clinical practice in a more consistent manner if this were used by clinicians, researchers, health policy makers and educators. There appears to be little evidence to support concern about stigmatisation of individuals who are diagnosed with CG. Recommendations: Terminology, Theories, and Diagnostic Criteria

It is recommended that any communication (written or web-based) from relevant areas within the Australian Government Department of Health & Ageing that refers to complicated grief use the most current definition as outlined in this report and be consistent in use of the term.

It is recommended that training be provided to health professionals involved in the care of the bereaved (e.g. GPs, psychologists, psychiatrists, counsellors, community health workers) regarding accepted criteria for diagnosing CG. Such training should be included in under-graduate and post-graduate courses.

We support the recommendation of Parkes that the concept of CG is now so wellsupported that it deserves to be recognised as a specific disorder rather than be assigned to any of the suggested residual categories with DSM V.

30

CHAPTER 3: MEASURES IN COMPLICATED GRIEF

An examination of instruments to measure CG was included to identify ways in which CG might be assessed. Particular attention was given to the extent to which the instruments indexed the concept in a way that was concordant with current definitions of CG. As well, the instruments were examined for validity and reliability and the practical considerations of using the various tools in the clinical context. The measurement of emotional and cognitive process in response to loss is extremely complex, with the use of tools that range from broad health measures to population and situation specific questionnaires. Neimeyer and Hogan (2001) [61] identify general psychiatric symptom instruments that index broad distress, such as the Brief Symptom Inventory (BSI) or the Symptom Checklist (SCL-90). Generic measures of symptomatology are included in literature reporting on bereavement studies and responses to trauma, focusing on specific symptom clusters such as depression and anxiety. Measures in complicated grief reviewed include the Texas Revised Inventory of Grief (TRIG) (1988) [62]; the Hogan Grief Reaction Checklist (HGRC) (2001) [63]; the Grief Evaluation Measure (2005) [64]; the Revised Grief Experience Inventory (REGI) (1993) [65]; the Core Bereavement Items (1997) [66], the Inventory of Complicated Grief (ICG) [15, 27], the Inventory of Complicated Grief-Revised [30] and the Parkes Bereavement Risk Index (1993) [67]. Three specific grief assessment scales are reported that focus on specific populations: the Grief Experience Questionnaire (GEQ) [68] focusing on responses to suicide, Perinatal Grief Scale [69], and the Perinatal Bereavement Scale (PBS) [70]. Texas Revised Inventory of Grief (TRIG) The Texas Revised Inventory of Grief is a 21-item scale designed to measure the extent of unresolved or pathological grief. It relates to two points of time: past (immediate of shortly after the death) and present (the time of data collection). It is comprised of two subscales, structured as a five-point Likert-type questionnaire [62]. Items are summed to produce a total score. The first 8-item subscale measures feelings and actions at the time of the death (i.e., the extent to which the death affected emotions, activities and relationships). The second 13-item subscale measures present feelings (continuing emotional distress, lack of acceptance, rumination, painful memories). Split-half reliability of 0.81 has been reported [71] and internal consistency estimates of the two subscales of 0.77 and 0.86 respectively have been reported [72]. These estimates are acceptable given the brevity of the subscales [73]. Some evidence of the construct validity of the instrument has been reported using criterion group analysis and is supported by the fact that the intensity of responses varies over time as would be anticipated (i.e., worsening over the first year and then gradually improving) [72].

31

One criticism of the first subscale is the retrospective nature of the questions and the potential for memory of past emotional states to be influenced by the respondents current state [72]. The instrument has been expanded to a 58-item measure, the Expanded Texas Inventory of Grief [74]; however, no estimates of the psychometric properties of this expanded scale have been reported. Although this expanded questionnaire examines a number of additional grief dimensions, the lack of data related to the validity and reliability of the tool prevents its recommendation for clinical use. As well, the length may be prohibitive in many clinical or research contexts due to participant fatigue or burden. Hogan Grief Reaction Checklist (HGRC) The HGRC is a 61 item instrument structured as a five-point Likert-type scale [63]. The development of this questionnaire was based upon interview data obtained from bereaved adults (number of informants, length of time since death of relative, and type of death not provided). Content analysis of the interviews resulted in the identification of six categories: despair, panic behaviour, blame and anger, disorganisation, detachment, and personal growth. An initial set of 100 items was developed and presented to a series of four purposively selected focus groups to maximise the range of types of respondents (type of death, relationship to deceased). Results from this analysis were presented to a panel of experts (nursing graduate students) to assess content validity of the items. Percent agreement was used to determine retention of items (>80% preset criterion). The 100 item HGRC was then administered to 586 adults recruited through a range of bereavement support groups (Level III-2). Factor analysis was conducted using principal axis factoring with varimax rotation, resulting in a revised 61 item scale with six subscales as described above. Internal consistency estimates using Cronbachs alpha coefficient ranging from 0.79-.90 were reported. Estimates of test-retest reliability were also reported and range from 0.56 to 0.84 (p<0.001). Construct validity was assessed with a subsequent sample of 209 parents from mutual bereavement support groups. Confirmatory factor analysis was undertaken with this sample, verifying the six-subscale structure. The HGRC subscales were compared to subscale scores on the Texas Revised Inventory of Grief (TRIG) (Faschingbauer, 1981), Grief Experience Inventory (GEI) (Sander et al., 1985), and Impact of Event Scale (IES) (Horowitz et al, 1979) to determine convergent and divergent validity. Statistically significant correlations were reported and were clinically interpretable. The instrument was also able to detect changes in bereavement responses over time. A total score for the HGRC cannot be calculated; therefore, each subscale is examined independently and appears to measure different aspects of a grief response. A range of grief theories are offered as rationale for the various types of grief responses to each subscale (i.e. Dual Theory, Attachment Theory, etc). The instrument consists of a list of thoughts and feelings that the bereaved person may have experienced since their loved one has died. Respondents are asked to consider their feeling in the previous two weeks. In a subsequent study Hogan and colleagues (2003-2004) [53] (Level III-3) undertook a study to empirically test the complicated grief criteria agreed upon in 2001 by a panel of experts [28]. Complicated

32

grief was conceptualised as a uni-dimensional factor with two categories: separation distress (necessary condition) and traumatic distress (sufficient condition) (Prigerson & Jacobs, 2001). The first criterion, separation distress, represents yearning, longing, and loneliness; and the second criterion, traumatic distress, represents the bereft feeling stunned, dazed, empty and in shock (Prigerson & Jacobs, 2001). A self-selected sample of 166 bereaved parents was recruited and completed a mailed protocol which included the Complicated Grief Disorder criteria (CGD) (Prigerson & Jacobs, 2001), the HGRC (to index normal grief reactions) and the Beck Depression Inventory (BDI-II) (Beck et al, 1996). The pattern of correlations between the CG factors of separation distress and traumatic distress, and the normal grief factors of the HGRC were significant and large (-0.41-0.86, p values not provided). Hogan and colleagues argue that these findings challenge the notion that CG and normal grief are conceptually distinct. These results are difficult to interpret because a new conceptual framework (CG Criteria) is being tested against an instrument that is also new and may not be conceptually distinct. A number of subscales within the HGRC appear to be based upon different theoretical formulations. Testing of the HGRC has been limited and has, for the most part included parents of deceased children. The timeframe for assessing their responses to the HGRC has varied and assessments have been cross-sectional. A review of these papers indicates that it would be difficult to recommend the HGRC as an instrument to measure CG. Further longitudinal work is needed with additional studies to confirm the predictive validity of the tool to confirm that is indeed measuring normal grief and not complicated grief. At present, empirical evidence does not allow firm conclusions on what exactly is being measured. Grief Evaluation Measure (GEM) The GEM is a new instrument designed to screen for the development of complicated mourning response in a bereaved adult [64]. The instrument is comprised of seven sections using quantitative and qualitative questions to assess risk factors, including the mourners loss and medical history, coping resources before and after the death, and circumstances surrounding the death. It is designed to provide an in-depth evaluation of the bereaved adults subjective grief experience and associated symptoms. Reliability and validity testing was undertaken (Level III-3) with two samples of adults (n=23 and n=92 respectively). These individuals were recruited from a range of clinical and support settings. Estimates of the internal consistency reliability of the tool are high (0.88-0.97). However, this reliability testing was based on the small sample of 23 participants, limiting conclusions about this psychometric property. As well, the GEM demonstrates good concurrent validity when assessed against established measures of trauma as measured by the Inventory of Traumatic Grief [27] Impacts of Events Scale [75] physical and psychiatric symptoms as indexed by the SF36 [76]and Treatment Outcome Package [77]. The tool was also found to be predictive of mourner adjustment one year following the initial assessment. The extent to which the tool is able to discriminate respondents according to the severity of their grief responses warrants further testing. The instrument is also very long, which may prohibit its use in this present form in clinical practice.

33

Core Bereavement Item (CBI) This 17 item tool was developed using qualitative data from a longitudinal study of three groups: bereaved spouses, bereaved adult children and bereaved parents [66]. Items were derived from the literature and from the clinical experience of the investigators (Level III-2). A principal components analysis with varimax rotation revealed seven subscales, three of which measured frequently experienced phenomena in the bereaved. These three subscales (images and thoughts, acute separation, grief) formed the basis of a single measure, labelled the Core Bereavement Items (CBI), demonstrating high reliability (Cronbachs alpha coefficient of 0.91) and sound face. The sub factor scores discriminated among bereaved parents, bereaved spouses, and bereaved adult children in the order of severity of symptoms. Inventory of Complicated Grief-Revised (ICG-R) This tool was developed to assess a distinct cluster of symptoms that have been found to predict longterm dysfunction [28]. This inventory is based on previous empirical literature that confirms the distinction between complicated grief, anxiety and depression [27]. Prigerson and colleagues developed and tested this 19-item inventory with 97 elderly bereaved men and women (Level 4). Exploratory factor analysis indicated that the ICG measured a single underlying construct of complicated grief. High internal consistency (Cronbachs alpha coefficient of 0.92 0.94) and test-retest reliability estimates (0.80) were obtained. The ICG total score correlated well with measures of depressive symptoms and a general measure of grief providing evidence for the validity of the tool. Respondents with ICG scores greater than 25 were significantly more impaired in social, general, mental and physical health functioning and in bodily pain than those with ICG scores less than or equal to 25. The inventory demonstrated good convergent and criterion validity and appears to be an easily administered tool to assess for complicated grief. The researchers acknowledge the limits of this cross-sectional study and call for longitudinal research to determine the extent to which the ICG is able to predict individuals at risk for complicated grief responses over time. Revised Grief Experience Inventory (REGI) This 22-item six-point scale is based upon Parkes framework of bereavement (Parkes, 1972) and includes four subscales: Depression (six items), Physical Distress (seven items), Existential (six items), and Tension/Guilt (three items). Internal consistency estimates as measured by Cronbachs alpha coefficient range from 0.72 to 0.87 with an estimate for the total scale of 0.93. Relationships between total and subscale scores according to demographic variables reported in previous literature were confirmed, providing evidence of construct validity of the tool [65]. Principal components analysis yielded a four factor solution, confirming the internal theoretical structure of the instrument. Correlations of elapsed time since the loss and length of illness with the RGEI resulted in mixed findings. The correlation between time since loss and the overall RGEI score was small but significant (r=0.10, p=0.02). Time since loss was also positively correlated with responses on the existential and physical subscales (r=0.10, p=0.02; r=0.12, p=0.01). The associations between length of time living with diagnosis and responses on the guilt

34

subscale was significant (r=0.10, p=0.01). The relationship between length of time living with diagnosis and the existential subscale was non-significant. The extent to which this instrument is able to discern a complicated grief response has not been reported. Severity of response over an extended period of time may be a useful way of indexing some level of complicated grief reaction. This hypothesis warrants further testing. The instrument is reported to be concise and simple to administer (Level IV). Bereavement Risk Index (BRI) Kristjanson and colleagues [78] recently undertook a study to test the validity, reliability and feasibility of using a modified version of Parkes (1993) [67] Bereavement Risk Index (BRI) and bereavement support protocol in an Australian home hospice care setting. A prospective, descriptive study was used (Level III2). One hundred and fifty bereaved family members participated. Bereaved family members were classified as high, medium or low risk and received a structured bereavement support protocol based on their level of risk as measured by the BRI. Results indicated that a shorter 4-item version of the BRI was more internally consistent than the longer version and demonstrated good predictive validity when correlated with outcome measures at three months following the patients death. The modified 4-item BRI demonstrated acceptable reliability and validity and was brief and simple to use. Nurses were able to use the instrument with minimal training and were able to adhere to a matched bereavement support protocol. Three instruments were identified that endeavour to assess grief in specific contexts: Grief Experience Questionnaire, Perinatal Grief Scale, and the Perinatal Bereavement Scale. Grief Experience Questionnaire (GEQ) The GEQ is a 55-item questionnaire that measures individual grief elements common within the experience of suicide survivors including physical reactions, general grief reactions, search for an explanation, loss of social support, stigmatisation, guilt, responsibility, shame, rejection, self-destructive behaviour, and reactions to unexpected death [68]. Initial results with the GEQ suggest that it has potential to differentiate grief reactions experienced by suicide survivors from those experienced by survivors of accidental deaths, unexpected natural death, and unexpected death. The tool is comprised of 11 subscales, but the structure of GEQ has not been confirmed by factor analysis. Perinatal Grief Scale (PGS) The PGS is a well-known device for measuring the intensity of affective symptomatology following the loss of a baby [69]. A short version of the PGS with 33 items was developed (PGS-S) [79]. The scale has demonstrated concurrent validity when compared with the Symptom Checklist-90 (SCL-90) depression scale [69] and convergent validity with other measures of parental distress [80].

35

Summary In summary, a number of instruments have been developed and tested in an effort to measure grief responses and identify those who may be at risk for a more complicated grief response. Overall, the instruments demonstrate good estimates of reliability and validity. They range in length with some being brief and simple to use and others lengthy and potentially more burdensome. The extent to which the instruments are able to measure complicated grief responses has not been well documented given the cross-sectional nature of the study designs. The match between the theoretical constructs being measured and the measurement model used to index the construct is poorly articulated, making it difficult to determine the validity of the instruments. The Texas Revised Inventory of Grief [81] and the Inventory of Complicated Grief-Revised (ICG) [27] appear to be theoretically grounded and empirically sound instruments that have potential for assessment of complicated grief in both research and clinical practice. As well, the Core Bereavement Item (CBI) [66] is a less widely used instrument, but one that demonstrates excellent psychometric properties, is brief and has been developed and tested within the Australian context and may be a simple tool for assessing a range of grief responses. The extent to which it assesses complicated grief has not been examined and warrants further testing. Recent testing of the BRI is encouraging and may provide a practical means of screening the broader bereaved population in a simple, brief manner. No longitudinal studies to determine the extent to which the tool screens for CG have been undertaken and further testing is clearly warranted. Recommendations for Measures in Complicated Grief

It is recommended that clinicians/counsellors assessing individuals for complicated grief use the criterion for complicated grief as specified by Prigerson and colleagues in the Inventory of Complicated Grief- Revised (2001)[28].

It is recommended that primary care health professionals screen bereaved individuals for possible complicated grief if they present with a set of symptoms identified as persistent (beyond six months post-death) and severe (marked intensity or frequency, such as several times daily).

However, further empirical work to evaluate the reliability, validity, sensitivity, specificity, and diagnostic efficiency of criteria proposed for CG is required.

36

CHAPTER 4: COMPLICATED GRIEF AS A CONSTRUCT DISTINCT FROM ANXIETY, DEPRESSION, AND PTSD
It is recognised that Complicated Grief is not the only complication that may follow from bereavement. Other psychiatric disorders such as Major Depression or Posttraumatic Stress disorder may develop in response to the death of a significant other. Current research has focused on distinguishing CG from depressed mood and anxiety. The distinction between complicated grief and bereavement-related depression or anxiety is made even more difficult by the tendency for the three syndromes to occur simultaneously. Prigerson argues that depressed mood, psychomotor retardation, and damaged self-esteem are all depressive symptoms whereas symptoms of yearning, disbelief about the death, difficulty moving on/a sense of feeling stuck, feeling detached and bitter and agitated about the death are all specific indicators of Complicated Grief. In this section we reviewed studies that were designed to investigate if CG is distinct from other psychiatric disorders in terms of clinical phenomenology, aetiology/correlates, outcomes, clinical course, and response to treatment. This distinction has implications for screening, diagnosis, treatment and health policy decisions. Complicated Grief and other mental disorders subsequent to bereavement An early study by Kim and Jacobs (1991) explored the relationship between pathologic grief and psychiatric disorders (Level IV). Of 25 bereaved spouses who were referred following a psychiatric interview 16 met the criteria for pathological grief [21] (Level IV). All had been bereaved over 6 months and 52% had passed the first anniversary of the death. Data were collected on several self-report measures including the 20-item Centre for Epidemiologic Studies-Depression Scale (CRS-D); the 10 item Psychiatric Epidemiology Research Instrument anxiety scale (PERI) along with measures of separation distress, numbness/disbelief, the short form of the Texas Grief Inventory and a self-rating severity of grief scale. Relevant sections of the Structured Clinical Interview for DSM-III were used to assess major depression, panic disorder, and generalised anxiety. The structured assessment of pathologic grief was based on descriptive data developed specifically for the study. The group with pathologic grief (n = 16) was significantly more likely to be diagnosed with major depression (98% vs. 33%, Yates-corrected =
2

7.65, p <0.01) and scored significantly higher on separation distress measures (t = 3.0, p 0.001), anxiety (t = 2.9, p 0.01) and depression (t = 2.4, p 0.05). Comparisons between the group with pathological grief and the group without on various risk factors found a significant difference for acute cardiac causes
2 of spouses death which was observed significantly more in the pathologic grief group ( = 5.8, p0.05).

The small sample makes any interpretation of findings difficult.

37

Prigerson and colleagues (1995) [24] investigated whether or not CG could be distinguished from bereavement-related depression (Level III-3). Baseline data were collected from 82 widowed elderly subjects participating in a study of changes in sleep physiology. Data were collected 3 to 6 months after the death of the subjects spouses. CG was measured using subscales from the following measures: the Hamilton Depression Scale, Brief Symptom Inventory, Grief Measurement Scale and the Texas Revised Inventory of Grief. A Principal-Components Analysis revealed a complicated grief factor and a bereavement-depression factor. Seven symptoms constituted complicated grief: searching, yearning, preoccupation with thoughts of the deceased, crying, disbelief regarding the death, feeling stunned by the death, and lack of acceptance of the death. The first component accounted for 26% of the variance and the second component for 20% of the variance. The agreement between significantly impairing complicated grief and syndromal-level depression was moderate (Cohens Kappa = 0.34, p < 0.01) suggesting that symptoms of complicated grief may be distinct from depressive symptoms. However, the limits of this study include the relatively small sample (56-82 subjects) the potential for an underestimation of CG reactions due to selection bias; and because depressed participants were treated with nortriptyline, the results could not be generalised to untreated populations [24]. Recognising these methodological limitations, Prigerson and colleagues [23] proceeded to replicate their 1995 study a year later in a non-clinical, community-based sample of 150 widowed individuals (Level III3). Participants were part of a larger study of 494 women whose spouses were admitted to hospital with life threatening illness. Interviews were conducted at the time of admission to hospital, and at 6 weeks, 6 months, 13 months and 25 months. Only those women whose spouse had died at the time of the 6 month interview were included in this analysis. Items for depression and anxiety factors were derived from the Center for Epidemiologic Studies Depression Scale (CES-D) Scale) and the Psychiatric Epidemiology Research Interview. Items for CG were obtained from the Grief Measurement Scale and selected to approximate as closely as possible the items contained in the Inventory of Complicated Grief. The factor structure was determined by Principal Axis Factoring with iterated commonalities based on squared multiple correlations and varimax rotations. Three factors emerged accounting for 90% of the variance. The symptoms of complicated grief (e.g. yearning, hallucinations and preoccupation) achieved high loadings on the first factor (values ranging from 0.50 to 0.82). The symptoms of depression (e.g. depressed mood, the blues) loaded on the second factor (0.57-0.71) and the symptoms of anxiety (anxiousness and restlessness) loaded on the third (0.46-0.81). Correlations between the summary scores of the items in the CG factor and the depression and anxiety factors were 0.15 and 0.33 respectively (pvalues not reported). The correlation between the items included in the depression and anxiety factors was 0.13 (significance levels not reported). Results confirmed their previous findings that CG symptoms are distinct, but not completely independent from bereavement related depression and anxiety [23].

38

In a further study to extend their research undertaken in 1995 and 1996 Prigerson and colleagues [82] tested the validity and utility of distinguishing symptoms of anxiety from those of depression and grief in 56 recently spousal bereaved elders (Level III-3). A sub-set of complicated grief symptoms from the ICG was available for selection and was used together with the Grief Measurement Scale, the Texas Revised Inventory of Grief. The Hamilton Rating Scale for Depression and the Brief Symptom Inventory anxiety sub-scale was used. A confirmatory factor analysis showed the BSI anxiety sub-scale loaded on the anxiety factor (0.780, 0.858, and 0.739 respectively). The symptoms of yearning and searching for the deceased, preoccupation disbelief, crying and being stunned by the death, all had factor loadings on the grief factor between 0.433 and 0.750. The symptoms of depressed mood had factor loadings between 0.623 and 0.800 on the depression factor. All inter-factor correlations were significant at p < 0.05 level. Path analyses showed that symptoms of anxiety, depression and grief all declined significantly over time (6 to 18 months post-loss) except for the "stunned by the death" measure (p<0.05). These findings have important implications for diagnosing CG because feelings of being stunned or shocked by the death have previously been seen to be an initial grief reaction that subsequently declines. Boelen and colleagues [83] sought to replicate Prigersons study [82] using a Dutch population of outpatients who had sought help after bereavement. They hypothesised that symptoms of traumatic grief (their definition) are distinct from those of bereavement-related depression and anxiety. One hundred and three participants completed the Dutch version of the Inventory of Traumatic Grief (Level IV). Depression and anxiety were measured with the Symptom Checklist. Symptoms were analysed using Principal Axis Factor analysis. Three distinct symptom clusters were replicated: traumatic grief, bereavement-related anxiety and bereavement-related depression. The first factor accounted for 30% of the variance, the second factor 18% and the third factor 16%. Symptoms of traumatic grief loaded on factor 1 (traumatic grief) (0.62 to 0.84); symptoms of anxiety on factor 2 (bereavement related anxiety) (0.74 0.78) and symptoms of depression on factor 3 (bereavement related depression) (0.62 0.80). The results indicate that complications of bereavement may include symptoms of traumatic grief that constitute a clinical entity distinct from bereavement-related depression and anxiety. The Dutch findings are comparable to studies by Prigerson and colleagues [82] in the USA suggesting cross-cultural generalisability. This Dutch group of investigators undertook a confirmatory factor analysis study to replicate their earlier findings that CG, depression and anxiety are distinct syndromes [84] (Level IV). They hypothesised that a limitation of earlier studies is that they relied on exploratory factor analysis to evaluate the latent structure of post-loss symptoms, a method that does not allow for the comparative evaluation of the fit of competing models of the latent structure [84, p. 2175]. Additionally, earlier studies did not distinguish the distinctiveness of the three symptom clusters across subgroups of bereaved individuals.

39

In this study, a sample of 1,321 bereaved individuals were recruited by grief counsellors, clergy and therapists and through an advertisement on a Dutch internet site [84]. Six hundred participants completed the questionnaires on a Dutch Internet site (43% response rate) and 260 returned a mailed questionnaire (53% response rate). The mean age was 43 years and the majority (82%) were female. The Dutch version of the Inventory of Complicated Grief was used that is an extended version of Prigersons ICG. CG, depression, and anxiety symptoms loaded on separate factors were superior to a one-factor model, had good fit, and was stable across subgroups even when severity of CG differed between subgroups. Correlations between factors were 0.78 for complicated grief with depression (p<0.001), 0.58 for complicated grief with anxiety (p<0.001) and 0.78 for depression with anxiety (p<0.001). The grief scores of victims of violent losses (mean 81.01, SD=21.33) on the ICG were significantly higher than those of victims of non-violent losses (mean 78.15, SD=20.35) (t=1.97, df=13.18, p<0.05). Additionally, the grief scores of bereaved partners (mean 80.18, SD=20.08) and parents (means=80.62, SD=22.44) were higher than those for other mourners (mean=76.32, SD20.40) (F=6.10, df=2, 1313, p<0.01). The authors acknowledged the following limitations: subjects were drawn from different sources, there was a relatively low response rate with subgroup analysis relatively small, lowering the power for subgroup analyses. Ogrodniczuk and colleagues (2003) examined whether dimensions of complicated grief could be distinguished from dimensions of depression [85]. Data from 398 patients from two previous studies by the same Edmonton Trial group in Canada were analysed (Level III-3). Measures previously gathered included the Texas Revised Inventory of Grief, a set of Pathological Grief items adapted from the work of Prigerson, the Impact of Events Scale and the Social Adjustment Scale. A Principal Component Analysis (PCA) found that among a sample of psychiatric out-patients, CG symptoms emerged as a distinct set of dimensions that were relatively independent of depressive symptoms. The PCA identified five dimensions accounting for 53% of the variance. The first dimension (grief symptoms) accounted for 15%, the second 13% (grief experience), the third 11% (depression-cognitive), the 4th 8.3% (grief avoidance) and the fifth accounted for 7% of the variance (depressive-somatic). Estimates of internal consistency using Cronbachs alpha coefficients for each dimension ranged from 0.83 to 0.94. The items that accounted for most of the variance included those that have been labelled complicated grief (i.e. intrusive thoughts and feelings about the lost person, yearning and searching for the lost person and numbness about the death) supporting the Diagnostic Criteria for CG reported by Prigerson in Chapter 2. Summary The studies reviewed in this section were consistent in their use of assessment scales (The Inventory of Complicated Grief) and measurements were taken over a period of time (e.g. 6 weeks to 25 months). Although early research used smaller samples, more recent studies using larger samples confirm earlier findings. Factor analysis results corroborate the construct validity of the term, complicated grief and the dimensions that comprise the concept. There is evidence that complicated grief can be distinguished from

40

depression and anxiety [23, 24, 82-84]. These studies demonstrate that although CG is frequently comorbid with other psychiatric disorders, a diagnosis based solely on the DSM-IV disorders of major depressive disorder and anxiety risks missing many cases of CG. There is some evidence to support the cross-cultural generalisability of the construct. Recommendation: Complicated Grief as a Construct Distinct from Anxiety, Depression, and PTSD It is recommended that clinicians endeavouring to diagnose an individual for potential

CG be alert to the distinctions between CG from other DSM-IV disorders of Major Depressive Disorder, Post Traumatic Stress Disorder and generalised anxiety.

41

CHAPTER 5: VIOLENT AND TRAUMATIC DEATH

Typically, violent death is characterised by one of three unnatural modes of dying: suicide, homicide, or accident [86]. There is consensus that the death of a loved one by violent death is associated with poor recovery for bereaved individuals [87-89]. From a trauma perspective, losses from violent deaths are likely to promote reactions resembling post-traumatic stress disorder (PTSD). Some researchers suggest that psychological trauma involves a violation of basic assumptive worldviews connected with the individuals survival and that of the social group [90]. Deaths by suicide, homicide, or accident are commonly conceptualised as a traumatic event that can lead to PTSD, thereby causing profound complications in grieving and difficulties with meaning-making [91]. Typically, loss by traumatic means is conceptualised as a traumatic stressor event that can lead to posttraumatic stress disorder (PTSD). The boundaries between traumatic stress and PTSD, complicated or chronic bereavement as a mental health outcome independent of the nature of the loss, and traumatic bereavement (loss by traumatic means) and traumatic grief (the unique mixture of trauma and loss) have not been examined fully. Researchers suggest that complicated grief following violent death generally triggers two concurrent but distinct syndromes: separation distress as a response to the lost relationship and traumatic distress in reaction to the manner of the dying [28, 92, 93]. Separation distress includes thoughts of reunion, feelings of longing, and searching behaviours for the deceased. Traumatic distress includes re-enactment thoughts, feelings of fear and behavioural avoidance. These two distress responses are often mixed in the course of complicated grief. Raphael and colleagues propose that two different reactive processes occur and describe these phenomena in terms of specific, frequently contrasting core reactions (e.g. affective reactions, avoidance phenomena and reactive processes) [94]. They argue that the phenomena differ in important ways. In terms of pathology, trauma leads to traumatic stress reaction and perhaps the development of PTSD, while bereavement leads to grief and perhaps chronic grief disorder [29, 94]. Rando described traumatic bereavement as one variation of complicated mourning, contending that any differences between uncomplicated acute grief and traumatic stress response are primarily in content and degree, and not necessarily in underlying dynamic processes [95]. However, some investigators have included the range of non-traumatic as well as traumatic bereavement experiences in developing their frameworks for CG [22, 27, 28]. They define the bereavement experience for inclusion in their criteria as bereavement (the loss of a spouse, other relative or intimate partner). No distinction is made between the type of death (traumatic versus non-traumatic types), rather they define complicated grief as a function of the intensity and symptomatology of distress [22, 27, 96].

42

The systematic review of the literature identified 10 studies published between 1995 and 2004 meeting the inclusion criteria that investigated the influence of violent or traumatic death on grief. These studies focussed on suicidal ideation and predictors of psychological distress in traumatic death. Complicated grief and suicide ideation Prigerson and colleagues (1999) [97] examined the influence of traumatic grief (their term) on suicidal ideation in 76 young adults who had experienced the suicide of a friend on average 6 years previously (Level III-3). Measures included the ICG, the Beck Depression Inventory and a modified version of the Beck-Kovacs Scale for Suicide Ideation. Logistic regression models estimated the main and interactive effects of syndromal traumatic grief and depression on the likelihood of suicidal ideation. Twenty participants (15%) with symptomatic levels of CG were found to be five times more likely to report suicidal ideation than participants with non-symptomatic levels (p = 0.006, OR5.08, CI 1.48-17.50). Levels remained high after controlling for depression, gender and time since death. Further longitudinal studies are needed to determine whether CG and depression are preludes to suicidal ideation. Mitchell and colleagues (2005) [98] examined CG and suicide ideation among 60 survivors of suicide of family member or a significant other (Level III-3). Participants were part of a larger crisis intervention study to examine the efficacy of a critical incident stress de-briefing intervention for survivors during the acute phase of bereavement (1 month after the death). Data were collected prior to the intervention. Participants completed the ICG, the Beck Depression Inventory and the suicidal ideation component of the BDI. Twenty-six of the 60 of the participants were classified as having CG. CG was significantly associated with suicidal ideation with subjects 10 times more likely to report suicidal ideation, after controlling for depression. CG was highly predictive of suicidal ideation in suicide survivors with 83.3% predictive success. Limitations include a small and homogenous sample, with possible selection bias because participants were taking part in crisis intervention study. In addition, family network effect was not examined fully. Longitudinal analyses are needed to draw conclusions about causality. Predictors of psychological distress in traumatic death Dyregrov et al. (2003) [99] compared the outcome and predictors of psychological distress of parents in Norway bereaved by young suicide, sudden infant death syndrome and child accidents (Level III-3). 232 parents completed the Impact of events Scale, the General Health Questionnaire and the Inventory of Complicated Grief 18 months after the death of their child. Between 57 and 78% of parents scored above the cut-off point for complicated grief. Self-isolation was found to be the best predictor of psychosocial distress and being female predicted complicated grief in the suicide and SIDS samples. There was no evidence of suicide survivors having greater difficulties in adapting to the death compared with survivors of SIDS or accidents. A noted limitation included the small sub-samples. This study supports the notion that the unique features of traumatic death, when present in suicide or in any other traumatic loss account for much of the variance in bereavement outcome in comparison to natural causes of death.

43

In 2001, Melhem et al. [100] administered the ICG to a group of 23 bereaved patients who presented for treatment for traumatic grief (their words) and were participating in a pilot study of exposure-based psychotherapy. (Level IV). Traumatic grief was measured by the Inventory of Complicated Grief. The study examined the rate of DSM-IV Axis 1 disorders in the sample. Measures also included the Structured Clinical Interview for DSM-IV, the Beck Anxiety and Depression Inventory and the Posttraumatic Diagnostic Scale and the Work and Social Adjustment Scale. Most subjects met criteria for a current or lifetime Axis 1 diagnosis. 52% (n=12) met criteria for current MDD and 30% (n=7) for current PTSD. Similar results were found for panic disorder. ICG scores and functional impairment were higher among patients with more than one concurrent Axis 1 diagnosis (F=3.48, df=3.17; p=0.039). The mean ICG score for patients with two diagnoses and for those with three or more was 42.3 (n=6) and 52.7 (n=3) respectively compared with 39 and 35 for those with no and only one concurrent diagnoses. ICG scores were also significantly correlated with self-reported anxiety (BAI; r=.55, n=16, p =0.028), selfreported depression (BDI; r=.53, n=16, p=0.035) and PTSD (r=.65, n = 14, p = .011). Results suggest that co-morbid depressive disorder and PTSD may be prevalent in patients presenting for treatment of traumatic grief. Limitations of this study include the fact that participants were referred or they were seeking help, making the level of psychiatric co-morbidity found to be higher than a general community sample. The sample was small precluding examination of group differences and there was no comparative group with low ICG scores [100]. Pivar and Field (2004) [101] examined the prominence and status of grief-specific symptoms from trauma and depressive symptoms in a sample of Vietnam veterans with PTSD (Level IV). One hundred and fourteen male Vietnam-era combat veterans admitted to an in-patient rehabilitation unit for treatment of PTSD completed measures such as the Texas Revised inventory of Grief, the Core Bereavement Items, the Mississippi Scale for combat-Related Post-Traumatic Stress Disorder, the Beck Depression Inventory and qualitative data on combat experiences were collected. Principal Component Analyses were conducted on grief, PTSD and depression subscale scores. Grief-specific symptoms accounted for 30% of the actual variance, PTSD accounted for 15% and depression explained 14% of the variance indicating that the grief-specific symptom subscales were distinguished from the PTSD and depression subscales. Multiple regression analyses was conducted to determine the unique relationship between each of the symptom measures and the extent of attachment to men in the unit during the war (r 2 = .07 (adjusted r 2 =.05, F(3,110) = 2.80, p<0.05). Grief symptoms were significantly positively associated with attachment to men in the unit (t=2.79; p<0.01) whereas no relationship was found for trauma symptoms or depression with attachment. Grief symptoms were also significantly associated with the number of combat losses (t = 3.14; p<0.001) and closeness to a buddy (t=4.35; p<0.01) but not with trauma or depression. The results provide support for the existence of grief-specific symptoms as distinct from other

44

war trauma-related symptoms. Limitations include the retrospective nature of the measures in assessing experiences of loss in combat that occurred 30 years ago, and the generalisability of the results. Predictors of complicated grief in traumatic death Melhem and colleagues (2004) examined predictors of complicated grief among adolescents exposed to a peers suicide (n=146) [102] (Level III-3). Multivariate analyses showed that complicated grief at 6 months was significantly associated with gender (female) (B 1.6, p=0.004, CI 1.7 13.8). PTSD at 6 months was significantly associated with a previous history of anxiety disorders (B 2.2, p=0.003, CI 2.238.7); feeling that they could have done something to prevent the death (B 2.1, p=0.001, CI 2.2 26.9); financial problems (B 1.6, p=0.006 CI 1.6 14.8) and a previous history of depression (B 1.5, p=0.012, CI 1.4 13.7). The presence of these variables was associated with an 81% risk of CG. Major depression at 6 months was significantly associated with gender (female) and a previous history of depression (B 1.5, p=0.012, CI 1.4 13.7). Limitations included a sample consisting of friends and acquaintances of suicide victims with high rates of previous psychiatric problems, the 6 month assessment may have been affected by recall bias leading to under-reporting, and their current mental status leading to over-reporting. Psychological outcomes of traumatic death Melhem and colleagues (2004) [103] then went on to describe the symptoms and course of traumatic grief among these adolescents who had been exposed to a peers suicide, and the relationship between grief and depression, and posttraumatic stress disorder (PTSD) in this population (Level III-3). The Texas Revised Inventory of Grief (TRIG) was administered at 6, 12 to 18, and 36 months; and the Inventory of Complicated Grief (ICG) was administered at a 6 year assessment. Principal Component Analysis on the TRIG resulted in two factors explaining 64% of the variance at six months. Similarly, factor analysis at 1218 months resulted in similar 2 factor solutions, with the cluster of symptoms loading on factor 1 reflecting traumatic grief whereas symptoms loading on factor 2 appeared to reflect a type of separation distress component. Correlations between scores on the traumatic grief factor at 6, 12-18, and 36 months after the suicide with scores on ICG administered 6 years after the suicide were 0.46 (p<0.001), 0.64 (p<0.001), and 0.72 (p<0.001) respectively. Of the participants who were depressed within one month of exposure to suicide (n=59), 61% (n=36) continued to be depressed at 6 months, CG was defined as scoring in the upper 25% of factor 1 at 1 and 6 months. Of the participants who met the criteria for CG at 6 months (n=29), 13.8% (n=4) continued to meet the criteria 12-18 months, and 7% (n=2) met the criteria throughout the assessment period. The presence of CG was found to be independent of depression and PTSD. CG at 6 months predicted depression at 12-18 months, (odds ratio=1.15, p=0.02) and CG at 12-18 months predicted depression at 36 months. Among participants with CG, 41.7%, 50.0% and 22.5% had PTSD at 6, 12 to18, and 36 months respectively with Pearson correlations of 0.50 (p<0.001), 0.61 (p<0.001), and 0.43 (p<0.001) respectively.

45

Ginzburg et al. (2002) [104] assessed patterns of grief reaction and adaptiveness in bereaved Israeli parents whose adult child had died in military service. Eighty-five parents completed a battery of questionnaires 2.5 years after the death (Level IV). Measures included the Texas Revised Inventory of Grief, the Symptom Checklist (SCL-90) and a modified version of the Psychological Adjustment to Illness Scale. One third of participants were classified as having prolonged grief reaction (n=31; 36%), one-third was identified as having an absence of grief reaction (n=28; 33%) and the remainder were divided into delayed (n=14; 17%) and resolved grief reaction groups (n=12; 14%). Results indicated that prolonged grief reactions and absence of grief were the most prevalent variants comprising one third of the sample respectively. Absent and delayed grief was associated with lower levels of psychosocial adjustment compared with prolonged reactions. For example, level of education (2 =24.84, df=9, p=<0.01) and religious attitudes (2 = 13.05, df=6, p<0.05) were associated with the type of grief reaction. However, use of multi-variate analyses revealed that these associations were not significant indicating that the association between general psychiatric symptomatology and type of grief reaction may not be explained by sociodemographic background. Multi-variate analyses found no significant association between occupation and social functioning and type of grief reaction. There are a number of limitations to this study. Firstly, the small sample, participants in the study were already participating in a support group organised by the Israeli Ministry of Defence. Secondly, the questionnaires were administered during one of the regular group sessions and parents participated on average 31 months after the death. Mitchell et al. (2004) [105] undertook a descriptive pilot study examining complicated grief in 60 adult survivors of suicide of a family member or significant other (Level III-3). Participants were part of a larger crisis intervention study (to examine the efficacy of a critical incident stress de-briefing intervention for survivors during the acute phase of bereavement) Assessment was taken within one month of the death (Level III-2). Complicated grief was measured by the Inventory of Complicated Grief. Closely related survivors (n = 27) experienced nearly twice the level of CG as distantly related survivors (n = 33) (F=47.66, p<.001). In particular spouses had significantly higher mean ICG scores than in-laws (p<.0001) and friends/co-workers (p=<.0001); parents had significantly higher mean ICG scores than in-laws (p = .004) and friends/co-workers (p=.002) and children had significantly higher mean ICG scores than in-laws (p=.013) and friends/co-workers (p=.006). Relationships classification to the deceased explained 43% of variance in CG scores suggesting that professional assessments and interventions should take into account the familial and/or social relationship of the bereaved to the deceased. The quality of the relationship was not measured. Saltzman et al. [106] used a pre-test-post-test design to evaluate the effectiveness of a school-based screening and group treatment protocol, trauma- and grief-focused group psychotherapy, for adolescents (n=26) exposed to community violence and trauma either due to losing someone to a traumatic death or witnessing a traumatic act (Level IV). Grief was assessed using the Grief Screening Scale and UCLA

46

Trauma-Grief Screening Interview. Group participation was associated with improvements in posttraumatic stress, CG symptoms (t7.38, p=0.015) and academic performance. However, this study had a small sample, no control group, and used a limited battery of measures to assess treatment outcomes. Interventions in traumatic death Although it is not the focus of this review to assess intervention studies in traumatic death, clinical observations and programs that have arisen out of the 9/11 terrorist attacks are worthy of mention. The broader literature on interventions after traumatic death has revealed that psychological de-briefing is the most common form of early intervention for recently traumatised people. However, a Cochrane Review reported that there is little evidence supporting its continued use with individuals who experience severe trauma [107. Based on available evidence, it is proposed that psychological first aid is an appropriate initial intervention, but that it does not serve a therapeutic or primary preventive function. When feasible, individual screening for PTSD is required so that targeted preventive interventions can be offered to those individuals who may have difficulty recovering on their own. Evidence-based CBT approaches are indicated for people who are at risk of developing post-traumatic psychopathology. Guidelines for managing acutely traumatised people are suggested and standards are proposed to direct future research that may advance understanding of the role of early intervention in facilitating adaptation to trauma {Litz, 2002 #37]. The extent to which exposure to trauma is associated with complicated grief was not measured in by the studies in this review [108]. The work by Harvey (1996) in the area of traumatic death or complex trauma using a recovery and resilience treatment approach has been recently re-visited in an effort to identify clinical approaches following the 9/11 attacks. She proposes that individual differences in posttraumatic response and recovery are the result of complex interactions among person, event, and environmental factors. These interactions define the interrelationship of individual and community and together may foster or impede individual recovery. The ecological model proposes a multidimensional definition of trauma recovery and suggests that the efficacy of trauma-focused interventions depends on the degree to which they enhance the person-community relationship and achieve "ecological fit" within individually varied recovery contexts. In attending to the social, cultural and political context of victimization and acknowledging that survivors of traumatic experiences may recover without benefit of clinical intervention, the model highlights the phenomenon of resiliency, and the relevance of community intervention efforts [109]. No recent empirical literature evaluating the effectiveness of this model was identified in our review. Moreover, the link between this intervention and complicated grief has not been empirically reported. The criteria for diagnosing complicated grief includes the requirement that symptomatology persists for at least six months, regardless of when those six months occur in relation to the death. This criteria is intended to apply to individuals who have experienced a non-traumatic grief. On the surface, this may

47

appear to be inconsistent with reports of longer-term responses to genocide and homicide. For example, following homicide, five years is considered a "normal" length of time for intense symptoms to continue. These symptoms can fluctuate with periods of time within that 5 years of no or reduced symptoms. In the first 12 months coping is often less of a problem, and exacerbations are often linked to issues around justice, perpetrator trials, similar events and meaning of life questions and secondary losses. Often studies only track people for a little over 12 months and most people do not show complicated reactions until late in the second year. Anecdotally, there are reports of a decrease in functionality, brain fog, situational anxiety, aggression, confidence issues, substance abuse, work-alcoholism, sleep disorders, problem solving problems, fatigue and suicidality. Therefore, it would be important in the event of a traumatic loss, for assessments of potential complicated grief to continue over a longer time period (J. Dunsmore personal communication). Summary The studies reviewed in this section were conducted with populations that had experienced a traumatic death and included adolescent, young adult and adult survivors of the suicide of a family member, friend or peer and those bereaved after armed conflict. The majority of the studies were consistent in the use of the Inventory of Complicated Grief. However, several had small samples and only one used a longitudinal design; hence, further study is required before any conclusions can be made about complicated grief and causality in traumatic death. Two studies found that CG was highly predictive of suicidal ideation in suicide survivors. Only one study examined predictors of complicated grief after suicide [102] and found complicated grief at 6 months was significantly associated with gender (female), participants feelings that they could have done something to prevent the death, and a previous history of depression. Three studies examined psychological outcomes after a traumatic death and complex trauma. Results indicated that individuals who met criteria for CG were at greater risk for depression and PTSD. These psychological effects continued for as long as 36 months in some instances. Proximity of the relationship to the deceased was also found to be associated with more complicated grief responses. However, a limitation in this study was that the quality of the relationship was not measured. Only one study reported the effectiveness of group participation following traumatic grief and found this to be associated with improvements in posttraumatic stress, CG symptoms and academic performance. Recommendations: Violent and Traumatic Death

Further research is required into the phenomenon of complicated grief as a mental health outcome independent of the nature of the loss.

48

Further research to examine situational factors (e.g. sudden, expected, traumatic, nontraumatic) associated with the death in the Australian context of complicated grief is warranted

Clinicians/counselors should be proactive in screening people for CG if they have experienced a traumatic and/or violent death because CG appears to be a predictor for suicidal ideation in these populations.

49

CHAPTER 6: RISK FACTORS FOR COMPLICATED GRIEF Over the past 20 years, potential risk factors for complications of grief have been identified from textbooks, theoretical papers and from observations in clinical practice. Our systematic review identified six studies that examined predictors of risk for CG within the construct, complicated grief as defined in this review. Two studies examined practitioners views of risk factors for CG. The majority of studies measured complicated grief using the Inventory of Complicated Grief or the Texas Revised Inventory of Grief and were within the construct of CG as defined in this review. Predictors of risk for complicated grief Silverman et al. (2001) [110] explored the effects of prior trauma and loss on the risk of developing psychiatric disorders in 85 recently widowed people (Level III-2). Traumatic Grief was measured by the Traumatic Grief Evaluation of Response to Loss (TRGR2L). Participants who reported prior adversities in their life were generally more distressed following bereavement than those who did not report adversities. In particular, adversities occurring in childhood such as death of a parent (OR=8.83, CI 1.9041.0, p<0.01) and abuse (OR could not be calculated) seemed to have a greater impact and were significantly associated with traumatic grief. Due to the cross-sectional design, recall bias could have inflated the associations between disorders and prior adversities. Also, because of the small sample and rarity of some adversities, the estimate of risk may not be reliable and the statistical power is too low to show a significant association. Chen and colleagues (1999) [111] examined gender differences in spousal bereavement on mental and physical health outcomes, by interviewing 92 future widows and 58 future widowers at the time of their spouses hospital admission and then at 6 weeks and 6, 13 and 25 month follow-ups (Level III-2). The modified Grief Measurement Scale was used and this included items contained in the Inventory of Complicated Grief. Findings suggested that high symptom levels of CG, depression and anxiety predicted different mental and physical outcomes for men and women. Widows had higher mean symptom levels for traumatic grief (F=10.33, df=1,94, p<0.01), depression, and anxiety at 6 weeks, 6 months, 13 months, and 25 months post-loss. Among widows, high symptom levels of traumatic grief was found to predict sleep changes at the anniversary of the death of the spouse (b=2.12, df=4, p<0.01, RR=8.39). For widowers, high symptom levels of traumatic grief predicted hospitalisation (b=0.28, df=4, p<0.01, RR=1.32), having a physical health event such as cancer, stroke or a heart attack (b=0.15, df=4, p<0.05, RR=1.16). Limits of this study include a reduced sample size due to stratification by gender; and the fact that the rarity of some health outcomes may make the estimate of risk unreliable. The following study examined patterns of bereavement following conjugal loss and associated predictors using the sampling frame of the Changing Lives of Older Couples (CLOC study). This was a prospective study of a two-stage area probability sample of 1,532 married individuals from the Detroit Standardized Metropolitan Statistical Area [3, 39, 112].

50

Bonanno et al. (2002) [112] examined bereavement patterns in relation to pre-loss predictors of complicated grief (Level III-3). The study gathered prospective data from 205 individuals (180 women and 25 men) several years prior to the death of their spouse and at 6- and 18- months post-death. Measures included the CES-D Scale for depression and grief symptoms were measured using the Bereavement Index, the Present Feelings About Loss Scale and the Texas Revised Inventory of Grief. Five core bereavement patterns were identified: common grief, depressed-improved, resilient, chronic grief, and chronic depression. Bonanno defined the chronic grief pattern as a change score based upon a low pre-loss depression score (as measured on the CES-D) and a grief reaction at six and 18 months bereavement. Appropriate grief measures were used to assess complicated grief, and this operational definition of chronic grief (whilst unique) is concordant with the theoretical definition of complicated grief used to guide this systematic review. The clearest bereavement pattern predictor of chronic grief was excessive dependency, both as dependency on the spouse (F=2.58, df = 4, n= 80, p < 0.05) and as a more general personality variable (F=3.30, df=4, n=80, p<0.05). Those in the chronic grief group also reported less instrumental support (F=3.17, df=4, n=80, p<0.05) and a greater likelihood of having a healthy spouse (14 of 32, 43.7%). This was assessed using Habermans (1978) standardized, adjusted residual statistic (HAR), resulting in the following outcome, HAR=3.6, p<0.001). Different interventions were suggested for the different bereavement patterns. The researchers hypothesised that in light of the three predictive factors found to be associated with a chronic grief response, people in this group may benefit more from cognitive and behavioural interventions. This hypothesis is based on a sound theoretical premise and the findings reported in this study provide a helpful foundation for future testing. The researchers acknowledge some limitations associated with the study. Data was gathered using selfreports and interview observations, rather than objective indicators of behaviours and health. They also caution that findings may not be generalisable to younger bereaved because the mean age of the sample was 72 years. Vanderwerker et al. (2006) [111] explored the associations between childhood separation anxiety (CSA) and CG later in life. The aim of this study was to explore the etiologic relevance of childhood separation anxiety to the onset of CG to Major Depressive Disorder (MDD), PTSD, and Generalised Anxiety Disorder (GAD). The Structured Clinical Interview for DSM-IV, Inventory of Complicated GriefRevised, and childhood separation anxiety items from the Panic Agoraphobic Questionnaire were administered to 283 recently bereaved community-dwelling residents at an average of 10.6 months postloss. Participants were either retired, widowed or inpatients at a local hospital. Findings showed that CSA was significantly associated with CG both bivariately (OR =3.3, 95% C.I. 1.3-8.2) and after controlling for

51

sex, level of education, kinship relationship, history of psychiatric disorders, and history of childhood abuse (OR = 3.2, 95% C.I. 1.2-9.0). Although CSA was bivariately associated with MDD, CSA was not significantly associated with MDD, PTSD, or GAD in the adjusted models. A study by van Doorn et al. (1998) [113] examined the relationship between marital quality and adjustment to the impeding death of a terminally ill spouse residing at home or in nursing homes in the Pittsburgh area, USA (Level IV). Study participants (n=59) were interviewed before the spouses death while they were caring for their terminally ill spouse, and after the spouses death at 3, 6 and 13 months post-loss, using semi-structured interviews. Grief was measured using the Inventory of Complicated Grief. Findings suggested that having a secure, supportive spouse and an insecure attachment style contribute independently, but not interactively, to the severity of CG symptoms (p<0.0001). These findings need to be confirmed in prospective studies of a larger number of caregivers of terminally-ill spouses who are followed from pre-loss to post-loss. In a study that investigated insomnia and complicated grief symptoms in 508 bereaved and 307 nonbereaved undergraduate psychology students, Hardison and colleagues [114] (2005) reported that 22% of the bereaved group and 17% of non-bereaved group reported insomnia (Chi squared=4.89, p<0.05) (Level III-2). The Inventory of Complicated Grief was administered to assess the severity of grief complications. For the insomniacs, middle insomnia was higher in bereaved group than the non-bereaved group (67% vs 50%, Chi squared=4.05, p<0.05) [114]. Bereaved insomniacs had significantly higher CG scores than bereaved non-insomniacs (t(492)=-2.93,p<0.01) Bereavement-related sleep variables (dreaming of deceased and ruminating about the deceased) were significantly related to CG symptomatology. Insomnia proved to be a significant predictor of CG (p<0.01) along with nature of death, whether violent or not, the younger age of the deceased, level of closeness with the deceased, recency of the loss, relationship to deceased and sex of bereaved with women showing greater grief. However, limitations included data being self-reported raising questions about accuracy; the population reflected college students; ethnic groups other than Caucasians and African Americans were not represented; and the majority of participants was female. Practitioners views on risk factors that may predict complicated grief In a descriptive study, Wiles et al. (2002) [115] examined the range of issues that a varied sample of GPs with access to practice-based counsellors take into account when making decisions about the referral of bereaved people (Level IV). Interviews were conducted with 50 GPs in two sites in England. The following topics were explored: views about normal bereavement and risk factors for bereavement problems; approaches to bereavement care; examples of bereavement issues with patients; criteria for referral to practice counsellor; the referral process and satisfaction with bereavement counselling. GPs held views about what a normal reaction to bereavement should be and used this template to identify cases of abnormal bereavement which might need referral for bereavement counselling. Three major

52

themes emerged: the nature of the death, where violent, unusual or traumatic deaths were looked upon as likely to result in an abnormal reaction among the bereaved; the level of social support provided by friends and family networks to the bereaved, and reaction to the death, the length of time since bereavement, and the use of props or drugs. In addition, GPs would refer patients who showed willingness to attend counselling and have realistic understanding of what can be achieved. The study did not provide a specific definition of complicated grief, but rather drew on GPs notions of abnormal bereavement. Findings indicate that GPs are highly subjective when making decisions about the referral of bereaved patients. Further education in assessment of individual at risk for complicated grief may assist GPs in understanding bereaved patients experiences and in developing their skills in making appropriate referrals. In a descriptive study, Ellifritt et al. (2003) [116] using clinical experience and a review of the literature developed a Bereavement Risk Questionnaire to rate 19 possible factors for assessing complicated grief among caregivers of seriously ill patients prior to the death of the patient (Level IV). The questionnaire was administered to 269 bereavement professionals (53% response rate), in ten states in the USA. Forty four per cent of respondents were social workers, 41% nurses, 13% chaplains and 13% counsellors. Median job experience was 4 years. Profession was not significantly associated with response to any of the 19 individual factors. Overall, most (70%) rated perceived lack of social support, caretaker history of alcohol/substance abuse (68%), poor caregiver coping skills (68%), caregiver history of mental illness (67%), patient was a child (63%) and caregiver experiencing a concurrent crisis (52%) carried a significant risk. Overall 61% chose perceived lack of social support and 47% rated poor coping skills of caregivers as top risk factors. They conclude that it is possible to assess bereavement risk thus allowing palliative care teams to allocate resources and services to those at greatest risk for complicated grief. The design may have influenced responses as a list of possible risk factors was provided to choose from. Summary Studies reviewed in this section were concordant with the definition of complicated grief used to guide this review. Risk factors specific to complicated grief suggest that insecure attachments play a crucial role. The predictors of complicated grief identified in these studies include adversities occurring in childhood such as death of a parent, childhood abuse [110] or childhood separation anxiety [111]. Gender was found to have a role, for example widowers with high levels of CG predicted hospitalisation and having a physical event such as cancer, stroke, or heart attack. Widows had higher levels of CG than widowers and this predicted sleep changes, anxiety and depression [111]. A clear predictor was found to be excessive dependency both as dependency on the spouse and as a more general personality [112] and having an insecure attachment style [113]. Insomnia was found to be a significant predicator of complicated grief along with the nature of the death (whether violent or not), the younger age of the deceased, levels of closeness with the deceased, recency of the death, relationship to the deceased and sex of the bereaved with women showing greater grief [114].

53

The limitations of these findings include potential selection bias, retrospective or selfreporting and generalisability. Interventions that promote secure alternative attachments to others and emotional reengagement are needed to address the detachment and disengagement that is symptomatic of CG.

Recommendation: Risk Factors for Complicated Grief

Further research is warranted to examine risk factors such as the role of attachment styles and cognitive functioning, using prospective, longitudinal designs and objective measures of CG.

54

CHAPTER 7: OUTCOMES OF BEREAVEMENT AND THE RELATIONSHIP TO COMPLICATED GRIEF

Our systematic review revealed eleven studies that assessed outcomes of bereavement and their relationship to CG and five studies that explored the effect of complicated grief on mental and physical health outcomes. These are summarised below. CG and Bereavement Outcomes Bonanno et al. (2004) [3] examined differences in how respondents (total of 185 widowed persons) in each bereavement trajectory group in the CLOC Study reacted to and processed the death (Level III-3). Using measures as indicated above, comparisons were made between resilient and depressed-improved individuals and common grievers, chronic grievers and chronically depressed individuals. Analyses suggested that the chronic grief group was the most likely to report searching for meaning at 6 months F(12,516) = 3.02, p<0.001) and at 18 months F(12,501) = 2.51, p<0.01) They also reported current yearning (AR=2.3,p<0.05) and current emotional pangs (AR=2.5, p<0.05) at six months post-loss and also thinking about (p<0.05) and talking about (p<0.05) the loss more often than did chronically depressed individuals. Chronic grievers also decreased significantly in the degree that they thought about (p<0.05) and talked about the loss (p<0.05) from 6 to 18 months bereavement. They were also significantly more likely to report finding meaning at 18 months post-loss (AR=2.6, p<0.05) relative to other participants, a pattern consistent with active engagement with the emotional aspect of bereavement. The high level of distress exhibited by the chronic grief group was due primarily to the cognitive and emotional upheaval surrounding the loss of a healthy spouse. Stated limitations were similar to the earlier study conducted in 2002. Boerner et al. [39] (2005) followed up the participants in Bonannos CLOC Study for 4 years to further examine patterns of distress or grief trajectory following conjugal loss in 92 older adults who initially showed high or low distress (Level III-2). Grief was measured by using the Bereavement Index, the Present Feelings about Loss Scale and the Texas Revised Inventory of Grief. Bereaved adults were assigned to five grief trajectory groups: common grief, depressed-improved, resilient, chronic grief and chronic depression. Repeated measures analyses of variance were conducted with one pre-loss and three post-loss assessments at 6 months, 18 months and 48 months post-loss. The chronically depressed group showed significantly higher scores in grief (t=87.00, -1.75, p< 0.05) and depression (t=16.82, -1.79, p< 0.05) over time supporting the hypothesis that signs of improvement would be more evident in the chronic grief compared with the chronically depressed group. The chronic grief group showed significant decrease over time in grief (t=9.00, 4.68, p < 0.01) and depression (t=9.00, 3.30, p< 0.01). Although the chronic grief group experienced a more intense and prolonged period of distress compared with other groups (eg., common grief group), improvements by 48 months suggest that this group does not remain chronically distressed as a result of the loss. However, no data is provided regarding the extent to which

55

this group or other participants may have sought or received bereavement interventions. In contrast, the chronically depressed group clearly demonstrated long-term problems, with little indication of improvement between 18 months and 48 months. There are two other limitations to this study: attrition and small sample size, which reduced statistical power and limited the investigation of long-term adaptation, particularly for the group with chronic grief and chronic depression. Further research is needed with larger samples to determine the long term trajectory for individuals with chronic grief. A study undertaken by Boelen et al. (2003) [83] in the Netherlands explored the role of grief reactions in contributing to emotional problems after bereavement, with 234 individuals who had been confronted with the death of a close relative (Level IV). The Negative Interpretation of Grief Scale (NIGS) was used to assess different negative interpretations of grief reactions and the Inventory of Traumatic Grief was used to assess symptom severity. Findings suggested that if mourners interpreted their grief reactions as indicating mental insanity, inadequate adaptation or personal incompetence, they were more likely to experience distress and discomfort (r=0.61, p<0.001). Mourners that assigned more negative meanings to their grief reactions were more inclined to avoid cues associated with the loss (r=0.09, F(26.39) p<0.001). They were also more likely to engage in rumination (r=0.09; F(30.48) p<0.001), thought suppression (r=0.05, F(15.77) p<0.001) and distraction (r=0.02, F(5.47) p<0.05). Results support the notion that negative interpretations of grief reactions, in themselves do not indicate disturbance, are likely to play a role in the development and maintenance of emotional problems after bereavement. This is because they influence the degree to which these reactions are experienced as distressing and consequently influence the degree to which mourners engage in avoidance strategies that are likely to impede recovery and may serve to exacerbate and prolong rather than ameliorate grief reactions. However, the cross-sectional design of the study precluded any causal interpretations and more insight would be obtained from prospective studies. Also, the self-selected nature of the sample provided another limitation as to problems being over- or under- represented and thus limited the generalisation of findings to the general population of bereaved individuals. Germain and colleagues (2005) [117] evaluated the severity of sleep disturbances in a group of 105 adults presenting with complicated grief with and without co-morbid major depression disorder and PTSD (Level IV). The Inventory of Complicated Grief was used to measure grief. The study found that ICG scores predicted poor sleep quality (Beta=0.20, p<0.05). Depression was found to worsen the sleep quality of individuals with CG, whereas PTSD did not. Limitations of the study include: the sample size was too small to generate group differences and the lack of control groups without CG, MDD and PTSD. Barry et al. (2002) [118] evaluated the association between bereaved persons perceptions of the death, such as extent of suffering, and preparedness for the death and psychiatric disorders (n=122), using a baseline interview at 4 months post-loss and a follow-up interview at nine months post-loss (Level IV).

56

Complicated Grief was measured with the Inventory of Complicated Grief-Revised. The persons perception of lack of preparedness for the death was associated with CG at baseline and follow-up (Wald 2 (5) 6.80, OR2.55, 95%CI 1.26-5.15, p<0.01). Because participants in this study had relatively good physical and mental health compared to earlier reports of CG prevalence, this study needs to be replicated to determine if results hold with more physically and mentally impaired bereaved populations. Also, most of the deaths were from natural causes; therefore, it is uncertain whether results can be generalised to those whose relative had a violent death. The study by Goodenough et al. (2004) [119] explored psychological functioning and bereavement outcomes in an Australian sample of fifty mothers and fathers who had experienced the death of their child from cancer in the preceding 1 to 5 years (Level IV). Complicated grief was measured by the Inventory of Complicated Grief. The sample included 30 parents whose child had died at home and 20 parents whose child had died in hospital. Results showed that fathers whose child died in hospital rather than at home exhibited relatively higher levels of depression (F(1,24)=4.49, p=0.046) , anxiety (F(1,24)=8.545, p=0.008 and stress (F(1,24)=5.214, p=0.056) whereas for mothers the place of death was not reflected in psychological outcomes. However, symptoms of CG were positively related to the time that had elapsed between diagnosis and death (r=0.66, p<0.017). Some of the limitations of this study included the self-selecting participation process and a greater heterogeneity of antecedent events for oncology patients dying in a hospital and therefore a greater variance in family anticipatory grieving. Mitchell et al. (2004) [105] undertook a descriptive pilot study of 60 survivors of suicide of a family member or significant other. This study is described in more detail in Chapter 3. Statistically significant differences as measured by the ICG were noted between closely related and distantly related survivors of the suicide victim (F=47.66, p<0.001). Closely related survivors (n = 27) experienced nearly twice the level of CG as distantly related survivors. Relationships classification to the deceased explained 43% of variance in CG scores suggesting that professional assessments and interventions should take into account the familial and/or social relationship of the bereaved to the deceased. In an exploratory study of the effects of CG on sleep, McDermott and colleagues (1997) [120] found that although mild subjective sleep impairment was associated with CG, no effect was detected using the electroencephalographic (EEG) sleep measures. CG was measured with the Inventory of Complicated Grief. Sixty-five bereaved spouses participated in a longitudinal investigation (Level III-3). CG was found to interact with depression to increase the proportion of participants REM sleep (t=3.479, p<0.001). Therefore, CG does not appear to have the same effect on sleep as depression, although only a proxy measure of CG was used rather than precise criteria for syndromal levels of complicated grief.

57

Silverman and colleagues (2000) [121] examined the association between a diagnosis of traumatic grief (TG) and quality of life impairments (n=67) (Level III-2). Traumatic grief was measured by the Traumatic Grief Evaluation of Response to Loss (TRGR2L) Bereaved widowed persons with a positive diagnosis for traumatic grief were found to have greater impairments in quality of life compared to those with a negative diagnosis, controlling for age, gender, time from loss and PTSD diagnosis. For example, TG was found to be significantly associated with lower social functioning scores (Beta=-0.33, p<0.01), lower mental health scores (Beta=-0.38, p<0.0001), and lower energy levels (Beta=-0.16, p<0.01). Limitations of this study include the cross-sectional design and small sample size. Boelen and colleagues (2003) [122]explored the relationship between emotional problems after bereavement for a first degree relative (n=329) and effect of the negative cognitive variables (ie. global negative beliefs, cognitions about self-blame, negative cognitions about other people's responses after the loss, and negative cognitions about one's own grief reactions) (Level IV). Grief was measured using the Inventory of Traumatic Grief. Each of the cognitive variables was significantly related to traumatic grief, depression, and anxiety symptom severity, even after controlling for background and loss-related variables. Forty-nine percent of variance in traumatic grief severity was explained by the cognitive variables global negative beliefs about life, threatening interpretations of grief reactions, negative beliefs about the world, and the future pointing to a need to address negative cognitions when treating traumatic grief. However, there was an under-representation of individuals with little or no grief reactions, other cognitive variables such as positive attitude toward death were not assessed, and symptoms were measured by self-report rather than interview-based assessment. Also, causality can not be drawn between cognitive variables and emotional problems due to the cross-sectional design of the study. Monk et al. (2006) [123] sought to quantify the disruption in peoples daily lives associated with complicated grief by using a diary of daily events. Comparisons were made with a healthy control group matched for age and gender. Complicated grief was measured by the Inventory of Complicated Grief. Sixty-four participants enrolled in an on-going CG treatment study were asked to complete a diary for 14 consecutive days recording information about the day just passed (Level III-3). Thirteen items were designed to give an index score of which events were missed or completed on a given day. The overall MANOVA was significant (F=13,114) =5.13, p<0.001) indicating that activity scores were lower for CG patients than for controls. Patients with complicated grief were significantly more likely than controls (p<0.05) to miss contact with another person (p=0.008), and to miss breakfast (p=0.006), lunch (p=0.0001), dinner (p=0.001), starting work (p=0.005), exercise (0p.002) and going outside (p=0.0001), and more likely to add an afternoon nap (p=0.005) and evening snack (p=0.004). The authors acknowledge that there is more to a persons daily life than 13 simple activities; however, it gave a methodology for quantitative assessment. They concluded that grieving individuals with complicated grief tended to neglect more social events and active events, whereas more passive and/or solitary activities

58

were done more frequently. They conclude that CG may be a unique syndrome associated with measurable lifestyle disruption. Such a pattern of daytime disruption may parallel the previously reported night-time sleep disruptions where excessive napping and night-time sleep disturbance is a clinically significant problem contributing to long-term mental and physical problems. Limitations include the use of a historical control and the lack of a simple grief control group. Complicated grief as a risk factor for adverse health outcomes The following five studies found that complicated grief is a risk factor for a number of adverse health outcomes (Prigerson et al, 1997; Ott, 2003; Latham & Prigerson, 2004; Prigerson et al, 1999; Simon et al, 2005). Prigerson et al. (1997) [124] confirmed that CG, as measured by the Inventory of Complicated Grief, is a risk factor for mental and physical morbidity with a study group consisting of 150 future widows and widowers who were interviewed at the time of their spouses hospital admission and then at 6 weeks and 6, 13 and 25 month follow-ups (Level III-3). Traumatic grief symptoms present 6 months post-loss were found to predict negative health outcomes at the 13- and 25-months follow-up assessments. These health outcomes included: cancer (Mantel-Cox=15.87, df=1, p<0.0001), heart disease (Wald Chi squared=7.38, p<0.01, relative risk=1.15), high blood pressure (Chi squared=3.94, p<0.05, relative risk=1.11), suicidal ideation (t=2.18, p=0.03), and changes in eating habits (Wald Chi Squared=6.69, p<0.01, relative risk=7.02). However, it must be noted that the lack of objectivity in obtaining the measures of physical and mental health, the rarity of outcomes being measured and the lack of a case-control design limited the generalisability of these findings. Ott (2003) [125] examined the impact of complicated grief (CG) on mental and physical health at various points of the spousal bereavement process (Level III-2). Although 112 participants provided data at four points in time, only 29 were identified with CG and this was measured with the Inventory of Complicated Grief. Mental health scores were found to be significantly lower for the CG group 6 months post-loss. This trend continued for the remainder of the study with assessment being made at 9-, 12-, 15-, and 18months post-loss. Compared with the group with non complicated grief, the CG group were found to have experienced more additional life stressors (t(110)=2.24, p=0.027), perceived less social support (t(110)=5.09, p=<0.001), and achieved less clinically significant changes in the mental health scores (Chi squared=23.81, df=2, p<0.001). The results indicated that those in the complicated grief group experienced a significant decrease in mental health, a decreased sense of well-being, decreasing functioning in life roles, and an increase in problematic symptoms compared with the group who did not meet criteria for complicated grief. However, the study may not represent a 'typical' grieving person due to the voluntary nature of the recruitment process. Additional limits of the study include: an overrepresentation of women, the cohort sequential design meant that data were not collected from one

59

timeframe, participants were not asked when health problems developed, and studies were not able to link CG to the development of new health problems. Latham and Prigerson (2004) [126] examined the influence of CG on suicidal thoughts or behaviours in 281 bereaved elders (74% female, median age 64 years) at an average of 6.2 months post-death at baseline and 10.8 months at follow-up (Level III-3). This study was also cited in Chapter 3. Cross-sectionally CG was associated with a 6.58 times greater likelihood of high suicidality at baseline, and an 11.3 times greater risk of high suicidality at follow-up after controlling for gender, race, major depressive disorder, PTSD and social support. Longitudinally, CG at baseline was associated with an 8.21 times greater likelihood of high suicidality at follow-up, controlling for the above confounders. They conclude that CG poses an independent psychiatric risk for suicidal ideation. Prigerson and colleagues (1999) [97] examined the influence of traumatic grief (also referred to as CG) on suicidal ideation in 76 young adults who had experienced the suicide of a friend on average 6 years previously (Level III-3). This study is reviewed in Chapter 3. Twenty participants (15%) with symptomatic levels of CG were found to be five times more likely to report suicidal ideation than participants with non-symptomatic levels. Levels remained high after controlling for depression, gender and time since death. A recent study by Simon et al. (2005) [127] investigated the frequency and implications of the death of a loved one and complicated grief on 103 patients with bipolar disorder. Participants were in a Systematic Treatment Enhancement Program for Bipolar Disorder, a large naturalistic study in order to identify frequency of loss and to examine the presence of CG and its clinical correlates. In the CG sample, more patients reported a lifetime history of a suicide attempt (58% vs 34%) (FET p = 0.054). This association of CG with suicide attempts did not diminish after controlling for lifetime panic disorder, with a more than doubling of the odds of a lifetime suicide attempt (OR=2.5, Z= 1.92, p = <0.06). CG patients had significantly higher PAS-SR total scores, as well as higher scores on specific subscales representing paniclike symptoms, substance and medication sensitivity, anxious expectation, agoraphobia, illness-related phobias and hypochondriasis, and reassurance sensitivity. CG continue to predict elevated total phobic avoidance in a regression model controlling for any current anxiety disorder and current mood state (B=10.2, t = 2.4, p = <0.02). Limitations of the study include reliance on self-selection for participation, self-report assessments for deaths without clinical validation and the cross-sectional design does not allow the direction of effect to be established.

60

Summary The outcomes of bereavement that led to CG and adverse health outcomes include lack of preparedness for the death [118]; the place of death (hospital versus home) and the time that elapsed from diagnosis to death [119]; the closeness of the relationship to the deceased [105]; cognitive variables such as global negative beliefs about life, threatening interpretations of grief reactions, negative beliefs about the world and the future [83]; sleep disturbances [117] [120]; and changes in daily routine such as missing meals., exercise, increased sleep, and lack of energy [123]. Although bereavement itself has been shown to pose an elevated risk for a variety of negative physical, mental and social outcomes, and death; some studies have found that complicated grief among the bereaved is associated with heightened risk of physical and mental impairments. These include cancer, heart disease, high blood pressure, suicidal ideation, and changes in eating habits [124]. Other studies found a significant decrease in mental health, a decreased sense of well-being, decreased functioning in life roles, more perceived additional life stressors, perceived less social support [125]. Increased panic attacks, alcohol abuse co-morbidity, higher rates of suicide attempts, greater functional impairment and poorer social support were found in a population of patients with bipolar disorder [127]. The studies that examined the role of sleep disturbances and complicated grief produced inconclusive results. An exploratory study by McDermott and colleagues (1997) found mild subjective sleep impairment but no effect was detected on EEG sleep measures [120]. Similarly, Germain and colleagues [117] found an association between complicated grief and poor sleep quality, with depression found to worsen sleep quality of individuals with CG whereas PTSD did not. Although these findings offer some insight regarding outcomes associated with CG to further our knowledge, limitations include small sample sizes, the use of self-report subjective measures, the crosssectional design of studies preventing conclusions about predicting causality, lack of case-controls and inconsistencies in the measures of CG used. Recommendations: Outcomes from Bereavement

Research is needed to assess the effect of CG on outcome measures using large, nonclinical samples, prospective controlled designs.

Additional research is needed to clarify the nature of the association between complicated grief and adverse health outcomes and to identify the specific psychological and biological pathways through which CG is expressed in poor health.

61

CHAPTER 8: COMPLICATED GRIEF IN SPECIFIC POPULATIONS The literature review identified studies on complicated grief within the following bereaved populations: children and adolescents, parents, spouses, family members who were either caring for loved ones receiving palliative care or who had received palliative care and died, adults whose partner died from HIV/AIDS, euthanasia, older adults, people with psychiatric illness and minority groups including different cultural groups and Indigenous populations. These are summarised below. Bereaved children and adolescents Over the past 25 years the psychological sequelae of grief in bereaved children have been described. Symptoms examined include depression, anxiety, posttraumatic stress disorder (PTSD), behavioural problems, suicidal ideation, and reduced psychological function. The work has not been characterised by consistent definitions of the term, complicated grief and the measures selected to assess the effect of loss on children have also varied making it difficult to draw firm conclusions. Four empirical studies are summarised below that appear to have relevance to the aims of this systematic review. Saltzman and colleagues [106] evaluated the effectiveness of a school-based screening and group treatment protocol, trauma- and grief-focused group psychotherapy, for adolescents (n=26) exposed to community violence and trauma either due to losing someone to a traumatic death or witnessing a traumatic act (Level IV). This study was also summarised in Chapter 5. Results of the study, which used a pre-test and post-test design, suggested that group participation was associated with improvements in posttraumatic stress, CG symptoms and academic performance. Finlay and Jones [128] tested a health promotion grief awareness programme for young offenders with CG (Level IV). This study was reported as a Brief report and lacked sufficient data and study information to draw sound conclusions. Seventeen male offenders aged 17 to 21 years were included in the study; of these eight offenders attended the programme, seven declined the programme and two were excluded. Participants were initially screened to identify those suffering CG, and then interviewed using a structured interview. Young offenders who reported coping poorly with bereavement were more likely to have been bereaved in late adolescence, to have lost a first degree relative, used drugs, had suicidal thoughts, and had anxiety and depression (no statistical data was reported in the paper). These findings are of limited value because of poor study design including use of a small sample, lack of a control group, lack of information regrading the screening tool and structured interview guide, lack information describing the programme, and finally a lack of any reported data. A large amount of literature related to traumatic loss was identified that specifically related to children and adolescents. This literature focused mainly on posttraumatic stress disorder, as a result of natural (e.g. earthquakes and tsunamis) and induced disasters (e.g. war and terrorist activities). A number of authors

62

(such as Pynoos, and Cohen) have undertaken extensive work in the area. Although these studies appear to examine complicated grief, they focus on the effects from exposure to trauma, rather than the trauma caused by grief as in CG, and use of the term in these studies is not concordant with the definition and criteria specified in this review. Given the scope of this review, examination of these papers was considered to be outside the scope of this review. However, two papers were considered to make a helpful contribution to the newly emerging condition of Childhood Traumatic Grief (CTG) in the wake of current community traumas, and highlight the urgent need for implementing and evaluating the efficacy of interventions to address the need for effective treatments specifically for CTG. Cohen and colleagues (2004) [129] examined the efficacy and specific timing of treatment response of individual child and parent trauma-focused cognitive-behavioural therapy for childhood traumatic grief (CTG) (Level III-3). CTG is the term given to the condition in which children have a loved one die under traumatic circumstances and develop trauma symptoms that may impinge upon the childs ability to successfully navigate the grieving process. This pilot study examined the clinical response to parallel individual child and parent focused cognitive behavioural therapy for CTG (CBTCTG) using a16-session treatment model. The authors hypothesised that the treatment would result in significant decreases in CTG, PTSD, depressive, anxiety and behavioural symptoms. Multiple assessment times were used to measure pre and post treatment changes and whether changes would correspond to planned changes in treatment instruments. Sessions 1 to 4 used interventions to improve affective modulation and stress reduction, sessions 5 to 8 interventions focused on naming and accepting what the child had lost, sessions 9 to 12 focused on preserving memories, and sessions 13 to 16 focused on making meaning of the loss. Twenty-two participants between 6 and17 years of age and had a loved one die from a variety of traumatic events: accident, medical cause, suicide, homicide, and drug overdose. Instruments included the Expanded Grief Inventory (EGI), Childrens PTSD Symptom Scale (CPSS), the Mood and Feelings Questionnaire (MFQ), and the Screen for Childrens Anxiety Related Emotional Disorders (SCARED). The following instruments were completed by parents: UCLA PTSD for DSM-IV Parent Report Version, Child Behaviour Checklist, the PTSD Diagnostic Scale, and BDI. Children experienced significant improvements in CTG (p<0.001), PTSD (p<0.001), depressive (p<0.01), anxiety (p<0.000), and behavioural problems (p<0.01), with PTSD improving only during the trauma-focused treatment components, and CTG improving during the grief focused treatments. Parents also experienced significant improvement in PTSD (p<0.000) and depressive symptoms (p<0.000). This was the first study to evaluate the potential efficacy of a parallel child and parent trauma and grief focused treatment protocol for resolving CTG, PTSD, depression, and other symptoms in traumatically bereaved children. Cohen et al. (2004) [130] described two psychotherapies for complicated traumatic grief (CTG) that are being used in a randomised study of two psychotherapies after the 9/11 disaster. This disaster is an example of dual tragedy that creates chaos around anger and sadness. Findings from this study are still in

63

their infancy, thus, the paper does not meet the criteria for inclusion in this study. The authors identified that approximately 11,000 children lost parents or other loved ones as a result of the 11 September 2002 disaster. At the time there was an absence of any practical precedent to use in the development of treatment planning interventions for CTG. A randomised comparison study of the following two psychotherapies for CTG has been commenced: Trauma Focused CTG (TGCBT) and Client Centred Therapy (CCT). Both interventions offer theoretically grounded approaches to treatment of CTG, and have been extensively studied previously, thus creating a therapeutic relationship critical to success. The authors underscore the need for collaboration among multiple levels within the health/community systems and individuals with diverse expertise, resources, support and self care by families. The authors described the following theoretical reasons for choosing TG-CBT and CCT. For TGCBT, the experience of a traumatic event and presence of some symptoms of PTSD required selection of a treatment specifically targeting PTSD that was theoretically sound. At the time of this study, CBT had the strongest evidence of efficacy and had been modified prior to 9/11 to include grief focused interventions (see previously described study by this author). CCT represented a common type of treatment provided by community therapist treating grieving children that comprised supportive, child centred model for treating grief that approximated community treatment-as-usual. CCT focused on empowerment and reestablishment of trust aimed to correct powerlessness, the sense of betrayal and helplessness experienced by bereaved and traumatised children, as well as bereaved parents. No firm recommendations can be made based upon this work in progress. As well, the findings will be specific to children who have experienced a traumatic event. Therefore, further testing would be needed to determine the extent to which the interventions used above would apply to a child diagnosed with CG as specified in this systematic review. Summary No intervention studies have been undertaken with infants, children or adolescents focused on treatment of complicated grief. One study provided information regarding the effect of an intervention in the context of traumatic grief; however, methodological limitations limit conclusions that can be drawn and the study is not specific to complicated grief. Much of the literature appears to be directed to developing evaluating interventions that can be used with children who have suffered a traumatic loss or experience. This work has led to the first randomised controlled trial in the area of traumatic childhood grief, the results of which may have implications for this situation. Further testing would be needed to assess the usefulness of this intervention in the treatment of complicated grief in children and adolescents.

64

Bereaved parents The literature search resulted in seven studies that focused on bereaved parents. The studies evaluated perinatal loss, bereavement outcomes, patterns of complicated grief, hospital-based and bereavement support for parents. Perinatal death is defined as the death of a foetus after 20 weeks gestation, through to the death of an infant up to the age of one month post partum. Perinatal bereavement is uniquely devastating for parents who expect to give birth to a healthy infant. It is an unanticipated and unnatural event in the course of human life. Perinatal death has long been overlooked by health professionals and researchers. As discussed earlier in this review, CG has been defined in a number of ways and has been labelled with various terms. There is a lack of clarity surrounding CG as it relates to perinatal loss. Janssen, Cuisinier and Hoogduin (1996) [131]undertook a critical review of empirical studies related to pathological grief following pregnancy loss according to four subtypes derived from general bereavement literature: chronic grief, delayed grief, masked grief, and exaggerated grief. These authors concluded that in the first six months following a pregnancy loss, common complaints reported by mothers can include psychological complaints, behavioural changes, and somatic complaints. Less common problems include psychiatric disorders during the first two years post loss in 10-15% of mothers, with less than 10% of these women seeking appropriate psychiatric care. Both parents often mourn the loss of their baby for more than one ear, with one in five women unable to accept pregnancy loss after two years. It was also found that delayed grief reactions occurred in 4% of parents, occurring most commonly in fathers. The authors suggested that CG for bereaved parents may be less common than once thought. Of note, they further proposed that the long held belief that parents are at higher risk for complicated grief following pregnancy loss, probably results from flawed empirical studies. Perinatal death has been distinguished from other forms of deaths; therefore, psychological trauma and grief outcomes following perinatal loss must be viewed according to individual trauma-related vulnerability, resiliency, and context. Further, it is generally accepted that the normal grief reactions to perinatal death do not differ greatly from those observed in other bereavement situations, and that it is not usually associated with complicated grief. No studies were found that related to perinatal death and CG as per the inclusion criteria for this review. However, a study by Murray evaluated the effects of an intervention program on parental distress following infant death [132, 133]. One hundred and forty-four parents were divided into an experimental group (n=84) and a control group (n=60) and were assessed in terms of their psychiatric disturbance, depression, anxiety, physical symptoms, dyadic adjustment and coping strategies. The program consisted of counselling from a trained grief worker and provision of resource materials appropriate to parents

65

needs. Findings indicated that the intervention was effective in reducing the distress of parents particularly those assessed to be at high-risk of developing mourning problems. Compared to the control group, the experimental group had a significant change in total psychiatric disturbance (F(2,44)=52.18, p<0.001), coping measures (F(3,73)=13.88,p<0.001) and adjustment (F(4,63)=6.30, p<0.001). Some of the limitations of the study relate to the recruitment of the two groups at different time periods raising issues of comparability. Also, the influence of hospital and emergency staff at the time of death on parental reactions to the death cannot be under-estimated. Hospital-based bereavement support programs in the paediatric setting are a relatively recent development and reflect both an acknowledgment by health professionals of the complexity of parental grief, as well as the role they have in the provision of support for bereaved families. Despite these advances, little empirical work has been undertaken to determine the effect of interventions on complicated grief in this high risk population. A survey of the ten major tertiary paediatric oncology units in Australia and New Zealand was conducted by deCinque and colleagues [134] to determine the current practices that existed in relation to hospital bereavement-based support programmes (Level IV). A 19-item questionnaire was developed for the purpose of the study. Nine questionnaires were returned. Findings showed the majority of hospitals (n=8, 80%) provided a multidisciplinary bereavement service for approximately one year after the death of a child. The most common programmes provided were counselling (n=7, 78%) and support groups (n=6, 67%). However, no formal evaluation of programmes had been undertaken. Approximately half the hospitals were found to be working from a limited theoretical basis (n=5, 55%) and no hospitals screened parents to determine those that may be at risk of CG. A limitation of the study was the non-inclusion of questionnaire items related to staff education about bereavement support and loss and grief. Findings from this study support the need for preliminary intervention studies with parents at high risk of complicated grief. Ginzburg, Geron and Solomon [104] assessed patterns of grief reaction and adaptiveness in bereaved parents whose adult child had died during military service. This study is reviewed in full in Chapter 3. The study results showed that absent and delayed grief reactions were the most prevalent variants and were associated with lower levels of psychosocial adjustment compared to delayed grief reactions. Circumstances associated with the loss, level of education, and religious attitudes were found to be associated with the type of grief reaction. Findings from this study should also be viewed cautiously because all participants were already participating in a bereavement support group and data were collected during this intervention. In addition, the sample and social context of the study population were unique. The long-term bereavement and psychological outcomes of parents who have lost a child to cancer was addressed in a study by Goodenough and colleagues [119]. This study is discussed in more detail under risk factors in Chapter 4. Results showed that fathers whose child died in hospital rather than at home

66

exhibited relatively higher levels of depression (F(1,24)=4.49, p=0.046) , anxiety (F(1,24)=8.545, p=0.008 and stress (F(1,24)=5.214, p=0.056) whereas for mothers the place of death was not reflected in psychological outcomes. However, symptoms of CG (as measured by the ICG) were positively related to the time that had elapsed between diagnosis and death (r=0.66, p<0.017). A study by Kempson [135] examined the effect of touch therapy on grieving mothers. This study is discussed in detail in Chapter 6 which reviews interventions. Touch therapy was found to significantly improve despair (F=8.290, p=0.005), depersonalization (F=4.904, p=0.031), and somatisation (F=6.833, p=0.012). However, the actual effect of the intervention on CG was not measured. Dyregrov et al. [99] compared the outcome and predictors of psychological distress of parents in Norway bereaved by youth suicide, sudden infant death syndrome and child accidents. Self-isolation was found to be the best predictor of psychosocial distress and being female predicted complicated grief in the suicide and SIDS samples. There was no evidence of suicide survivors having greater difficulties in adapting to the death compared with survivors of SIDS or accidents. This study is discussed in more detail in Chapter 3, which reviews traumatic death. Summary No studies examined CG in the context of perinatal death. One study tested a counselling and educational support approach with parents who had experienced an infant death. Although positive outcomes were reported for the experimental group, no diagnosis of CG was used, limiting conclusions about the extent to which this study has relevance in the context of CG. A review of bereavement services provided in hospitals provided no evaluation data and limited descriptions of current practices. One of the most useful studies reviewed in this section was Goodenough and colleagues (121) study of the long-term bereavement and psychological outcomes of parents who have lost a child to cancer. Place of death and time since diagnoses were found to be important variables in predicting CG. Spouses Families are the primary source of long-term care for the sick and elderly, with as much as 60 to 80 percent of home care for the elderly reported to be provided by spouses. The death of a spouse can be painful and debilitating, and is considered one of the most stressful events a person may endure. To date, research has centred on adaptation to conjugal loss and has revealed several basic patterns of outcome including: depression, chronic grief, delayed grief responses and the absence of grief symptoms. Fourteen studies were identified within this category that met the systematic review criteria. Most of this work has been conducted after bereavement, thus knowledge of divergent trajectories of grieving and antecedent predictors is lacking. Three studies examined patterns of bereavement following conjugal loss and associated predictors. These studies are summarised below.

67

Bonanno and colleagues [112] examined the bereavement patterns in relation to pre-loss predictor variables. The study gathered prospective self-report data on 205 individuals several years prior to the death of their spouse and then at 6 and 18 months post-loss. This study is examined in more detail under risk factors in Chapter 4. A study by Boerner and colleagues [39] examined patterns of distress or grief trajectory following conjugal loss of 92 older adults who initially showed high or low distress following conjugal loss (Level III-2). The chronically depressed group showed significantly higher scores in grief and depression over time supporting the hypothesis that signs of improvement would be more evident in the chronic grief compared with the chronically depressed group. The chronic grief group showed significant decrease over time in grief. Although the chronic grief group experienced a more intense and prolonged period of distress compared with other groups (e.g., common grief group), improvements by 48 months suggest that this group does not remain chronically distressed as a result of the loss. In contrast, the chronically depressed group clearly demonstrated long-term problems, with little indication of improvement between 18 months and 48 months. This study is examined in more detail under risk factors in Chapter 4. Bierhals and colleagues (1996) [136] explored whether the symptoms of CG could be mapped onto a staged theory of grief, and the extent to which widowed spouses through the same stages of grief, and whether the progression was similar for both genders (Level III-2). Of the 97 participants, 26 were widowers and 71 were widows. CG was measured using the ICG, grief via the TRIG, and depression via the BDI. Widowers bereaved for three or more years were found to have increased bitterness (t=-2.97; DF=6; p<0.05), compared with widows who were found to have lower levels of CG (t=3.03; DF=16; p<0.01). These findings indicated that symptoms of CG remained stable for the first three years of bereavement for both genders. The study was limited by use of a cross sectional design. Prigerson and colleagues (1995) [24] investigated if CG could be distinguished from bereavement-related depression in 82 widowed elderly subjects participating in a study of changes in sleep physiology (Level III-3). Seven symptoms constituted complicated grief: searching, yearning, preoccupation with thoughts of the deceased, crying, disbelief regarding the death, feeling stunned by the death, and lack of acceptance of the death. The first component accounted for 26% of the variance and the second component for 20% of the variance. The agreement between significantly impairing complicated grief and syndromal-level depression was moderate suggesting that symptoms of complicated grief may be distinct from depressive symptoms. In 1995 Bonanno and colleagues explored whether avoiding painful emotions during bereavement leads to either prolonged grief, delayed grief, or delayed somatic symptoms [137] (Level III-3). Emotional avoidance was measured in 42 bereaved participants 9 (aged between 21 and 55 years) 6 months after a

68

conjugal loss. Negative dissociation at 6 months was found to be associated with minimal grief symptoms at 14 months (F(1,36)=21.13, p<0.001). Negative dissociation scores were also linked to high levels of somatic symptoms, that fell to a low level by 14 months (F(1,36)=14.61, p<0.001). The findings, therefore, were consistent with the competing hypothesis that emotional avoidance during bereavement may serve adaptive functions. In 1999 Bonanno and colleagues examined emotional avoidance (via verbal autonomic response dissociation) and its effect on grief symptoms [138]. Participants comprised 42 conjugally bereaved (3-6 months prior to study commencement) spouses who had participated in the above described study [137]. A battery of questionnaires was mailed to participants between 3-6 months post-loss. A structured grief symptoms interview was then carried out 6 months post-loss, followed by a semi-structured narrative interview 17 days later. A post-loss grief symptom interview was conducted at 14 and 25 months respectively (Level III-2). Findings from this study confirmed those of the previous study and showed that verbal-autonomic dissociation was linked to the mildest grief course with no evidence of delayed grief. This predictive relationship remained significant even when initial levels of grief were controlled. No evidence was found for enduring or delayed health difficulties in association with verbal-autonomic dissociation. The main limitation of this study was that data related to a specific type of loss, the death of a spouse at a specific point in the life span, midlife. The sample was relatively homogenous with regard to culture and ethnicity. Although gender differences in emotional disassociation were not evidenced in the present study, gender difference may emerge in a larger sample size. In addition, the study did not use an objective measure of health outcome and it is unclear whether alternative findings may have emerged if a more objective measure of health was available at the time of the study. Bonanno and Field (2001) examined different theoretical positions regarding grief responses using prospective data from a sample of 39 midlife, conjugally bereaved adults in their fifth year of bereavement [139] (Level III-3). No cases of delayed symptom elevations were observed and data on the emotional processing of the loss at six months failed to support the traditional assumption that minimal emotional processing of the loss would lead to delayed grief. The small sample size is an obvious limitation of this study. However, the capacity to track participants longitudinally over five years is a strength, and this work represents one of the few studies in this area that has been able to examine changes in grief responses over time, challenging some longstanding theoretical postulates. Clearly this work needs to be replicated with larger samples to confirm these results and provide additional data regarding the process of grief responses and the stability of grief responses. Maercker and colleagues (1998) [140] used narrative research to explore the thematic parameters of CG among bereaved spouses. Forty-four participants (27 widows and 17 men) who had been bereaved between 3 and 6 months prior to the study were recruited. The prevalence of, and interrelationship

69

between, positive and negative themes were investigated using narrative interviews 6 months post bereavement. A number of assessments were also made at the time of interview including: The Diagnostic Interview Model for Grief, TRIG, impact of Event Scale (IES), BDI and BAI. The relationship of the identified themes to various symptom measures was then examined at 6 and 14 months post bereavement. Eight positive and 8 negative themes were identified. Findings revealed no systematic relationship between corresponding negative and positive themes. Only small inter-correlations among and within the positive and negative themes were found. The strongest correlations found (r=.36 to r = .56, p = <0.01) were between adjacent themes (e.g. generativity and integrity, stagnation and despair/regret, autonomy and initiative). Aggregated positive and negative themes showed a significant relationship with measures of intrusion (r=-0.33, p<0.05) and avoidance (r=-0.36, p<0.05) taken at 6 months and measures of grief-specific, anxiety, and depression taken at 14 months. Levels of grief at 14 months post loss were predicted by 6 month grief symptom ratings (predicting 56% of total variance) and overall frequency of positive themes. This study is limited in its generalisability due to the sample size and homogenous sample given the potential effect of cross cultural factors. Field and Horowitz (1998) used the Gestalt empty-chair technique to assess unresolved grief and its relation to later adjustment [141] (Level III-3). Bereaved individuals who experienced the death of a spouse on average six months previously participated in an empty-chair monologue task in which they were instructed to speak to their deceased spouse, imagining that they had one last opportunity to do so. Participants completed the Beck Depression Index (BDI) and Impact of Event Scale (IES) at time of the intervention and at 14 months post-loss. As hypothesised, the extent of unresolved grief as assessed by the monologue questionnaire at six moths post-loss was predictive of 14-month post-loss symptoms. A study by Prigerson and colleagues [124] examined the extent to which symptoms of CG are predictors of future physical and mental health outcomes and confirmed that CG is a risk factor for mental and physical morbidity with a study group consisting of 150 future widows and widowers. Traumatic grief symptoms present 6 months post-loss were found to predict negative health outcomes at the 13- and 25months follow-up assessments. These health outcomes included: cancer, heart disease, high blood pressure, suicidal ideation, and changes in eating habits. However, it must be noted that the lack of objectivity in obtaining the measures of physical and mental health, the rarity of outcomes being measured and the lack of a case-control design limited the generalisability of these findings. This study is reviewed in more details under adverse health outcomes in Chapter 4. A prospective longitudinal study by Beery and colleagues [142] examined the possible factors associated with depression and CG among caregivers of terminally ill spouses (Level III-3). This study is discussed in more detail in the next section of this chapter on palliative care.

70

A cross-sectional, descriptive, pilot study by Brintzen-hofeSzoc et al. [143] investigated the relationship between family functioning, psychological distress and grief reaction in surviving spouses whose spouse died of cancer (Level IV). A convenience sample of 37 (52% response rate) completed measures including the Texas Revised Inventory of Grief (TRIG), the Family Adaptability and Cohesion Scale and Brief Symptom Inventory (BSI). Significant relationships were found between the level of family functioning, psychological distress and grief reactions. Specifically, the more enmeshed the family, the more complicated the current grief reaction; the higher the anxiety the more likely an enmeshed level of family functioning, and the more depressed, distressed, and anxious the surviving spouse the more likely the grief was complicated. Limitations include the retrospective reporting of family function, a small sample size, the time since death ranged from 5 month to 19 months and the use of the TRIG to measure complicated grief. Johnson et al. (2006) [144] developed and validated an instrument for assessing dependency on the deceased. The Bereavement Dependency Scale (BDS) was tested within the Yale Bereavement Study among 170 widowed participants. Other measures used in this study were the Dyadic Adjustment Scale (DAS), the Structured Clinical Interview for DSM-IV Axis 1 disorders, the Inventory of Complicated Grief (ICG), the Interpersonal Support Evaluation List (ISEL), the Yale Evaluation of Suicidality Scale (YES) and the Relationships Styles Questionnaire (RSQ). Data analyses were conducted to investigate whether the BDS demonstrated satisfactory reliability and convergent, discriminant and criterion-based validity. Respondents with complicated grief, general anxiety disorder and major depressive disorder, and respondents who reported significant suicidal ideation (i.e., a YES score >3) had significantly elevated Bereavement Dependency Scale scores relative to those without these conditions. These associations were statistically significant after controlling for age, gender and level of education. A significant correlation was obtained between continuous BDS and YES scores, after the covariates were controlled statistically (partial correlation=0.30; p<.0001). BDS scores were positively correlated with CG symptoms among both the men and women in the sample. Summary Most studies on spouses has been conducted after bereavement, thus knowledge of divergent trajectories of grieving and antecedent predictors is lacking. In summary, a number of studies examine issues associated with grief in older adults focusing on widows/widowers and survivors of individuals who died from progressive illnesses (eg. cancer, heart disease, etc). Most of the research has been descriptive and correlational with few methodologically sound intervention studies targeted toward older adults with complicated grief. Outcome studies have produced mixed, but somewhat positive results. For example, interventions that provide support and encourage expression of emotions in the months after a traumatic or sudden death appear to be associated with reduced symptoms [145, 146]. Similarly, high risk bereaved

71

individuals who received support in death with their grief immediately after the death appear to benefit, as indicated by lower levels of physical symptoms, depression, worrying, and use of health care services [147]. Palliative Care Bereavement follow-up is considered to be an essential component of any palliative care program [148]. Palliative care services offer a range of bereavement follow-up services ranging from mailing a sympathy card to survivors, to intense one-to-one grief counselling. However, little research has been undertaken to identify the factors that might predict those family members who may be at risk for a more complicated grief reaction [78]. To date, most efforts to provide bereavement services have been undertaken using trial-and-error approaches, with little evidence to guide service delivery. Bouton [149] has called for the development of bereavement programs that are based upon a continuum ranging from very little need to extreme need, to allow service providers to tailor services, make optimal use of limited resources and ensure that they are not over-treating or under-treating the bereaved populations that they serve. According to Parkes [150] the most common form of bereavement intervention relies on the bereaved individual to contact the service. This form of intervention relies on the bereaved to make a rational, objective decision which may be difficult at the time when they are most in need. Consequently, individuals who may not be coping well may delay finding help or may become more distressed, which can result in an under-treatment of complicated grief responses [78]. Recent reviews of bereavement interventions indicate that grief counselling may not be helpful for many people experiencing normal grief and may even have negative effects (eg. Jordan et al and Schut et al. [151, 152]). Reviews indicate that interventions may be more helpful for individuals experiencing complicated grief. Therefore, use of a proactive screening tool to identify individuals who may be at risk for a complicated grief response in the context of palliative care may offer preventive health benefits [153]. Nine studies involving palliative care populations were identified that met the criteria for inclusion in this systematic review. A recent prospective, descriptive study was undertaken by Kristjanson and colleagues [78] to test the reliability and validity of the Bereavement Risk Index (BRI) [154] in assessing grief reactions of bereft family members in a home hospice care setting. The study also endeavoured to identify the types of family members most likely to experience a more difficult grief reaction. One hundred and fifty bereaved family members participated. Participants received one of three types of bereavement support: follow-up based on BRI assessment from nurses who had received a bereavement education program, follow-up from nurses who had received a bereavement education program only, or standard care. Families in the

72

first group were classified as high, medium or low risk and received a structured bereavement follow-up protocol based on their level of risk. Results indicated that a shorter 4-item version of the BRI was more internally consistent than the longer 8-item version and demonstrated good predictive validity when correlated with outcome measures at three and six months following the patients death. All bereaved family members in the study reported poorer health scores as measured by the SF36 at three and six months following the patients death, compared with normative data for the same aged groups. This study identified a very small number of individuals who were classified in the high risk category (approximately 7%). Spouses/partners experienced greater bereavement distress, as did younger (< 45 year-old) family members. Other researchers have reported that the quality of the relationship is predictive of bereavement outcomes [28, 155]. One descriptive study by Beery and colleagues [142] examined the effects of changes in role function, care giving tasks, caregiver burden and gratification on symptoms of depression and traumatic grief (Level III3). Data were derived from rater-administered and self-report questionnaires completed by 70 spouses of terminally ill patients. Caregiver burden was significantly associated with the spouses level of depression and traumatic grief. Results also indicated that changes in role function, specifically changes in restriction of sport and recreational activities were associated with caregivers level of depression. The number of instrumental activities of daily living tasks (IADL) performed for the terminally ill spouse was negatively associated with the caregivers level of depression. This was an unexpected finding. Fewer IADLs may have been found to be associated with higher levels of depressive symptoms because care giving activities may become routine, providing a sense of structure and purpose for the caregiver. Caregivers with fewer of these tasks to perform may therefore experience more depression and less sense of purpose. In this sense, IADLs may work as a buffer against depression in bereaved widows and widowers [156]. The researchers [142] found no significant association between activities of daily living tasks performed for the ill spouse, caregiver gratification, duration of care giving, and time spent care giving each week with mental health outcomes (ie, traumatic grief response or depression ratings). These results may be limited by the fact that data were obtained from a sample of spouses who provided care to individuals suffering from a wide range of terminal illnesses. The relatively small sample size and descriptive design also limit conclusions. Barry et al. evaluated the association between bereaved persons perceptions of the death (e.g. extent of suffering, violent versus peaceful death) and preparedness for the death and psychiatric disorders [118] (Level IV). Barry et al. [118] have shown that the perception of death as more violent was associated with Major Depressive Disorder at 4 months post-loss. More importantly, this work indicated that the

73

perception of lack of preparedness for the death was associated with complicated grief at 4 and 9 months post loss. Kissane and colleagues (1996a) undertook a study to identify patterns of family functioning in adult families after the death of a parent due to cancer [157] (Level III-3). One hundred and fifteen families completed measures of family functioning, grief (Bereavement Phenomenology Questionnaire), psychological state and social adjustment at 6 weeks, 6 months, and 13 months post-death. Cluster analysis methods were used to develop a typology of perceptions of family functioning during bereavement. Five types of families emerged from dimensions of cohesiveness, conflict and expressiveness on the Family Environment Scale. Thirty-six percent of families were considered supportive and another 23% resolved conflict effectively. Two types were dysfunctional: hostile families and sullen families. In a companion article Kissane and colleagues [158] reported that sullen families displayed the most intense grief and the most severe psychological morbidity (Level III-3). Wellfunctioning families (supportive and conflict-resolving) resolved their grief and adjusted more adaptively than their dysfunctional counterparts (intermediate, sullen and hostile). Although Kissane did not set out to specifically measure complicated grief, this work offers some interesting new ways of classifying families and treating the health of the family as a unit as a potential predictor of grief reactions. In 1998 Kissane and colleagues reported on the development of family grief therapy based upon their prior work and ability to classify families and identify those most vulnerable for more difficult grief responses. Again, the term complicated grief was not used the focus of this work [159] The researchers recommended use of the Family Relationship Index (Moos & Moos, 1981) as a quick and simple way of screening families that might benefit from family grief therapy. Case studies of various family typologies are provided as an illustration of the therapy. Specific outcomes measures are not provided to evaluate the intervention; however, the work raises helpful questions about how to best structure a family oriented intervention. In 2003 Kissane and colleagues used the Family Relationships Index (FRI) to screen families and the Family Assessment Devise (AD) as an independent family outcome measure [160]. The Brief Symptom Inventory (BDI), and Social Adjustment Scale (SAS) were used as psychosocial measures. No measure of complicated grief or grief response was included. Screening of 257 families revealed 74 well-functioning families and 183 at some risk of morbid outcome. Of the latter, 81 agreed to participate in a randomised controlled trial of family focused grief therapy. Results confirmed the predictive validity of the FRI as a measure to screen families for psychosocial morbidity. Follow up results of the effectiveness of the intervention have not yet been reported.

74

Recent work by Bradley and colleagues [161] suggests that earlier hospice enrolment may reduce the risk for Major Depressive Disorder during the first six-eight months of bereavement. This is consistent with Barrys work because earlier hospice enrolment might indicate more preparedness for the death. These data strongly suggest that advanced preparation for the loss may help to reduce the risk of developing complicated grief and make the grieving process less painful for the survivors. Christakis and Iwashyna [162] investigated whether hospice use by patient was associated with decreased risk of death in surviving bereaved spouses using a matched, retrospective cohort design. A populationbased sample of 195,533 elderly couples in the United States was obtained: 30,983 couples that used hospice and a matched cohort of 30,838 that did not. The mean length of hospice care was 22 days for male patients and 25 for women, which is consistent with national norms. After adjustment for other measured variables, 5.4% of bereaved wives died by 18 months after the death of their husband when their deceased husband had not received hospice care, compared with 4.9% that died when their deceased husband had received hospice care (adjusted odds ratio of 0.92 in favour of hospice use). Similar results were reported for bereaved husbands, indicating that palliative care may impart some type of protection for the surviving spouse in both relative and absolute terms. Summary Despite the view that palliative care services should include bereavement services to address the needs of families, little empirical work to guide the identification of those at risk for complicated grief reactions in this care context has been undertaken. Some theoretical and descriptive work has been undertaken to identify factors that might predict those family members who may be more at risk of developing complex grief reactions, such as functioning, quality of the patients death, family care satisfaction during the palliative phase of illness and pre-morbid state of family members health [163-165]. Preliminary descriptive research has been undertaken to identify outcome measures that might be useful in assessing the effectiveness of bereavement interventions, such as family members health, family functioning, psychological health indicators [150, 164-166]. With the exception of the recent work by Kristjanson and colleagues [78], little systematic empirical work had been undertaken to test a clinical bereavement assessment tool for use in a palliative care setting. As well, there is a paucity of research to guide development of bereavement interventions and services directed toward this population. The literature that does exist suggests that there may be a small group of individuals at risk for a complicated grief response. The apparent protective effect of palliative care services for widows and widowers whose spouse received this type of care has been reported in a large, well designed study, suggesting that preparation for death and practical help with end of life care may be preventive [151].

75

HIV/AIDS AIDS is an illness that develops over a long period of time, becoming more debilitating and stressful for both the patient and supporting family and friends. With progressive deterioration come increasing demands for emotional and practical support, loss of social and sexual relations, financial pressures and challenges of end of life symptoms. Three features of this difficult illness are pertinent to the question of complicated grief: the long period of anticipation and care giving prior to death, the harsh and unpredictable course of the disease, and the possibility that some survivors may also have HIV and may witness the death of their loved one as a rehearsal for their own death. The literature review revealed three articles that focused on individuals who had experienced a death of a loved one because of HIV/AIDS [167-169]. These articles are summarised below. Goodkin and colleagues [167] examined the impact of a semi-structured bereavement support group among HIV homosexual males in the United States (Level II). One hundred and sixty-six men were randomly assigned to either a bereavement support group or a control group (97 HIV-1 seropositive and 69 HIV-1 seronegative). Participants were assessed at entry to the study and 10 weeks following the intervention using psychosocial questionnaires, a semi-structured interview for psychopathology, a medical history and physical examination, urine collection and phlebotomy. Men in the bereavement support group reported significantly reduced bereavement-related distress. Control subjects showed no significant decrements in overall distress, although a significant decrement in grief level was observed. The researcher concluded that a brief group intervention can significantly reduce overall distress and accelerate grief reduction in a sample of bereaved subjects unselected for psychopathology or at high risk for subsequent mal-adjustment. Conclusions based on this study must be considered cautiously given the lack of detail regarding the support group intervention. As well, the generalisability of the intervention results to a population with more severe distress and or psychopathology remains to be established. Summers and colleagues [169] used a longitudinal design to examine AIDS-related grief resolution and psychiatric morbidity in 286 HIV-positive (n=222) and HIV-negative (n=64) homosexual men examined between 1989 and 1993 in San Diego, USA (Level III-2). One hundred and seventy-one men reported a loss within the previous 12 months. Based upon self-report scores from the Grief Resolution Index obtained from the Texas Revised Inventory of Grief (TRIG) [71], bereaved men were classified into two groups: resolved grievers (N=140, 82%) and unresolved grievers (n=31. 18%). No difference was found between the resolved and unresolved grief groups in relation to multiple losses, weeks since death, intimacy of relationship to the deceased, or lifetime psychiatric disorders. Men with unresolved grief were significantly more likely to have major depression and panic disorders.

76

Results of this study are limited by the fact that the sample included predominantly white, middle-class educated gay males. Higher rates of distress may be present in populations that are less advantaged (ie, drug users, individuals with other medical conditions, lower income groups). The time frame for this study (ie., data collected between 1989-1993) is another limitation that must be considered when interpreting these findings. As well, the study was conducted overseas and may have less relevance to the Australian culture. Despite these limitations, the finding that a relatively small number of participants were classified as having a more difficult grief reaction provides further evidence to suggest that screening is necessary to identify those at greatest risk so that allocation of limited bereavement resources can be most appropriately used. A third study reported earlier by Van den Boom (see Euthanasia section below), undertaken within the Netherlands suggested that survivors of individuals who had experienced euthanasia may not be at greater risk for complicated grief [168] (Level IV). However, the quality of the death appeared to be an important factor associated with complicated grief responses. Measurement difficulties and sample size limitations cloud interpretations of this study. Summary Little research has been undertaken to test the effectiveness of different bereavement interventions with individuals at risk for complicated grief reactions associated with HIV/AIDS cause of death. The few studies that have been reported have been conducted in non-Australian care settings and are not recent making it difficult to draw firm conclusions regarding future service development for the Australian HIV/AIDS associated bereaved population.

Euthanasia Although grief is a normal reaction to the death of a loved one, a complicated grief response may be more likely when the cause of death is perceived to be unnatural (eg., suicide, homicide, etc) [170, 171]. Because euthanasia is considered an unnatural type of death, it is possible that survivors of this type of death may be at greater risk for a complicated grief response [172]. However, family members of individuals who have received euthanasia may be different from family members of individuals who have committed suicide because the former group of survivors will have had time to prepare for the patients death and may have been part of the decision for this action [173]. Each year approximately 3200 people die in the Netherlands as a result of euthanasia [174]. Euthanasia requires the active termination of life with socalled thanatic drugs [168]. To date, only two studies have been reported that describe the effect of this type of death on the grief response of survivors [168, 171]. These are reported here. In 1995 in the Netherlands, approximately 50% of people with AIDs made the necessary arrangement for a possible death by administration of thanatic drugs (Level III-2). In approximately 50% of these

77

instances, euthanasia was performed. Van der Bloom [168] interviewed 60 relatives of 52 deceased AIDS patients (Level IV). The majority of participants were partners of the deceased patients (18 homosexual and 6 heterosexual partners) and family members (16 females, 8 males). Reactions from survivors following the death of the patient included feelings of sadness and sorrow (95%), loneliness (67%), sleeping problems (56%), apathy (27%), eating disorders (17%), psychomotor agitation (9%), psychosomatic problems (22%), depressive episodes (20%), alcohol abuse (13%) and use of sleeping medications/tranquilizers (31%). These interviews appear to have been conducted within the first year of the death of the patient. The investigators report no significant association between the prevalence of depression in survivors and the method of death. The researchers report that when the euthanasia process was complicated, grief became complicated. They indicated that in six of the 12 cases of euthanasia examined, grief was complicated (not defined explicitly) in the following situations: the patient died the moment the injection was given, after injection the patient remained conscious for another 4 to 6 hours, at the moment of euthanasia the physician asked the relative to administer the medication and relatives had to decide when euthanasia should be performed [168]. The researchers reported that in two cases relatives developed serious psychopathology (not defined). This study is limited by the small sample size, lack of measurement and definition precision and the fact that the study is now 10 years old. In a more recent study conducted in the Netherlands by Swarte and colleagues [171], 189 bereaved family members and close friends of terminally ill cancer patients who had died by euthanasia and 316 bereaved family members and close friends of comparable cancer patients who died of natural death between 1992 and 1999 were compared (Level III-2). Symptoms of traumatic grief were assessed using the Texas Revised Grief Inventory and post-traumatic stress was assessed using the Impact of Event Scale. The bereaved family and friends of cancer patients who died by euthanasia coped better with respect to grief symptoms and post-traumatic stress reactions than the bereaved of comparable cancer patients who died a natural death [171]. This difference was independent of other risk factors. Summary Results from studies in euthanasia must be considered within the context of Australian health care. Euthanasia is not legal within Australia. However, from the perspective of this review, the question of preparedness for death and the degree of trauma and suffering associated with the patients death may be pertinent issues when examining family members that may be at greater risk for a complicated grief response. Preparation for the patients death and a sense that the death was peaceful and not distressing may be factors associated with a persons bereavement response.

78

Older adults Only a small proportion of bereaved persons in any age group suffer complicated grief. Among older adults baseline levels of health and adaptive capacity are lower and can be exacerbated by the stress associated with bereavement [175]. They may also have fewer economic and social resources to help them buffer the effects of loss. Among the elderly, spousal bereavement is most common and is most frequently studied. The death of a spouse in old age may interact with, or intensify the consequences of, other stressors that tend to cluster in late-life, such as chronic illness and disability, retirement and involuntary change of residence [176]. There is evidence that older widowed persons benefit from social support, and over time increase the percentage of other widowed persons in their friendship networks [177]. As well, those who use the time during a spouses lingering illness to prepare for how to deal with the practical consequences of the death (eg, learning how to drive, handle finances, make new friends), report less emotional disruption at the death and increased success in coping with the practical consequences of the loss [175]. Wells and Kendig [178] report that a period of care-giving by the bereaved spouse increases the spouses sense of competence and coping abilities. Five relevant empirical studies were identified that focused on older adults and complicated grief. These are summarised here. Beery and colleagues [142] examined factors associated with depression and traumatic grief among caregivers of terminally ill spouses. This study was summarised earlier in this report (See Palliative Care section, pp.93). The finding reported from this study most relevant to this age group is the association between caregiver burden and traumatic grief and subsequent depression. It is possible that the caregiver gratification measure used in this study may not have adequately measured the benefits associated with the caregiving experience. As well, the heterogeneity of diagnostic groups of patients in the study was an acknowledged limitation. Boerner and colleagues [39] undertook a descriptive study to examine whether patterns of coping with stress continued over a four year time period following the loss of a spouse. Data was obtained from 92 widows or widowers with one pre-loss and three post-loss follow-up assessments. Individuals with initially low distress continued to do well four years following the death of their loved one. Individuals who were chronically depressed were more likely to be experiencing distress at 48 months post-loss. Participants with high distress initially and over time suggested that this pattern remained chronic only for those who had reported high distress pre-loss. This study is limited by the small sample size and attrition (those who participated were less depressed) and the fact that participants with poorer long-term adaptation may have been under-represented in this study.

79

Prigerson et al. [24] obtained data from 82 participants diagnosed as depressed who were recruited through a geriatric centre to determine if symptoms interpreted as CG could be identified and distinguished from bereavement related depression. All participants were being treated with nortriptyline. She collected data at 3 to 6 months and again at 18 months. Two distinct clusters were identified, one reflecting the symptoms of complicated grief and the other depression. CG scores were significantly associated with impairments in global functioning, mood, sleep, and self-esteem at 18 months. The authors acknowledge that study group selection may have resulted in an underestimation of CG reactions and results cannot be generalised beyond this very specific clinical population. Szanto and colleagues [179] studied suicidal ideation in a sample of 130 elderly bereaved widows/widowers. Measures were administered over various time points during an 18 month time period including the ICG, the Becks-Kovacs Scale of Suicidal Ideation and the Hamilton Rating Scale for Depression (Level III-2). Groups of active, passive and suicide ideators as well as non-ideator controls were compared. Fifty-seven percent of participants with high CG scores were found to be ideators during the follow-up period versus 24% of participants with low CG scores. Thirty-nine percent of men were active or passive ideators compared with 26% of women. Active suicide ideators had higher CG scores than passive- or non-ideators. Active/passive ideators had significantly higher levels of depression, feelings of hopelessness, CG, anxiety and society support than non-ideators. Depression was associated with ideations of suicide and a history of suicide attempts was associated with an increased likelihood of suicidal ideation. Multivariate analyses were not undertaken to look at effectiveness of different interventions on suicidal ideation due to small sample size in each treatment cell. The researchers concluded that high levels of CG and depression increase vulnerability to suicidal ideation. Latham and Prigerson (2004) [126] examined the influence of CG on suicidal thoughts or behaviours in 309 bereaved elders (74% female, median age 64 years) at an average of 6.2 months post-death at baseline and 10.8 months at follow-up (Level III-3). Participants completed the Yale Evaluation of Suicidality, the Inventory of Complicated Grief, the Structured Clinical Interview for DSM-IV and the Interpersonal Support Evaluation List. Cross-sectionally CG was associated with a 6.58 (95% CI: 1.74-18.0) times greater likelihood of high suicidality at baseline, and an 11.3 (95% CI: 3.33-38.10) times greater risk of high suicidality at follow-up after controlling for gender, race, major depressive disorder, PTSD and social support. Longitudinally, CG at baseline was associated with an 8.21 (95%CI: 2.49-27.0) times greater likelihood of high suidicality at follow-up, controlling for the above confounders. They conclude that CG poses an independent psychiatric risk for suicidal ideation. Summary A number of studies examine issues associated with grief in older adults focusing on widows/widowers and survivors of individuals who died from progressive illnesses (e.g. cancer, heart disease, etc). Most of the research has been descriptive and correlational with few methodologically sound intervention studies

80

targeted toward older adults with complicated grief. Outcome studies have produced mixed, but somewhat positive results. For example, interventions that provide support and encourage expression of emotions in the months after a traumatic or sudden death appear to be associated with reduced symptoms [145, 146]. Similarly, high risk bereaved individuals who receive support with their grief immediately after the death appear to benefit, as indicated by lower levels of physical symptoms, depression, worrying, and use of health care services [147]. Of importance with regard to this population, is the fact that psychological distress associated with bereavement is likely to involve physical and environmental co-determinants in addition to the expected emotional loss [175]. Older adults may be at increased risk because their baseline health is poorer and they may have fewer resources and supports. And although the fact that many older individuals may have had time to prepare for the death of their spouse, the loss of a life-long relationship that may have continued for many decades may leave the survivor with a deep sense of loss and loneliness [153].

Mental Illness Five studies addressed complicated grief consistent with our definition in populations with serious mental illnesses. Macias et al. [180] explored the prevalence of prolonged severe grief among adults with serious mental illness by studying retrospective accounts of 33 bereaved individuals who reported the death of a close friend or family member (Level IV). The effects of situational factors were tested as predictors of severe and prolonged grief. These included residing with the close friend or family member at the time of the death, the suddenness of the death, having low social support and having concurrent stressors. Findings confirmed that the more situational factors occurred at the time of death, the more severe the grief reaction and this was not related to psychiatric symptomatology. Limitations include the small sample and reliance on participants self-reports. A similar study by Jones et al. [181] with the same participants (33 bereaved with serious mental illness) found that most of the participants with prolonged or severe grief had not received any preparation for parental loss (Level IV). Limitations of this study include the fact that data was self-reported; poor measures of grief were used; the frequency of complicated situational factors where measured rather than CG; and the small sample. An exploratory study by Piper and colleagues [182] investigated the prevalence of loss and complicated grief among patients from psychiatric outpatient clinics (Level IV). Measures included the Beck Depression Inventory, the Trait Anxiety Scale, the Brief Symptom Inventory and The Impact of Events Scale. Grief was assessed by the Texas Revised Inventory of Grief. Two variables, intrusion or avoidance,

81

and social dysfunction were used to differentiate between two levels of complicated grief moderate and severe. In addition, a minimum period of 3 months was required for a patient to meet the criteria for CG. Thirty-one percent (73 of 235) of psychiatric outpatients who experienced the death of a significant other met the criteria for moderate CG and 29% (69 of 235) for severe CG. The average time since death was approximately ten years, indicating that these participants had been suffering from long-term CG. The severe CG group reported significantly higher levels of social dysfunction and depression disturbance variables. Depression, anxiety and grief symptomatic distress were found to be significantly higher for the severe CG group compared to the moderate CG and those who had not experienced any loss. However, there was no control group consisting of people in the general population who did not seek psychiatric services [182]. Melhem and colleagues [100] examined the rate of lifetime and current psychiatric disorders, DSM-IV Axis I disorders, in a group of participants (n=23) with traumatic grief (i.e. co-morbidity of traumatic grief with other psychiatric disorders) (Level IV). See Chapter 3 for a full review. Forty-four percent had only one concurrent diagnosis, 48% had two or more additional psychiatric disorders, and eight percent had a lifetime diagnosis. In addition, 52% had a major depressive disorder and 30% had PTSD. Fifty-two percent of all the participants had a prior psychiatric history. The ICG scores and functional impairment were found to be higher among participants who had more than one concurrent Axis I diagnosis. Hence, prior psychiatric illness may be a risk factor for traumatic grief but not necessarily non-traumatic CG. Simon et al. [127] investigated the frequency and implications of the death of loved ones and complicated grief on 103 patients with bipolar disorder. Among those who reported a significant death, 25% met criteria for current CG (Level III-3) (see Chapter 4). The presence of CG was associated with increased rates of panic disorder, alcohol abuse co-morbidity, higher rates of lifetime suicide attempts, greater functional impairment and poorer social support. In a new study currently in press, Johnson and colleagues [183] sought to investigate attitudes about grief symptoms, receptivity to mental illness and stigmatisation attributable to grieving (Level IV). Participants included 135 recently bereaved persons (1-3 months post-loss) recruited as part of the Yale Bereavement Study recruited through obituaries in the local paper, newspaper advertisements, flyers, personal referrals, chaplain referrals (24%) and widowed persons residential service (76%). Interviews were undertaken with interviewers required to demonstrate nearly perfect agreement (Cohens Kappa = 0.90) with the Principal Investigator in regard to their diagnosis of psychiatric disorders (e.g. MDD) in a series of five interviews before the study commenced. Measures included the Structured Clinical Interview for DSV-IV and the Inventory of Complicated Grief-Revised [28]. Attitudes about grief, receptivity to treatment and concerns about stigmatization were assessed using the Stigma Receptivity Scale (SRS). (Cronbachs alpha =0.64). Sixteen participants (12%) had a psychiatric disorder and 16 (12%) had had CG at some point within the

82

study observation period. Six of 16 people had both CG and a psychiatric disorder, 10 had only a psychiatric disorder and another 10 had CG only. Both a psychiatric diagnosis and CG were independent predictors of recent mental health services use (p=0.02 and p=0.05 respectively). Two individual SRS items significantly predicted recent use of mental health services among bereaved individuals with and without CG. First, receptivity to a bereavement support group significantly increased the odds that they had used any mental health services in the past 60 days compared with individuals that were not receptive to the support group (adjusted OR = 5.14, 95% CI: 1.11, 23.85). Second, individuals who were concerned about meeting criteria for a mental illness were significantly less likely to have received any mental health treatment than those who were not concerned about meeting these criteria (adjusted OR = 0.07, 95% CI: 0.01, 0.58). Of the three SRS subscales, only subscale 1, attitudes towards a CG diagnosis predicted recent use of mental health services, specifically that negative attitudes towards a CG diagnosis significantly decreased likelihood of recent treatment (adjusted OR = 0.34, 95% CI: 0.15, 0.75). The SRS as a whole was a significant predictor of service use in unadjusted analyses only (OR: 0.70, 95% CI: 0.48, 0.93). Results suggest that, receptivity to a bereavement support group is a strong predictor of use of that service and concerns about having a mental health diagnosis substantially decreases likeliness of service use. Receptivity to other psychological treatments such as medication and psychotherapy does not predict of use of mental health services. Furthermore, knowledge of a mental health diagnosis and beliefs about others perceptions and reactions towards a CG diagnosis does not increase the likelihood of service use. These findings highlight the need for both assessing receptivity to supportive services for the bereaved and working with bereaved individuals to minimise their concerns about a mental health diagnosis. Limitations include a small sample size in the sub-group analyses and lack of a comparison group of non-bereaved. Additional research conducted on significant others of those diagnosed and/or treated for complicated grief is needed to determine whether family and friends withdraw support and/or develop a greater appreciation and understanding of the severity of the illness, how and why it manifests itself and the bereaved persons need for help in adjusting to the loss. Summary The findings of studies in mental illness suggest that the same situational factors that are associated with a more complicated and prolonged grief in the general population also have an impact on the lives of people with mental illness, further complicating their lives in addition to their psychiatric conditions. These results remain inconclusive due to the lack of control groups and the reliance on self-reports. Cultural groups Only three studies were located that examined complicated grief in a sub-section of cultural groups. Nakao et al [184] (Level III-2) examined the relationship between grief reactions and alexithymia in 54 Japanese women (33 outpatients attending a Psychosomatic Clinic and 21 healthy volunteers). Measurements included the Texas Inventory of Grief (TIG), the Toronto Alexithymia Scale and the

83

Profile of Mood States. Multiple regression analysis, controlling for effects of age, the length of time since the death, and the group (psychosomatic or control) indicated that complicated grief reactions (as measured by the TIG) may be closely associated with both alexithymic character and mood states in bereaved Japanese women. The most significant factor was difficulty in identifying feelings and higher ratings on the POMS tension-anxiety and depressions scale. Limitations include generalisability, the nonvalidation of the Japanese version of the TIG and the analysis of patient and non-patient data together.

Rates of complicated grief among 151 psychiatric clinic patients in Karachi were determined in a study by Prigerson et al. [185]. In addition, the influence of risks (such as mode of death, age, gender, time from death and the relationship to the deceased) on the likelihood of meeting diagnostic criteria for CG was examined (Level IV). A third of patients were diagnosed with CG. Although violent deaths did not increase the risk of CG, gender and kinship of the deceased did. Women were 3.3 times more likely to meet criteria for CG. Spouses followed by parents were the most likely to meet criteria for CG compared to non-first-degree relatives. Interpretation of results is limited because the researchers did not conduct standardised psychiatric assessments of participants. As well, there was a lack of information obtained related to factors surrounding the death. Momartin and colleagues investigated a possible phenomenologic overlap of complicated grief with PTSD and depression in 126 Bosnian refugees [186] recruited from a community centre in Sydney, Australia (Level III-3). The sample was supplemented by a snowball method (86% response rate) and participants completed the Clinician-Administered PTSD Scale (CAPS) for DSM-IV (a structured clinical interview to assess PTSD), and the Structured Clinical Interview for the DSM-IV (SCID) to assess the presence of major depressive disorder and dysthymia; the Core Bereavement Items (CBI) to assess complicated grief symptoms. All measures were translated and back-translated. The average time since exposure to severe traumatic experiences was 5 years (range 2 7). Results showed the rate of PTSD to be 63%. Widowhood, the dimension of traumatic loss, and human rights violations were significantly associated with complicated grief, however PTSD was unrelated, adding support to growing evidence that the two syndromes are largely distinct. A substantial association was found between depression and complicated grief. The authors concluded that complicated grief and PTSD, while sharing symptoms of intrusion, can be distinguished from each other. Complicated grief appeared to be one pathway leading to persisting depression. Stated limitations included the modest sample size and the non-random selection of subjects. The translation of measures may have contributed to semantic and linguistic errors. Indigenous populations No studies were identified in this review that specifically addressed complicated grief in Indigenous populations. The bulk of the research material identified focussed on intergenerational grief, historical

84

grief, or grief associated with the stolen generation. We briefly comment on studies that provide background information relevant to grief in Indigenous populations. According to the definition of complicated grief within this review if integration of the loss does not occur and acute grief is prolonged it could be extrapolated that Indigenous populations do indeed represent an at risk population [4, p. 253]. The high rate of death within Australian Indigenous communities means that individuals are continuously grieving for relatives who have died [187]. No other group of people within Australia experience the number of early deaths and death from non-natural causes as Indigenous Australians. Aboriginal people are exposed to more perceived high-risk bereavements due to the high rate of premature mortality and the types of losses (accident, violent or illness) [188]. Additionally, more people in the community tend to be affected by a death due to the closeness and connectedness of Indigenous communities. The high rate of funerals that are attended results in an immersion in death and grief [188, p. 9]. Being in a situation where death is more frequent and the types of death more traumatic may result in a sense of an overwhelming burden of stress, making it more difficult for people to deal with individual losses [188, p. 10]. The burden of generational losses and separations may add an additional burden that increases vulnerability. Swan states that it must be acknowledged that Australian Indigenous people have dealt with overwhelming loss in resilient and positive ways. However, there is a lack of appropriate mechanisms or avenues for the bereaved to seek grief counselling in many Indigenous communities. Suicide in this context is labelled as the grieving suicides by Tatz [187]. Tatz calls for urgent attention to the severe lack of counselling that is needed to break the perpetual cycle of grief within Indigenous communities. The national consultancy to develop an Aboriginal Mental Health Policy found trauma and grief are amongst the most significant mental health problems for Aboriginal people [189]. For the majority of Aboriginal people consulted, counselling was identified as a major area of need. In addition, there is a need for policy and programs to support Aboriginal people to deal with trauma and grief [188, p. 10]. Narrative therapy has been suggested as a possible suitable counselling program for Aboriginal people as it fits well with Aboriginal culture [188]. In the context of the inclusion of complicated grief in the DSM V, Walle believes that within the context of American or Canadian first peoples that the DSM is culturally bound and that the diagnosis category of bereavement needs to be expanded to deal with Indigenous peoples experience of broad-based cultural losses [59, p. 52]. Walle argues that the current bereavement category is not adequate as it does not deal with a number of specific circumstances experienced by Indigenous people. Four distinct categories are suggested that include bereavement due to the loss of: an individual; a group; a way of life; and a persons

85

niche in society. Walle states that many mental disorders suffered by native people stem from comparable, recurring, and predictable impacts [59, p. 63]. Despite cultural differences many Indigenous peoples have similar dysfunctions. By recognising this reality and upgrading the DSM accordingly, native people can be more effectively served by counsellors and therapists [59, p. 65]. A feature of studies within Indigenous populations in the United States and Canada is the concept of historical trauma (HT). HT is a cumulative emotional and psychological wounding over the lifespan and across generations, emanating from massive group trauma experiences; the historical trauma responses (HTR) is the constellation of features in reaction to this trauma [190, p. 7]. PTSD is inadequate in capturing the influences and attributes of native peoples historical trauma. This historical unresolved grief is passed down from one generation to the next [191]. The massive loss experienced by the American Indians for example, is postulated to have contributed to the current social pathology of high rated of suicide, homicide, domestic violence, child abuse, alcoholism and other social problems among American Indians [191, p. 56]. Indeed, the high rate of suicide and suicide attempts in the Canadian Artic are though to be an expression of complicated grief [192]. In the context of Australian Indigenous populations, Wanganeen [193, p. 13] states that we need to become aware that we cannot carry our Spiritual Ancestors grief any longer as we have become who we are today because of that brutal invasion. Believe it or not, we today are the carriers of their grief. We must become consciously aware of how we are passing down the same grief to our children and grandchildren. The recently published report from the Western Australian Aboriginal Health Survey [194] provides an important insight into the health and wellbeing of Aboriginal and Torres Strait Islander children in WA aged between 0-17 years of age (n=5289). While this report does not meet the inclusion criteria for this review it was deemed worthy of consideration as it provides data describing the mental health and social and emotional wellbeing of the children surveyed in this study. Mental health was described as representing one part of the concept of social and emotional wellbeing and includes mental health; suicide and self harm; emotional, psychological and spiritual wellbeing and issues impacting specifically on wellbeing in Aboriginal and Torres Strait Islander communities such as grief, loss, trauma and issues surrounding the forced separation of children from their families. The specific issue of loss and grief was not examined in isolation but included in a cluster termed Life Events whose effects were measured using the Strengths and Difficulties Questionnaire. Findings showed that 24% of Aboriginal children compared with 15% of similarly aged non aboriginal children were considered at high risk of clinically significant emotional or behavioural difficulties (emotional symptoms, conduct symptoms, hyperactivity, peer problems and pro-social problems). Similar findings were shown for children in the 4-11 years age group, and the 12-17 years age group. It was concluded that this was a major disruptor of childrens development as a higher proportion of children at risk of clinically significant emotional or behavioural

86

difficulties had problem behaviours compared with children at low risk of significant emotional or behavioural difficulties. These findings suggest that Aboriginal and Torres Strait Islander children may potentially be at higher risk for CG than non aboriginal children and that further research (specifically related to CG in this population) as well as new approaches/interventions are required to address the needs of this culturally disadvantaged population where normal processes of child development and future life prospects are compromised. A National Indigenous Palliative Care Needs Study released in April 2003 [195] reported that many palliative care services are moving away from a standard bereavement counselling model (for example programmed phone calls and cards) to conducting individual risk assessment followed by more tailored approaches where needed. For example, the Mid North Coast Palliative Care Service in NSW identified the following risk categories based on McKissocks criteria [9]: death of a child; sudden death; trauma; ambivalence in the relationship; pre-existing psychopathology (including unresolved losses, alcohol and drug dependence, history of depression, personality disorder); concurrent crises; centrality, perceived preventability; decreased (or lack of) role diversity; lack of reality and overly prolonged dying. No evaluation of these risk categories within Indigenous populations has been undertaken. What is noticeable about this risk assessment model is the likelihood that nearly every Aboriginal or Torres Strait Islander client would be assessed as high risk because of a burden of unresolved losses, current crises and other risk factors. Unresolved losses are likely to include not only other recent deaths in the family but other unresolved losses stemming from invasion and stolen generation issues. This underlies the need for action to provide better services for Aboriginal and Torres Strait Islander clients.

Summary No studies were identified in this review that specifically addressed complicated grief in Indigenous populations. The exposure of Aboriginal people to more perceived high-risk bereavements due to the high rate of premature mortality and the types of losses (accident, violent or illness) and the closeness and connectedness of Indigenous communities would suggest that they represent a vulnerable high-risk group for CG. Recommendations: Complicated Grief in Specific Populations

Further research is needed to identify empirical evidence related to children and adolescents and in particular: risk factors that may predispose them to complicated grief in later life the criteria used to define CG in the context of childhood and adolescent experiences, instruments most appropriate for measurement of CG in child and adolescent populations

87

the extent to which CG is distinct from traumatic experiences in the child and adolescent population interventions most appropriate to the grief experiences of children and adolescents

Research is needed to identify risk factors for CG in parents associated with perinatal death, infant death and child and adolescent death, and the death of an adult child.

Research is needed to identify most appropriate interventions to respond to CG in the following populations: parents who have experienced the death of a perinatal or neonatal infant, parents who have experienced the death of an infant, parents who have experienced the death of a child/adolescent and parents who have experienced the death of an adult child.

Further research is warranted to identify the elements of a palliative approach to care that may be instrumental in achieving positive family bereavement outcomes.

Research is needed to examine preparation for death and perceptions of the quality of death by bereaved survivors as factors associated with CG.

Further research is needed within the Australian health care setting focused on individuals who have experienced the death of a loved one due to HIV/AIDS to examine their risk for CG.

Further research is warranted to examine the effectiveness of a support group intervention for individuals with CG following the death of a loved one due to HIV/AIDS.

Further research is needed to better understand the needs of older adults who are not in a spousal relationship. The bereavement needs and grief risks for individuals that have never married, are divorced, have lost an adult child, friend, sibling or other relative are notably absent in the empirical literature and should become a future research priority.

Research is needed to examine the experience of CG for people with a mental illness Research to identify risk factors specifically related to CG in the Indigenous population is needed.

88

CHAPTER 9: GRIEF INTERVENTIONS

Schut and colleagues (2001) [152] grouped bereavement interventions into three categories: primary preventative, secondary preventative and tertiary preventative. Primary interventions focus on people considered to be experiencing uncomplicated (normal) bereavement. Secondary interventions target those who are at risk of complications of bereavement and tertiary interventions are targeted to those experiencing bereavement-related problems (complicated grief). For the purpose of this review, we focused on tertiary preventative interventions in keeping with the aim of the report. The literature search identified 25 studies (1990-2005) that investigated the effectiveness of various treatments and interventions for CG and met the inclusion criteria. These 25 studies selected for detailed review evaluated diverse types of interventions designed to ameliorate the adverse physical and psychological outcomes associated with CG and were classified under four categories: 1. Pharmacotherapy 2. Support groups or counselling 3. Psychotherapy-based interventions which included o o o o o Group therapy Cognitive-behavioural therapy Psychodynamic therapy Behavioural therapy Interpersonal therapy Touch therapy Eye movement desensitization and reprocessing

4. Other interventions such as o o

1. Pharmacotherapy A comparison of the effectiveness of paroxetine and nortriptyline for symptoms of traumatic grief by Zygmont et al. [196] in an open-trial pilot study with an archival contrast group, found both drugs to have comparable effects in improving depression and grief intensity symptoms (n=15) (Level III-3). Traumatic grief was measured by ICG with subjects having a baseline score of greater or equal to 20. The level of grief symptoms decreased by 53% using both drugs (p=0.0002). However, Zygmont favoured the use of paroxetine in general psychiatric practice because of the higher rate of diagnostic co-morbidity in that population, greater chronicity of symptoms and greater safety of paroxetine in overdose. Study limits included the small study size, the use of an archival control group, the heterogenous nature of the paroxetine group, potential confounding effects of psychotherapy, and the significantly different median time since death between the two groups.

89

2. Support groups or counselling Goodkin et al. (1999) [167] assessed the impact of a semi-structured bereavement support group among HIV 1 seropositive and seronegative homosexual men who had lost a close friend/partner to AIDS within the previous six months (Level II). Grief level was measured using the Texas Inventory of Grief. A total of 166 subjects were randomly assigned to intervention and control groups and assessed at entry and at 10 weeks. The intervention consisted of 90-minute weekly sessions led by two co-therapists, while controls were allowed to continue the level of psychosocial and medical care used prior to baseline. No data was provided on the gender or level of training of the therapists. A significant reduction in grief level was found between the two time periods if other factors affecting distress level were controlled. A repeated measures analyses of covariance showed a higher statistically significant intervention effect on grief scores for the intervention group (F=52.07, P<0.001) compared to the control group (F=20.75, p<0.001). Control subjects showed no significant decrements in overall distress although a significant decrement in grief level was observed. The trial entry criteria for this study limited generalisability of findings. 3. Psychotherapy-based interventions Group therapy was the main type of psychotherapy-led intervention. Of the nineteen studies that tested group therapy, seventeen were based on one major parent study carried out by Piper and colleagues, referred to as the Edmonton Trial in Canada (2001 2005). A matched control design was used with patients with CG matched for personality characteristics, gender and age; then randomising the participants into either an interpretive or supportive psychotherapy group. The groups: In interpretative therapy, the primary objective is to enhance the patients insight about repetitive conflicts and trauma that are associated with the losses and that are assumed to serve as impediments to experiencing normal mourning process. In supportive therapy the primary objective is to improve the patients immediate adaptation to their life situation. It is assumed that improvements in symptomatology and social (role) function can be achieved through the provision of support and problem solving. Patients were scheduled to weekly 90 minute sessions for 12 weeks. The therapists: Therapists (1 male, 2 female) were experienced in group therapy (10-14 years), had participated in a pilot group, followed a technical manual for both form of support groups (interpretative and supportive) for loss patients. Adherence to the manual was independently observed and rated. All therapy sessions were audio-taped and observed through a one-way mirror. Recruitment: Patients were assessed to see if they met CG criteria by completing three brief questionnaires including the Pathological Grief Items (PGI) adapted from work by Prigerson (1995), the Impact of Events Scale and the Social Adjustment Scale. The 17 papers connected to the Edmonton trial [85, 197-210] examined how various patient characteristics and group processes interacted to influence treatment outcome. All 17 papers had a level of evidence of III-1. Outcomes: Assessment of outcomes included 14 measures that covered 15 variables in the areas of grief symptoms, interpersonal distress, social (role) functioning), psychiatric symptoms, self-esteem, life

90

satisfaction, and physical functioning. Severity of disturbance for individual target objectives was also assessed. Some of the predictors of these outcomes examined in the 17 papers were psychological mindedness, quality of object relation, alexithymia (deficit in cognitive processing and regulation of emotions), gender, therapists reactions to patients, patients affect, therapeutic alliance, group process variables and their relationship to dropout, perceived social support, patients interpersonal functioning (attachment to lost person, quality of object relations, level of recent social role functioning), and perceived group climate (engagement, avoidance, conflict). In general, these papers explored the effect of such patient and other characteristics on psychotherapy outcome of interpretive versus supportive therapy. Key findings from these studies are highlighted below. Piper et al. (2001) [208] used a randomised clinical trial design to investigate the interaction of two patient personality characteristics, quality of object relations (QOR) and psychological mindedness (PM), with two forms of time limited short-term group therapy, interpretive and supportive (Level III-1). One hundred and thirty-nine patients were randomly assigned to one of the treatment groups, supportive and interpretive. The outcome variables were grief symptoms, interpersonal distress, social functioning, psychiatric symptoms, self-esteem, life satisfaction and physical functioning. Patients with high QOR in the interpretive therapy group showed greater improvements than patients with high QOR in the supportive therapy arm of the study. Patients with low QOR showed greater improvement with supportive therapy. Both therapies resulted in improvements for patients with high PM scores (F(1,91)=7.55, p<0.007) However, the study did not include a control group and most patients did not include grief as part of their presenting complaints. In 2003, a similar study that highlighted the importance of assessing patient personality to predict response to short-term group therapy was undertaken by Ogrodniczuk et al. [204] using the NEOPersonality Inventory (Level III-1). The grief symptoms scales consisted of the Intrusion and Avoidance subscales of the Impact of Events Scale, a set of pathological grief items by Prigerson et al, (1995) and six subscales of the Social Adjustment Scale Self Report. For patients in both groups (n=107), extraversion (F=6.88; df=1,89; p=0.002), and conscientiousness (F=8.33, df=1,89, p=0.005) were directly associated with a decrease in grief symptoms. Limitations included lack of examination of different types of loss and its association with NEO-Personality dimensions. Ogrodniczuk et al. [203] investigated changes in perceived social support before group therapy treatment onset, after treatment completion and 6 months post treatment (Level III-1). Sixty-one patients were randomly assigned to either interpretive or supportive group therapy. Perceived support improved after psychiatric treatment of CG for both groups (F(1,60)=7.66, p=0.008). A reduction in depression

91

severity was associated with improvement in perceived social support (t(60)=2.03, p=0.047). This study lacked a control group comprised of individuals with CG who did not receive treatment. A measurement of gender differences by Ogrodniczuk et al. [206] found that women generally had better outcomes in both supportive and interpretive short-term psychotherapy compared to men (F(1,43)=4.67, p=0.36)(Level III-1). Grief symptoms were measured by a set of seven pathological grief items by Prigerson et al, 1995, a 7-item Intrusion subscale and 8-item Avoidance subscale of the Impact of Event Scale and 13-item Present Feelings subscale of TRIG. This gender effect may be partially mediated by mens lack of commitment to the group and perceived incompatibility with other group members. Limitations of the study related to the small sample size (n=47) and that the effect of patient preference for a particular treatment was not investigated. Psychotherapy-based interventions other than group therapy were conducted by Shear et al. [211, 212]. A pilot study by Shear and colleagues [211], assessed the effectiveness of traumatic grief treatment and found CG scores to decrease significantly for both completers (n=13) and intent-to-treat (n=21) groups, compared to interpersonal therapy alone (Level IV). The grief treatment protocol comprised imaginal re-living of the death, in vivo exposure to avoided activities and situations, and interpersonal therapy. Shear et al. [212] later compared the efficacy of an approach in CG treatment with a standard interpersonal psychotherapy, using a randomised controlled clinical trial stratified by manner of death of loved one and treatment type. The study included 83 women and 12 men (Level II). Although both treatments produced improvement in CG symptoms, CG treatment was an improved treatment over interpersonal psychotherapy, showing higher response rates (51% for CG treatment compared to 28% for interpretive psychotherapy, p=0.02) and faster time to response (p=0.02). Limitations of this study included the high proportion of patients using psychotropic medications (45%), heterogeneity of sample and attrition. Saltzman et al. [106] evaluated the effectiveness of a school-based screening and group treatment protocol, trauma- and grief-focused group psychotherapy, for adolescents (n=26) exposed to community violence and trauma either due to losing someone to a traumatic death or witnessing a traumatic act (Level IV). Grief was assessed using the Grief Screening Scale and UCLA Trauma-Grief Screening Interview. Results of the study, which used a pre-test and post-test design, suggested that group participation was associated with improvements in posttraumatic stress, CG symptoms (t(7)=3.38, p=0.015) and academic performance. However, this study had a small sample size, did not have a control group, and used a limited battery of measures to assess treatment outcomes.

92

4. Other interventions A study by Kempson [135] examined the effect of touch therapy had on grieving mothers who experienced the death of a child within the last 6-60 months (n= 31) compared to a control group (n=34) (Level III-2). Grief was measured by the Grief Experience Inventory. The study employed a quasiexperimental non-equivalent pre-test-post-test control group design. Touch therapy was found to significantly improve despair (F=8.290, p=0.005), depersonalization (F=4.904, p=0.031), and somatisation (F=6.833, p=0.012). However, the actual effect of the intervention on CG was not measured. The major limitations included not having a specific instrument to measure the experience of touch and the use of a non-verbal intervention that they attempted to measure verbally. Sprang [213] set out to determine the differential effects of treatment on traumatic stress and complicated mourning by comparing eye movement desensitisation and reprocessing (EMDR) (n=23) and guided mourning (GM) (n=27) (Level III-2). Grief was measured by The Texas Revised Inventory of Grief. Clients completed measures designed to assess psychosocial and behavioural symptoms of loss before and after treatment and at a 9-month follow-up period. Even though the intensity of grief decreased significantly over the 9-month period for both groups, there were no remarkable differences according to treatment group (F(1,50)=0.87, p=0.659). A lack of a non-treatment group and random assignment made it difficult to distinguish the effect of treatments from the natural bereavement process. Layne et al. (2001) [214] investigated whether participation in trauma/grief-focused group psychotherapy was associated with reduced posttraumatic stress, complicated grief and depressive symptoms (Level IV). Participants consisted of 55 war-traumatized secondary school students in Bosnia. Students completed pre-group and post-group self-report measures of posttraumatic stress, depression and grief symptoms. Complicated grief was measured by the Grief Screening Scale. They further completed post-group measures of psychosocial adaptation and group satisfaction. Preliminary results of the evaluation are promising, with the finding that half of the students showed reliable improvements in the primary outcome measures, particularly a reduction in grief scores (F(1,35)=22.90, p<0.001). However, the study did not involve a control group or random assignment to treatment and relied only on self-reported instruments. Summary Most of the evaluated studies adopted tertiary preventive interventions for complicated grief. Although outcomes are positive, effects are only modest and must be viewed with caution due to inherent methodological problems in study design and implementation.

Seventeen studies used a set of pathological grief items adapted from Prigersons earlier 1995 Inventory. The validity of these items as a sub-scale has not been tested and so results must be viewed with caution.

93

Definitive outcomes associated with group psycho-therapy are not clear and are not embedded within the complicated grief symptomatology, which is the framework of this review. No data was found on the detrimental effects of interventions in complicated grief.

Several of the design flaws in studies identified for this review were related to the recruitment of participants who were already taking part in a support group, were receiving medication for clinical anxiety or depression or had been referred to a psychiatric out-patient clinic for assessment and counselling. In addition, few studies used control groups. Therefore, it is difficult to assess the interplay in the relationship between general psychosocial care and specialist bereavement services before and /or after the death.

Whilst this review of the literature focussed on interventions in complicated grief, it is important to consider them in the light of previous reviews undertaken. These are discussed in more detail in Chapter 10, however we make the following observations. The reviews by Allumbaugh and Hoyt [9] and Kato and Mann [215] showed that Primary Interventions, that is psychological interventions for normal bereavement are not effective interventions. Schut et al [152] found more evidence of efficacy for Secondary Interventions, that is interventions that focused on bereaved persons who had experienced the sudden, traumatic death of a loved one; those that were in a high-risk category (e.g. bereaved parents) but the effects were quite modest in comparison to traditional psychotherapy outcome studies.

Intervention studies for people with complicated grief (i.e. were already suffering from clinical levels of depression, anxiety and other bereavement-induced disorders) at the time of entry into the study (Tertiary Interventions) showed a moderate effect and indicated some proven effectiveness and hold promise for complicated grief [216] (p.488).

The National Bereavement Workshop held in Canberra 2003 made an important qualification in providing support to bereaved individuals that in determining appropriate interventions for the bereaved, interventions should be a need-based assessment and not simply a risk-based assessment. We support this view.

Recommendations: Interventions in Complicated Grief

Future research needs to examine links between assessment, intervention and outcomes that are targeted to well-defined patient populations at well-defined phases of bereavement. Research is needed to demonstrate the efficacy of pharmacotherapy for the reduction of symptoms of complicated grief.

94

Future research is needed into the role of the level of practitioner training, as well as client characteristics such as age, gender, time since loss, and relationship to the deceased in conducting individual or group counselling. Future research is needed into the comparison of client outcomes from individual counselling and group counselling. Future research is needed into gender differences in the effectiveness of individual or group counselling.

95

CHAPTER 10: SUMMARY AND RECOMMENDATIONS

The nature of complicated grief and its relationship to other syndromes and conditions and questions about how complicated grief should be defined, assessed, and classified, have been topics of significant and persistent debate [58]. Researchers have hypothesised that a small, though significant percentage of the population (approximately 9% - 12%) experience complicated grief, and that these individuals are at greatest risk for adverse health effects [5-7]. However, there has been a lack of evidence for good practice in bereavement research and services, especially for those who might be 'at risk' of complicated grief following bereavement. In addition, there is great diversity in the use of adjectives used to describe variations from normal grief and the conceptualisations of complicated grief differed according to the theoretical approach taken by the investigators. This definitional and theoretical confusion has created uncertainty for health care providers and services that endeavour to make sense of the complex and apparently conflicting literature. We conclude this review of the literature in complicated grief by addressing specific questions and making recommendations for future policy, clinical practice and research. When can a diagnosis of complicated grief be made? The criteria for Complicated Grief proposed for inclusion in the DSM-V specify that the particular symptomatic distress must persist for at least 6 months, regardless of when those 6 months occur in relation to the death. Hence, chronic and delayed subtypes of grief are both included in this conception of Complicated Grief, as long as the chronicity refers to, and the delay includes, at a minimum the required 6 months of symptomatic distress. Requiring the distress to last longer than 6 months minimises type 1 error and ensures a higher rate of true positive cases of Complicated Grief. In addition, as a further attempt to be conservative in diagnosing bereaved persons with CG, the stipulation that the symptoms be associated with significant functional impairment should be included [30]. What are the implications of defining someone as at risk for complicated grief? The clear implication of defining someone at risk is the potential for misdiagnosis. For example, people may be diagnosed too early in their grieving process, or they may more accurately meet the criteria for diagnosis of a major depressive disorder as opposed to complicated grief [52]. However, our review of the literature suggests while there is the need for further empirical testing and evaluation for the criteria for complicated grief proposed for inclusion in DSM-V, the measures developed by Prigerson and Horowitz and colleagues are the most statistically rigorous. Horowitz argues that the use of self-report measures to make a diagnosis of complicated grief needs to be supported by clinical observation of patients self-report and clinical interviews. This will enable the clinician to assess the individuals subjective experience and gather other salient information [60].

96

The second implication of defining someone at risk is the boundary between normality and pathology. Our review of the literature indicates there is disagreement over the issue of pathology. Those who disagree that the diagnostic criteria will pathologise grief reactions argue that what the criteria will pathologise is grief reactions that become chronic and predict enduring distress and impairment and not the usual range of suffering that may ensue following the death of a significant other [30]. Prigerson argues that the concerns raised about psycho-pathologising would apply equally well to the diagnosis of major depressive disorder, generalised anxiety disorder, bipolar disorder and post-traumatic stress disorder [30]. Another view is that the evidence for long-term mental and physical health consequences of complicated grief suggests that the lack of access to clinical care is of greater concern than pathologising grief [52]. The third implication is one of stigmatisation. Stroebe and Schut [33, 58] are concerned that the potential of withdrawal of family support is a real concern when a diagnosis and treatment for complicated grief occurs. However, in a new study currently in press, Johnson and colleagues [183] sought to investigate attitudes about grief symptoms, receptivity to mental illness and stigmatisation attributable to grieving (Level IV). Participants included 135 recently bereaved persons (1-3 months post-loss) recruited as part of the Yale Bereavement Study {recruited through obituaries in the local paper, newspaper advertisements, flyers, personal referrals, chaplain referrals (24%) and widowed persons residential service (76%)}. Interviews were undertaken with interviewers required to demonstrate nearly perfect agreement (Cohens Kappa = 0.90) with the Principal Investigator in regard to their diagnosis of psychiatric disorders (e.g. MDD) in a series of five interviews before the study commenced. Measures included the Structured Clinical Interview for DSV-IV and the Inventory of Complicated Grief-Revised [28]. Attitudes about Grief, Receptivity to Treatment and Concerns about Stigmatization were assessed using the Stigma Receptivity Scale (SRS). (Cronbachs alpha =0.64). Sixteen participants (12%) had a psychiatric disorder and 16 (12%) had had CG at some point within the study observation period. Six of 16 people had both CG and a psychiatric disorder, 10 had only a psychiatric disorder, and another 10 had CG only. Both a psychiatric diagnosis and CG were independent predictors of recent mental health services use (p=0.02 and p=0.05 respectively). Responses indicated that 87.5% of those who met criteria for Complicated Grief said that a diagnosis of CG would make them relieved to know that they were not going crazy, 93.8% said that they would be relieved to know that they had a recognizable problem, and 100% said that the diagnosis would help their family members to understand better what they were experiencing. Based on these preliminary results it appears that people diagnosed with Complicated Grief think that the diagnosis would enhance the ability of others to comprehend their suffering. The fourth implication identified within the context of US health system, is the lack of access to health insurance if complicated grief is not identified as a category within DSM-V because the patient does not

97

receive a listed diagnosis that requires professional help [60]. The extent to which this might become a future health insurance funding issue in Australia is unknown. What interventions currently are in use by services in Australia and internationally? Three Australian studies met criteria for inclusion in this review [78, 119, 134] None of these studies assessed interventions. However, a survey of the ten major tertiary paediatric oncology units in Australia and New Zealand was conducted by deCinque and colleagues [134] to determine the current practices that existed in relation to hospital bereavement-based support programmes found that the majority of hospitals provided a multidisciplinary bereavement service for approximately one year after the death of a child. The most common programmes provided were counselling and support groups. However, no formal evaluation of programmes had been undertaken. Does the literature support the effectiveness of interventions in decreasing risk and/or treating complicated grief? From our systematic review, different interventions were suggested by various authors for different bereavement patterns. For example, chronically depressed individuals might benefit from pharmacologic interventions, whereas those struggling with CG may benefit more from cognitive and behavioural interventions [112]; bereaved elders who show a trajectory of chronic depression might benefit from a different intervention focus than those with a CG pattern [39]; and that professional assessments and interventions should take into account the bereaved persons familial and/or social relationship to the deceased [105]. These findings highlight the importance of tailoring interventions, suggesting that the intervention may need to be as individual as the bereavement pattern. What evidence is there exploring the links and pathways between assessment, intervention and intervention outcomes for complicated grief? Currow [153] provides an argument for the necessary links between the prevention of complicated grief, screening and focused interventions from the perspective of the wider population. The article places the question of linkages in the context of a public health framework and provides a thoughtful theoretical framework to guide future research and direction in this area. No studies specific to this research question were identified. However, we identified some studies that addressed aspects of these linkages and some reviews that offered some direction regarding various levels of intervention. Kristjanson and colleagues [78] endeavoured to assess a brief, Australian version of Colin Murray Parkes BRI (1993) and match level of risk with a specific bereavement protocol in a community palliative care setting, using a primary care approach. One hundred and fifty bereaved family members were followed from time of the patients death until six months post-loss. The incidence of individuals in the high risk category was low (7%) and matching of the bereavement protocol to level of risk appeared to be feasible and appropriate. Further testing is required to confirm that the level of intervention provided is

98

concordant with the risk. As well, longer term follow up is needed to determine the extent to which individuals cope with their grief over time. Further research in this area should include a measure of CG. In a meta-analysis of 35 bereavement interventions by Allumbaugh and Hoyt [9] the authors reported an effect size for bereavement interventions of 0.43. They suggested that this relatively small effect may be due to a general ineffectiveness of grief counselling; to the low statistical power of many of the studies, or to one or more intervening variables that masked real effects of the interventions. They then examined 12 of these potential moderator variables (e.g. level of practitioner training and treatment modality, as well as client characteristics such as age, gender, time since loss, and relationship to the deceased). They concluded that more highly trained practitioners produced a better result (particularly when compared to non-professional therapists) and individual therapy produced better results than group treatment. However, these two variables were confounded because studies using individual treatment also tended to use professionally trained therapists rather than paraprofessional volunteers. Kato and Mann [215] used strict selection criteria to review general bereavement intervention studies that required random assignment to treatment and control groups, similar recruitment procedures for both groups, and initiations of the intervention after the death had occurred. They reviewed 13 articles, breaking the sample into studies that used individual, family or group intervention. They found that three of the four studies using individual therapy interventions produced only slight changes in physical health, and one found improvement in stress reactions of the participants. They also concluded that one family therapy study and six of the eight group studies reviewed found no beneficial effects of the intervention. Overall effect sizes of .052, .273 and .095 were reported for the reduction of depressive symptoms, somatic symptoms and all other psychological symptoms respectively. They concluded that psychological interventions for normal bereavement are not effective interventions (p.292). Schut et al [152] evaluated seven studies that focused on bereaved persons who were defined as being at high risk for developing bereavement related problems (secondary prevention). These included populations that had experienced the sudden, traumatic death of a loved one; those that were in a highrisk category (e.g. bereaved parents), and those who showed high levels of symptomatic distress on preintervention measures or on clinical assessment. They concluded that although there is more evidence of intervention efficacy for this population, the effects are still quite modest in comparison to traditional psychotherapy outcome studies. They also emphasised the importance of doing gender specific analyses because several of the studies showed differential effectiveness of the interventions for men and women. Importantly, they found that studies that specifically screened for high levels of distress (rather than simply selecting on the basis of membership in a high-risk category, such as bereaved mothers) tended to show better results for the intervention.

99

Schut et al (2000) also reviewed seven intervention studies for people with complicated grief (i.e. were already suffering from clinical levels of depression, anxiety and other bereavement-induced disorders) at the time of entry into the study (tertiary interventions). These participants tended to be self-referred for help (as opposed to being recruited to participate) and the interventions were typically delivered a longer time after the death. Despite some methodological limitations they found this level of intervention was generally successful, as indicated by reductions in levels of psychiatric symptoms and grief-related distress when compared with control participants.

Allumbaugh and Hoyt [2] found that the more effective interventions included a greater number of sessions and began closer to the time of the death. However, the mean time since the death across all studies reviewed was two years, suggesting that earlier after the death may have been a relatively long time post-death. Kato and Mann [215] in their review of intervention studies did not find significant effects for time since the loss. In addition, Neimeyer (2000) found that interventions the occurred sooner after the death had significantly smaller effect sizes. Schut et al (2000) reached a similar conclusion that interventions offered too soon in the mourning process may be less effective, or even counter productive. There are similar contradictory results around referral patterns. Allumbaum et al [2] found that in general bereavement intervention studies using clients who were self-identified and specifically seeking help had much larger effect sizes than studies where participants were recruited by the investigators. However, Schut et al found that self-referral in interventions in normal bereavement was less effective, and that intervention programs using an active, outreaching approach are much more likely to have no effect or negative effects than programs in which one waits for the bereaved person to initiate contact. No studies were found that determined the best health professional or service for the delivery of interventions in complicated grief. Only one study (the RCT by Shear and colleagues) assigned participants by the manner of the death and few of the studies assigned participants by the time since the death. All these factors are considered to be crucial to the effective running of bereavement support groups [9]. What screening and/or assessment tools exist and what evidence is there to support their use? A number of instruments have been developed and tested in an effort to measure grief responses and identify those who may be at risk for a more complicated grief response. Overall, the instruments demonstrate good estimates of reliability and validity. They range in length with some being brief and simple to use and others lengthy and potentially more burdensome. The extent to which the instruments are able to predict complicated grief responses has not been well documented given the cross-sectional nature of the study designs.

100

Does the nature of the death (i.e. expected or unexpected) affect the risk of complicated grief? and is the risk of complicated bereavement different for those where the death is expected versus unexpected? Our review relating to situational factors associated with the death provides some evidence that survivors of suicide have an increased risk of complicated grief. This supports the notion that the unique features of traumatic death, when present in suicide or in any other traumatic loss account for much of the variance in bereavement outcome in comparison to natural causes of death. The only studies on complicated grief identified for children and adolescents focused on children exposed to trauma. Studies relating to circumstances surrounding the death provide some evidence that complicated grief is an independent risk factor for suicidal ideation. There were a number of limitations in these studies, and the authors call for longitudinal data to determine whether CG and depression are preludes to suicidal ideation. Further investigation of this phenomenon is warranted. Of note in the studies on suicidal ideation was the consistent use of the Inventory of Complicated Grief in assessment of this group [97, 98, 217]. Other situational factors around the death identified in this review associated with complicated grief include: the time from diagnosis to death [119], perceptions of the death being more violent and lack of preparedness for the death [118]: a pattern of high distress pre-death [39] and persistent feelings of being stunned or shocked by the death. The question of preparedness for death and the degree of trauma and suffering associated with the patients death may be pertinent issues when examining family members that may be at greater risk for a complicated grief response. Preparation for the patients death and a sense that the death was peaceful and not distressing may be factors associated with a persons bereavement response. There is insufficient information on the phenomenology, clinical symptoms, clinical needs, and risk factors associated with unexpected and traumatic death together with the added burden of direct traumatisation. The dual burden of loss by traumatic means on top of direct traumatisation is evidenced in disasters generally, and the aftermath of terrorist attacks especially, yet the combined consequences have been understudied [218]. No empirical studies have been conducted examining mediators and moderators of responses to bereavement from traumatic means., although some studies have suggested that experiencing past trauma or previous loss may complicate or prolong the bereavement process [29, 219, 220]. What are the protective and risk factors that predispose an individual to complicated grief? The framework developed by Stroebe and Shut provides a succinct categorisation for the consideration of risk factors [217]. These are determined as situational factors related to the death; person factors such as

101

gender and characteristics prior to the death; and interpersonal factors such as the availability of social and emotional support from family and friends.

Situational factors related to the death include: place of death (eg hospital), the time from diagnosis to death [119]; perceptions of the death being more violent and lack of preparedness for the death [118]; a pattern of high distress pre-death [39]; and persistent feelings of being stunned or shocked by the death [82]. Personal factors include gender (higher levels of depression and anxiety and stress in men when their child died in hospital) [119]; having an insecure attachment style [113]; excessive dependency, both as dependency on the spouse and as a more general personality traits [112]; interpreting grief reactions as indicating mental insanity, inadequate adaptation or personal incompetence [83]; assigning negative meanings to grief reactions [83] and cognitive and emotional upheaval surrounding the death of a healthy spouse [3]. Interpersonal factors such as perceived lack of social support and poor coping skills were identified by bereavement professionals in the US [116]. Although it is necessary to evaluate the familial connection and the apparent degree of intimacy inferred from the description individuals give of the nature of the relationship to the deceased, it is also critical to evaluate the meaning and implication of the loss for the individual [218]. Only one study identified protective factors citing an association between resilience and pre-loss acceptance of death and belief in a just world in and lower levels of distress [3]. Limitations of these findings include potential selection bias, retrospective or selfreporting and difficulty with generalisability. Few studies of sufficient rigour were undertaken to examine inter-personal factors such as social and emotional support from family and friends. Some theoretical and descriptive work has been undertaken to identify factors that might predict those family members who may be more at risk of developing complex grief reactions, such as functioning, quality of the patients death, family care satisfaction during the palliative phase of illness and pre-morbid state of family members health [163-165]. The limitations in these studies include small sample sizes, the time that measurements were undertaken, inconsistency in the use of measures and the use of self-report data. It has been recommended that the duration of six months of the specified symptomatic distress (not from the time of the death) be used in assessing complicated grief and that the symptoms be associated with significant functional impairment [30].

102

When might heightened risk be determined (i.e. is there an optimal time to intervene and how often is intervention required? What are the potential timelines and timings?) Schut et al. [152] reviewed and critically assessed grief and bereavement intervention efficacy studies at three levels: primary, secondary and tertiary. Although the evidence is inconclusive, they reported that the timing of the intervention appears to play a role in efficacy for three reasons: Early intervention may disrupt the natural course of grieving, as emotional, social and practical consequences of the loss still need to take their natural course Interventions could interfere with support networks triggering friends and family to withdraw Bereaved people may be prevented from finding their own solutions

As the CG diagnostic criteria are predicated upon the death of a significant other as a prerequisite for diagnosis, studies have attempted to evaluate CG using time limits to define the on-set of symptoms. Periods of two, four, six, and eight weeks as well as six, 12 and 14 months have been used. Jordan and Neimeyer (2003) postulate that there may be a critical window of time, neither too soon nor too long after a death, when mourners are most responsive to, and able to use, formal support services. One possibility is that services may be most effective when delivered in a 6-18 month period following the death. This may be the time when complicated grief is both diagnosable and prognostic of later difficulties, but before problematic patterns of adjustment have become entrenched (p. 774). Further empirical work is needed to test this postulation. However, this review supports the conclusion that the more complicated the grief process appears to be or becomes, the better the chances of interventions leading to positive results and that tertiary preventive interventions take place at longer durations from the death [152, p. 731]. How do the questions posed above apply to or accommodate the grief experiences of children and adolescents? No intervention studies have been undertaken with children or adolescents to address CG. Only one study provided information regarding the effect of an intervention on symptoms of CG. The paper by Saltzman et al. [106] reports findings from an intervention study with middle school children who had been exposed to severe trauma and/or traumatic loss. The intervention used was a trauma- and grieffocused psychotherapy protocol. The study was considered to be of poor methodological quality due to the non use of a control group, a small sample size, the wide variation in the type of trauma and traumatic loss experienced by the children, and potential confounding of outcomes. How do the findings apply to or have implications for minority groups such as Indigenous people, culturally and linguistically divers groups and isolated families (eg remote

103

farmers/graziers) and identify any studies that deal with minority groups rather than the average/ mainstream experience)? No studies were identified in this review that specifically addressed complicated grief in Indigenous populations. The bulk of the research material identified focussed on intergenerational grief, historical grief, or grief associated with the stolen generation. Recommendations The challenges of undertaking research to investigate complicated grief have been well documented and include inconsistent use of definitions, instruments, cross-sectional designs, heterogeneous samples, high attrition, demographic differences between cases and controls, differences in length of time since death; differences in types of death experienced, and use of recruitment techniques that may contribute to biases in sample characteristics. Notwithstanding these difficulties, this systematic review has confirmed the need for targeted research to address the gaps in knowledge that exist in the area of complicated grief. Without systematic and trustworthy investigations, health professionals and service providers endeavour to provide interventions and services based on anecdotal experiences and trial-and-error approaches. Therefore, the following recommendations are offered: Information

Recommendation 1:

It is recommended that any communication (written or web-based) from relevant areas the Commonwealth Department of Health & Ageing that refers to Complicated Grief (CG) use the most current definition as outlined in this report and be consistent in use of the term.

Professional Development

Recommendation 2:

It is recommended that training be provided to health professionals involved in the care of the bereaved (e.g. GPs, psychologists, psychiatrists, counsellors, community health workers etc.) regarding accepted criteria for diagnosing CG. Such training should be included in under-graduate and post-graduate courses.

Clinical Practice Implications

Recommendation 3:

Clinicians/counselors should be proactive in screening people for CG if they have experienced a traumatic and/or violent death because CG appears to be a predictor for suicidal ideation in these populations.

104

Recommendation 4:

It is recommended that clinicians/counsellors assessing individuals for CG use the diagnostic criterion for complicated grief as specified by Prigerson and colleagues in the Inventory of Complicated Grief-Revised (2001).

Recommendation 5:

It is recommended that the Inventory of Complicated Grief-Revised be used to screen individuals for possible CG if they present with persistent (beyond six months post-death) and severe symptoms (marked intensity or frequency, such as several times daily)

Recommendation 6:

It is recommended that clinicians be aware of the distinction between CG and the DSM-IV disorders of MDD, PTSD and generalised anxiety.

Research
Further research is required to:

Recommendation 7:

Evaluate the reliability, validity, sensitivity, specificity, and diagnostic efficiency of criteria proposed for CG.

Recommendation 8: Recommendation 9:

Assess CG as a mental health outcome independent of the nature of the death. Examine the situational factors (e.g. sudden, expected, traumatic, non-traumatic) associated with death in the Australian context of CG

Recommendation 10: Examine risk factors such as the role of attachment styles and cognitive
functioning, using prospective, longitudinal designs and objective measures of CG.

Recommendation 11: Assess the effect of CG on outcome measures using large, non-clinical samples,
prospective controlled designs.

Recommendation 12: Clarify the association between CG and adverse health outcomes and to identify
the specific psychological and biological pathways through which CG is expressed in poor health.

Recommendation 13: Identify risk factors for CG associated with:

perinatal death, infant death child and adolescent death

105

death of an adult child

Recommendation 14: Identify most appropriate interventions to respond to CG in the above

populations

Recommendation 15: Identify empirical evidence related to children and adolescents and in particular
the grief experiences of children and adolescents, risk factors that may predispose them to complicated grief in later life, the criteria used to define CG in the context of childhood and adolescent experiences, instruments most appropriate for measurement of CG in child and adolescent populations the extent to which CG is distinct from traumatic experiences interventions most appropriate to the grief experiences of children and adolescents

Recommendation 16: Identify the elements of a palliative approach to care that may be protective
factors associated with CG

Recommendation 17: To examine preparation for death and perceptions of the quality of death as
factors associated with CG

Recommendation 18: Examine the risk for CG within the Australian health care setting focused on
individuals who have experienced the death of a loved one due to HIV/AIDS Recommendation 19: Examine the effectiveness of a support group intervention for individuals with CG following the death of a loved one due to HIV/AIDS. Recommendation 20: Better understand the needs of older adults who are not in a spousal relationship

Recommendation 21: Examine the experience of CG in the context of mental health populations. Recommendation 22: Identify risk factors specifically related to CG in the Indigenous populations Recommendation 23: Examine links between assessment, intervention and outcomes that are targeted
to well-defined patient populations at well-defined phases of bereavement.

Recommendation 24: Demonstrate the efficacy of pharmacotherapy in complicated grief

106

Recommendation 25: Determine the role of the level of practitioner training, as well as client
characteristics such as age, gender, time since loss, and relationship to the deceased in individual or group counselling.

Recommendation 26: Compare client outcomes from individual counselling and group counselling. Recommendation 27: Investigate gender differences in assessing the effectiveness of individual or
group counselling To inform future research into complicated grief it is recommended that a systematic review of the literature in grief and bereavement be undertaken in the following areas: children and adolescents violent and traumatic death e.g. murder, suicide, homicide, genocide, natural disasters and acts of terrorism

107

ACKNOWLEDGEMENTS
We acknowledge the guidance and support of the project staff and reference group at the Australian Government Department of Health and Ageing. We would also like to acknowledge and thank the members of our Expert Panel for their comment and review. Members Ms. Julie Dunsmore Mr. Chris Hall Mr. Mal McKissock Consumer Professor Richard Bryant Ms. Kate Sullivan Ms. Trudy Hansen Ms. Jane Mowll Affiliation President of the National Association of Loss and Grief, NSW Director, Centre for Grief Education, Victoria Co-Director - Bereavement CARE Centre, NSW Co-Director, Clinical Services National Centre for Childhood Grief Dr. Anne Atkinson School of Psychology, University of New South Wales Consultant on Indigenous affairs NALAG (Dubbo, NSW) Senior Forensic Counsellor Dept. of Forensic Medicine, Western Sydney Area Health Service, NSW

We also acknowledge our research and administrative staff involved in the preparation of this report Project Officer: Ms. Anna Davies Administrative Assistant: Ms. Helen Morris Dr. Georgia Halkett NBCF Post Doctoral Research Fellow

108

REFERENCES

1.

2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

14. 15. 16. 17. 18. 19. 20.

21. 22. 23.

24.

25. 26.

Christ, G., Bonanno, G.A., Malkinson, R., and Rubin, S.S., Bereavement experiences after the death of a child, in When children die: Improving palliative and end-of-life care for children and their families, M. Field and R. Berhman, Editors. 2003, National Academy Press: Washington, DC. p. 553-579. Allumbaugh, D.L. and Hoyt, W.T., Effectiveness of grief therapy: A meta-analysis. Journal of Counseling Psychology, 1999. 46(3): p. 370-380. Bonanno, G.A., Wortman, C.B., and Nesse, R.M., Prospective patterns of resilience and maladjustment during widowhood. Psychology and Aging, 2004. 19(2): p. 260-271. Shear, K. and Shair, H., Attachment, loss, and complicated grief. Developmental Psychobiology, 2005. 47(3): p. 253-267. Byrne, G.J. and Raphael, B., A longitudinal study of bereavement phenomena in recently widowed elderly men. Psychological Medicine, 1994. 24(2): p. 411-421. Prigerson, H. and Jacobs, S., Caring for bereaved patients: "All the doctors just suddenly go". The Journal of the American Medical Association, 2001. 286(11): p. 1369-1376. Middleton, W., Burnett, P., Raphael, B., and Martinek, N., The bereavement response: A cluster analysis. The British Journal of Psychiatry, 1996. 169(2): p. 167-171. Clark, S.E., Bereavement after suicide: How far have we come and where do we go from here? Crisis, 2001. 22(3): p. 102-108. McKissock, D. and McKissock, M., Bereavement counselling: Guidelines for practitioners. 2003, Terrigal: The Bereavement C.A.R.E. Centre. NHMRC, A guide to the development, implementation and evaluation of clinical practice guidelines. 1999. Campbell Collaboration, Campbell collaboration guidelines. 2002. Cochrane Collaboration, Cochrane reviewers' handbook 4.2.2. 2002. Egger, M., Juni, P., Bartlett, C., Holenstein, F., and Sterne, J., How important are comprehensive literature searches and assessment of trial quality in systematic reviews? Empirical study. Health Technology Assessments, 2003. 7(1): p. 1-76. Abernethy, A.P., Currow, D., and Butler, T. Searching electronic bibliographic databases for palliative care articles. in 5th Conference and Scientific Meeting. 2002. Townsville, Queensland. Prigerson, H. and Maciejewski, P., A call for sound empirical testing and evaluation of criteria for complicated grief proposed for DSM-V. Omega: Journal of Death & Dying, In press. Neimeyer, R.A., Defining the new abnormal: Scientific and social construction of complicated grief. Omega: Journal of Death & Dying, In press. Walter, T., What is complicated grief? A social constructionist perspective. Omega: Journal of Death & Dying, In press. Parkes, C.M. and Weiss, R.S., Recovery from bereavement. 1983, New York: Basic Books. Raphael, B., The anatomy of bereavement. 1983, New York: Basic Books. Brent, D.A., Perper, J., Moritz, G., Allman, C., Schweers, J., Roth, C., Balach, L., Canobbio, B.S., and Liotus, L., Bereavement or depression? The impact of the loss of a friend to suicide. Journal of the American Academy of Child and Adolescent Psychiatry, 1993. 32(6): p. 1189-97. Kim, K. and Jacobs, S., Pathologic grief and its relationship to other psychiatric disorders. Journal of Affective Disorders, 1991. 21(4): p. 257-263. Horowitz, M.J., Siegel, B., Holen, A., Bonanno, G.A., Milbrath, C., and Stinson, C.H., Diagnostic criteria for complicated grief disorder. American Journal of Psychiatry, 1997. 154(7): p. 904-910. Prigerson, H., Bierhals, A.J., Kasl, S.V., Reynolds, C.F., Shear, K., Newsom, J., T., and Jacobs, S., Complicated grief as a disorder distinct from bereavement-related depression and anxiety: A replication study. American Journal of Psychiatry, 1996. 153(11): p. 1484-1486. Prigerson, H., Frank, E., Kasl, S.V., Reynolds, C.F., Anderson, B., Zubenko, G.S., Houck, P.R., George, C.J., and Kupfer, D.J., Complicated grief and bereavement-related depression as distinct disorders: Preliminary empirical validation in elderly bereaved spouses. American Journal of Psychiatry, 1995. 152(1): p. 22-30. Marwit, S.J., DSM-III-R, grief reactions, and a call for revision. Professional Psychology: Research and Practice, 1991. 22(1): p. 75-79. Marwit, S.J., Reliability of diagnosing complicated grief: A preliminary investigation. Journal of Consulting and Clinical Psychology, 1996. 64(3): p. 563-568.

109

27.

28.

29.

30. 31. 32. 33.

34. 35. 36. 37. 38.

39.

40. 41. 42. 43. 44. 45. 46. 47. 48. 49.

50.

51. 52.

53.

Prigerson, H., Maciejewski, P., Reynolds, C.F., Bierhals, A.J., Newsom, J., T., Fasiczka, A., Frank, E., Doman, J., and Miller, M., Inventory of complicated grief: A scale to measure maladaptive symptoms of loss. Psychiatry Research, 1995. 59(1-2): p. 65-79. Prigerson, H. and Jacobs, S., Traumatic grief as a distinct disorder: A rationale, consensus criteria, and a preliminary empirical test, in Handbook of Bereavement: Consequences, Coping, and Care, M.S. Stroebe, et al., Editors. 2001, American Psychological Association Press: Washington, DC. p. 613-645. Raphael, B. and Martinek, N., Assessing traumatic bereavement and Posttraumatic Stress Disorder, in Assessing Psychological Trauma and PTSD, J.P. Wilson and T.M. Keane, Editors. 1997, Guilford Press: New York. p. 373-395. Prigerson, H. and Vanderwerker, L.C., Final remarks. Omega: Journal of Death & Dying, In press. Freud, S., Mourning and melancholia, in The Standard Edition of the Complete Psychological Works of Sigmund Freud, J. Strachey, Editor. 1961, Basic Books: New York. p. 243-258. Stroebe, M., Bereavement research and theory: Retrospective and prospective. American Behavioral Scientist, 2001. 44(5): p. 854-865. Stroebe, M., Schut, H., and Finkenauer, C., The traumatization of grief? A conceptual framework for understanding the trauma-bereavement interface. The Israel Journal of Psychiatry and Related Sciences, 2001. 38(3/4): p. 185-201. Bonanno, G.A. and Kaltman, S., The varieties of grief experience. Clinical Psychology Review, 2001. 21(5): p. 705-734. Wortman, C.B. and Silver, R.C., The myths of coping with loss. Journal of Consulting and Clinical Psychology, 1989. 57(3): p. 349-357. Stroebe, M. and Stroebe, W., Does 'grief work' work? Journal of Consulting and Clinical Psychology, 1991. 59(3): p. 479-482. Klass, D., Silverman, P.R., and Nickman, S.L., eds. Continuing bonds: New understandings of grief. 1996, Taylor & Francis: Washington, DC. Wortman, C.B. and Silver, R.C., The myths of coping with loss revisited, in Handbook of Bereavement Research: Consequences, Coping, and Care, M.S. Stroebe, et al., Editors. 2001, American Psychological Association Press: Washington, DC. p. 405-431. Boerner, K., Wortman, C.B., and Bonanno, G.A., Resilient or at risk? A 4-year study of older adults who initially showed high or low distress following conjugal loss. The Journals of Gerontology, 2005. 60B(2): p. 67-73. Bowlby, J., Attachment and loss, Vol 1. Attachment. 1969, New York: Basic Books. Bowlby, J., Attachment and loss, Vol 2: Separation, anxiety and anger. 1973, London: Hogarth Press. Bowlby, J., Attachment and loss, Vol 3: Loss, sadness and depression. 1980, New York: Basic Books. Field, N.P. and Friedrichs, M., Continuing bonds in coping with the death of a husband. Death Studies, 2004. 28(7): p. 597-620. Stroebe, M. and Schut, H., Complicated grief: A conceptual analysis of the field. Omega: Journal of Death & Dying, In press. Neimeyer, R.A., ed. Meaning reconstruction and the experience of loss. 2001, American Psychological Association Press: Washington, DC. Janoff-Bulman, R. and Berger, A.R., The other side of trauma, in Loss and Trauma, J.H. Harvey and E.D. Miller, Editors. 2000, Brunner Mazel: Philadelphia. Genevro, J.L., Marshall, M.P.H., and Miller, T., Report on bereavement and grief research. 2003, Centre for the Advancement of Health: Washington. Neimeyer, R.A., Prigerson, H.G., and Davies, B., Mourning and meaning. American Behavioral Scientist, 2002. 46(2): p. 235-251. Folkman, S., Revised coping theory and the process of bereavement, in Handbook of Bereavement: Consequences, Coping, and Care, M.S. Stroebe, et al., Editors. 2001, American Psychological Association Press: Washington, DC. p. 563-584. Bonanno, G.A., Grief and emotion: A social-functional perspective, in Handbook of Bereavement: Consequences, Coping, and Care, M.S. Stroebe, et al., Editors. 2001, American Psychological Association Press: Washington, DC. p. 493-515. Stroebe, M. and Schut, H., The dual process model of coping with bereavement: Rationale and description. Death Studies, 1999. 23(3): p. 29-31. Goodkin, K., Lee, D., Molina, R., Zheng, W., Frasca, A., O'Mellan, S., Asthana, D., Shapshak, P., and Khamis, I., Complicated bereavement: Disease state or state of being? Omega: Journal of Death & Dying, In press. Hogan, N.S., Worden, J.W., and Schmidt, L.A., An empirical study of the proposed complicated grief disorder criteria. Omega: Journal of Death & Dying, 2004. 48(3): p. 263-277.

110

54. 55. 56. 57. 58. 59. 60. 61.

62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72.

73. 74. 75. 76. 77.

78.

79. 80.

81.

Parkes, C.M., Complicated grief: Introduction to a symposium. Omega: Journal of Death & Dying, In press. Prigerson, H., Bierhals, A.J., and Zonarich, D.L., Complicated grief: Nosology, treatment, and outcomes. Directions in Psychiatry, 1996. 16(8): p. 1-11. Parkes, C.M., Guest editor's conclusions. Omega: Journal of Death & Dying, In press. Lichtenthal, W.G., Cruess, D.G., and Prigerson, H.G., A case for establishing complicated grief as a distinct mental disorder in DSM-V. Clinical Psychology Review, 2004. 24(6): p. 637-662. Stroebe, M., van Son, M., Stroebe, W., Kleber, R., Schut, H., and van den Bout, J., On the classification and diagnosis of pathological grief. Clinical Psychology Review, 2000. 20(1): p. 57-75. Walle, A., Native people and the DSM IV-TR: Expanding diagnostic criteria to reflect minority trauma. Journal of Ethnicity in Substance Abuse, 2004. 3(3): p. 49-65. Horowitz, M.J., Meditating on complicated grief disorder as a diagnosis. Omega: Journal of Death & Dying, In press. Neimeyer, R.A. and Hogan, N.S., Quantitative or qualitative? Measurement issues in the study of bereavement., in Handbook of Bereavement Research: Consequences, Coping, and Care, M. Stroebe, et al., Editors. 2001, American Psychological Association Press: Washington, DC. p. 89-118. Faschingbauer, T.R., The Texas Revised Inventory of Grief manual. 1981, Houston: Honeycomb. Hogan, N.S., Greenfield, D.B., and Schmidt, L.A., Development and validation of the Hogan Grief Reaction Checklist. Death Studies, 2001. 25: p. 1-31. Jordan, J.R., Baker, J., Matteis, M., Rosenthal, S., and Ware, E.S., The Grief Evaluation Measure (GEM): An initial validation study. Death Studies, 2005. 29(4): p. 301-332. Lev, E., Monro, B.H., and McCorkle, R., A shortened version of an instrument measuring bereavement. International Journal of Nursing Studies, 1993. 30(3): p. 213-226. Burnett, P., Middleton, W., Raphael, B., and Martinek, N., Measuring core bereavement phenomena. Psychological Medicine, 1997. 27(1): p. 49-57. Parkes, C.M., Bereavement, in Oxford Textbook of Palliative Medicine, D. Doyle, G.C.W. Hanks, and N. MacDonald, Editors. 1993, Oxford University Press: Oxford. p. 663-678. Barrett, T.W. and Scott, T.B., Development of the grief experience questionnaire. Suicide & LifeThreatening Behavior, 1989. 19(2): p. 201-215. Toedter, L.J., Laser, J.N., and Alhadeff, J.M., The Perinatal Grief Scale: Development and initial validation. American Journal of Orthopsychiatry, 1988. 58(3): p. 435-449. Theut, S.K., Pederson, F.A., Zaslow, M.J.U., Cain, R.L., Rabinovich, B.A., and Morihisa, J.M., Perinatal loss and parental bereavement. American Journal of Psychiatry, 1989. 146(5): p. 635-639. Faschingbauer, T.R., Devaul, R.A., and Zisook, S., Development of the Texas Inventory of Grief. American Journal of Psychiatry, 1977. 134(6): p. 213-226. Hansson, R., Carpenter, B.N., and Fairchild, S.K., Measurement issues in bereavement, in Handbook of Bereavement: Theory, Research, and Intervention, M.S. Stroebe, W. Stroebe, and R.O. Hansson, Editors. 1993, Cambridge University Press: New York. p. 62-74. Nunnally, J.A. and Bernstein, I., Psychometric Theory. 1994, New York: McGraw-Hill. Zisook, S., DeVaul, R.A., and Click, M.A., Measuring symptoms of grief and bereavement. American Journal of Psychiatry, 1982. 139(12): p. 1590-1593. Horowitz, M.J., Wilner, N., and Alvarez, W., Impact of Events Scale: A measure of subjective stress. Psychosomatic Medicine, 1979. 41: p. 209-218. Ware, J.E., SF-36 physical and mental health summary scales: A user's manual. 1994, Boston: The Health Institute, New England Medical Center. Kraus, D.R., Seligman, D., and Jordan, J.R., Validation of a behavioural health treatment outcome and assessment tool designed for practice research networks and other naturalistic settings: The treatment outcome package. Journal of Clinical Psychology, In Press. Kristjanson, L.J., Cousins, K., Smith, J., and Lewin, G., Evaluation of the Bereavement Risk Index (BRI): A community hospice care protocol. International Journal of Palliative Nursing, 2005. 11(12): p. 610-618. Potvin, I., Lasker, J., and Toedter, L., Measuring grief: A short version of the Perinatal Grief Scale. Journal of Psychopathology and Behavioral Assessment, 1989. 11(1): p. 29-45. Zeanah, C.H., Daily, J., Rosenblatt, M.J., and Saller, N., Do women grieve following termination of pregnancy for tetal anomalies? A controlled investigation. Obstetrics and Gynecology, 1993. 82(2): p. 270-275. Faschingbauer, T.R., Zisook, S., and De Vaul, R., The Texas Revised Inventory of Grief, in Biopsychosocial aspects of bereavement, S. Zisook, Editor. 1987, American Psychiatric Press: Washington, DC. p. 111-124.

111

82.

83.

84.

85.

86. 87. 88.

89.

90. 91. 92.

93.

94. 95.

96. 97.

98. 99. 100.

101. 102. 103. 104. 105.

Prigerson, H., Shear, M.K., Newsom, J.T., Frank, E., Reynolds, C.F., Maciejewski, P.K., Houck, P.R., Bierhals, A.J., and Kupfer, D.J., Anxiety among widowed elders: Is it distinct from depression and grief? Anxiety, 1996. 2(1): p. 1-12. Boelen, P.A., van den Bout, J., and van den Hout, M.A., The role of negative interpretations of grief reactions in emotional problems after bereavement. Journal of Behavior Therapy and Experimental Psychiatry, 2003b. 34(3-4): p. 225-238. Boelen, P.A. and van den Bout, J., Complicated grief, depression, and anxiety as distinct postloss syndromes: A confirmatory factor analysis study. American Journal of Psychiatry, 2005. 162(11): p. 2175-2177. Ogrodniczuk, J.S., Piper, W.E., Joyce, A.S., Weideman, R., McCallum, M., Azim, H.F., and Rosie, J.S., Differentiating symptoms of complicated grief and depression among psychiatric outpatients. Canadian Journal of Psychiatry, 2003. 48(2): p. 87-93. Rynearson, E.K. and McCreery, J.M., Bereavement after homicide: A synergism of trauma and loss. American Journal of Psychiatry, 1993. 150(2): p. 258-261. Rando, T., The increasing prevalence of complicated mourning: The onslaught is just beginning. Omega: Journal of Death & Dying, 1992. 26(1): p. 43-59. Rando, T., Complications in mourning traumatic death, in Living with Grief after Sudden Loss: Suicide, Homicide, Accident, Heartattack, Stroke, K. Doka, Editor. 1996, Hospice Foundation of America: Washington, DC. p. 139-159. Stroebe, W., Stroebe, M., and Schut, H., Risk factors in bereavement outcome: A methodological and empirical review., in Handbook of Bereavement Research, R.O. Hansson, et al., Editors. 2001, American Psychological Association: Washington, DC. p. 349-373. Matthews, L.T. and Marwit, S.J., Examining the assumptive world views of parents bereaved by accident, murder, and illness. Omega, 2004. 48(2): p. 115-136. Neimeyer, R.A., Searching for the meaning of meaning: Grief therapy and the process of reconstruction. Death Studies, 2000. 24(6): p. 541-558. Rynearson, E.K. and Sinnema, C.S., Supportive group therapy for bereavement after homicide, in Group Treatments for Post-Traumatic Stress Disorder, D. Blake and B. Young, Editors. 1999. p. 137147. Prigerson, H., Shear, M.K., Jacobs, S., Reynolds, C.F., Maciejewski, P.K., Davidson, J.R.T., Rosenheck, R., Pilkonis, P.A., Wortman, C.B., Williams, J.B.W., Widiger, T.A., Frank, E., Kupfer, D.J., and Zisook, S., Consensus criteria for traumatic grief: A preliminary empirical test. British Journal of Psychiatry, 1999. 174: p. 67-73. Raphael, B., The interaction of trauma and grief, in Psychological Trauma: A Developmental Approach, D. Black, et al., Editors. 1997, Gaskell/Royal College of Psychiatrists: London. p. 31-43. Rando, T., On the experience of traumatic stress in anticipatory and postdeath mourning, in Clinical Dimensions of Anticipatory Mourning, T.A. Rando, Editor. 2000, Research Press: Champaign. p. 155221. Horowitz, M.J., Bonanno, G.A., and Holen, A., Pathological grief: Diagnosis and explanation. Psychosomatic Medicine, 1993. 55(3): p. 260-273. Prigerson, H., Bridge, J., Maciejewski, P., Beery, L.C., Rosenheck, R., Jacobs, S., Bierhals, A.J., Kupfer, D.J., and Brent, D.A., Influence of traumatic grief on suicidal ideation among young adults. American Journal of Psychiatry, 1999. 156(12): p. 1994-1995. Mitchell, A., Kim, Y., Prigerson, H.G., and Mortimer, M., Complicated grief and suicidal ideation in adult survivors of suicide. Suicide & Life-Threatening Behavior, 2005. 35(5): p. 498-506. Dyregrov, K., Nordanger, D., and Dyregrov, A., Predictors of psychosocial distress after suicide, SIDS and accidents. Death Studies, 2003. 27(2): p. 143-165. Melhem, N., Rosales, C., Karageorge, J., Reynolds, C.F., Frank, E., and Shear, K., Comorbidity of axis 1 disorders in patients with traumatic grief. The Journal of Clinical Psychiatry, 2001. 62(11): p. 884887. Pivar, I.L. and Field, N.P., Unresolved grief in combat veterans with PTSD. Journal of Anxiety Disorders, 2004. 18(6): p. 745-755. Melhem, N., Day, N., Shear, K., Day, R., Reynolds, C.F., and Brent, D.A., Predictors of complicated grief among adolescents exposed to a peer's suicide. Journal of Loss & Trauma, 2004. 9(1): p. 21-34. Melhem, N., Traumatic grief among adolescents exposed to their peer's suicide, in Graduate School of Public Health. 2001, University of Pittsburgh. Ginzburg, K., Geron, Y., and Solomon, Z., Patterns of complicated grief among bereaved parents. Omega: Journal of Death & Dying, 2002. 45(2): p. 119-132. Mitchell, A., Kim, Y., Prigerson, H., and Mortimer-Stephens, M., Complicated grief in survivors of suicide. Crisis, 2004. 25(1): p. 12-18.

112

106.

107. 108. 109. 110.

111. 112.

113.

114. 115. 116. 117. 118.

119.

120.

121.

122.

123. 124.

125. 126.

127.

128.

Saltzman, W.R., Pynoos, R.S., Layne, C.M., Steinberg, A.M., and Aisenberg, E., Trauma and grieffocused intervention for adolescents exposed to community violence: Results of a school-based screening and group treatment protocol. Group Dynamics: Theory, Research, and Practice, 2001. 5(4): p. 291-303. Rose, S., Bisson, J., and Wessely, S., Psychological debriefing for preventing post-traumatic stress disorder (PTSD). Cochrane Review, The Cochrane Library, 3, Oxford: Update Software, 2001. Litz, B., Gray, M., Bryant, R., and Adler, A., Early intervention for trauma: Current status and future directions. Clinical Psychology: Science and Practice, 2002. 9(2): p. 112-134. Harvey, M.R., An ecological view of psychological trauma and trauma recovery. Journal of Traumatic Stress, 1996. 9(1): p. 3-23. Silverman, G.K., Johnson, J.G., and Prigerson, H.G., Preliminary explorations of the effects of prior trauma and loss on risk for psychiatric disorders in recently widowed people. The Israel Journal of Psychiatry and Related Sciences, 2001. 38(3-4): p. 202-215. Vanderwerker, L.C., Jacobs, S.C., Murray-Parkes, C., and Prigerson, H.G., An exploration of associations between separation anxiety in childhood and complicated grief in later-life. In press. Bonanno, G.A., Wortman, C.B., Lehman, D.R., Tweed, R.G., Haring, M., Sonnega, J., Carr, D., and Nesse, R.M., Resilience to loss and chronic grief: A prospective study from preloss to 18-months postloss. Journal of Personality and Social Psychology, 2002. 83(5): p. 1150-1164. van Doorn, C., Kasl, S., Beery, L.C., Jacobs, S., and Prigerson, H.G., The influence of marital quality and attachment styles on traumatic grief and depressive symptoms. Journal of Nervous & Mental Disease, 1998. 186(9): p. 566-573. Hardison, H.G., Neimeyer, R.A., and Lichstein, K.L., Insomnia and complicated grief symptoms in bereaved college students. Behavioral Sleep Medicine, 2005. 3(2): p. 99-111. Wiles, R., Jarrett, N., Payne, S., and Field, D., Referrals for bereavement counselling in primary care: A qualitative study. Patient Education & Counseling, 2002. 48(1): p. 79-85. Ellifritt, J., Nelson, K.A., and Walsh, D., Complicated bereavement: A national survey of potential risk factors. American Journal of Hospice and Palliative Care, 2003. 20(2): p. 114-120. Germain, A., Caroff, K., Buysse, D.J., and Shear, K., Sleep quality in complicated grief. Journal of Traumatic Stress, 2005. 18(4): p. 343-346. Barry, L.C., Kasl, S.V., and Prigerson, H.G., Psychiatric disorders among bereaved persons: The role of perceived circumstances of death and preparedness for death. American Journal of Geriatric Psychiatry, 2002. 10(4): p. 447-457. Goodenough, B., Drew, D., Higgins, S., and Trethewie, S., Bereavement outcomes for parents who lose a child to cancer: Are place of death and sex of parent associated with differences in psychological functioning? Psycho-oncology, 2004. 13: p. 779-791. McDermott, O.D., Prigerson, H.G., Reynolds, C.F., Houck, P.R., Dew, M.A., Hall, M., Mazumdar, S., Buysse, D.J., Hoch, C.C., and Kupfer, D.J., Sleep in the wake of complicated grief symptoms: An exploratory study. Biological Psychiatry, 1997. 41(6): p. 710-716. Silverman, G.K., Jacobs, S., Kasl, S.V., Shear, M.K., Maciejewski, P., Noaghiul, F.S., and Prigerson, H.G., Quality of life impairments associated with diagnostic criteria for traumatic grief. Psychological Medicine, 2000. 30(4): p. 857-862. Boelen, P.A., van den Bout, J., and van den Hout, M.A., The role of cognitive variables in psychological functioning after the death of a first degree relative. Behaviour, Research and Therapy, 2003a. 41(10): p. 1123-1136. Monk, T., Houck, P.R., and Shear, K., The daily life of complicated grief patients - what gets missed, what gets added? Death Studies, 2006. 30(1): p. 77-85. Prigerson, H., Bierhals, A.J., Kasl, S.V., Reynolds, C.F., Shear, M.K., Day, N., Beery, L.C., Newsom, J., T., and Jacobs, S., Traumatic grief as a risk factor for mental and physical morbidity. American Journal of Psychiatry, 1997. 154(5): p. 616-623. Ott, C.H., The impact of complicated grief on mental and physical health at various points in the bereavement process. Death Studies, 2003. 27(3): p. 249-272. Latham, A.E. and Prigerson, H.G., Suicidality and bereavement: Complicated grief as psychiatric disorder presenting greatest risk for suicidality. Suicide and Life-Threatening Behavior, 2004. 34(4): p. 350-362. Simon, N.M., Pollack, M.H., Fischmann, D., Perlman, C.A., Muriel, A.C., Moore, C.W., Nierenberg, A.A., and Shear, M.K., Complicated grief and its correlates in patients with bipolar disorder. Journal of Clinical Psychiatry, 2005. 66(9): p. 1105-1110. Finlay, I.G. and Jones, N.K., Unresolved grief in young offenders in prison. British Journal of General Practice, 2000. 50(456): p. 569-570.

113

129.

130. 131. 132. 133. 134.

135. 136.

137.

138.

139. 140. 141. 142.

143. 144.

145. 146. 147. 148. 149. 150. 151. 152.

153. 154.

Cohen, J.A., Goodman, R.F., Brown, E.J., and Mannarino, A.P., Treatment of childhood traumatic grief: Contributing to a newly emerging condition in the wake of community trauma. Harvard Review of Psychiatry, 2004. 12(4): p. 213-216. Cohen, J.A., Mannarino, A.P., and Knudsen, K., Treating childhood traumatic grief: A pilot study. Journal of the American Academy of Child and Adolescent Psychiatry, 2004. 43(10): p. 1225-1233. Janssen, H.J.E.M., Cuisinier, M.C.J., and Hoogduin, K.A.L., A critical review of the concept of pathological grief following pregnancy loss. Omega: Journal of Death & Dying, 1996. 33(1): p. 21-42. Murray, J.A., Caring for families affected by infant death: Evaluating an intervention. Grief Matters: The Australian Journal of Grief and Bereavement, 1998. 1(3): p. 6-11. Murray, J.A., Terry, D.J., Vance, J.C., Battistutta, D., and Connolly, Y., Effects of a program of intervention on parental distress following infant death. Death Studies, 2000. 24(4): p. 275-305. deCinque, N., Monterosso, L., Dadd, G., Sidhu, R., and Lucas, R., Bereavement support for families following the death of a child from cancer: Practice characteristics of Australian and New Zealand paediatric oncology units. Journal of Paediatrics and Child Health, 2004. 40(3): p. 131-135. Kempson, D.A., Effects of intentional touch on complicated grief of bereaved mothers. Omega: Journal of Death & Dying, 2000-2001. 42(4): p. 341-353. Bierhals, A., Frank, E., Prigerson, H.G., Miller, M., Fasiczka, A., and Reynolds, C.F., Gender differences in complicated grief among the elderly. Omega: Journal of Death & Dying, 1995. 32(4): p. 303-317. Bonanno, G.A., Keltner, D., Holen, A., and Horowitz, M.J., When avoiding unpleasant emotions might not be such a bad thing: Verbal-autonomic response dissociation and midlife conjugal bereavement. Journal of Personality and Social Psychology, 1995. 69(5): p. 975-989. Bonanno, G.A., Znoj, H., Siddique, H.I., and Horowitz, M.J., Verbal-autonomic dissociation and adaptation to midlife conjugal loss: A follow-up at 25 months. Cognitive Therapy and Research, 1999. 23(6): p. 605-624. Bonanno, G.A. and Field, N.P., Examining the delayed grief hypothesis across 5 years of bereavement. The American Journal of Behavioral Scientist, 2001. 44(5): p. 798-816. Maercker, A., Bonanno, G.A., Znoj, H.J., and Horowitz, M.J., Prediction of complicated grief by positive and negative themes in narratives. Journal of Clinical Psychology, 1998. 54(8): p. 1117-1136. Field, N.P. and Horowitz, M.J., Applying an empty-chair monologue paradigm to examine unresolved grief. Psychiatry, 1998. 61(4): p. 279-287. Beery, L.C., Prigerson, H.G., Bierhals, A.J., Santucci, L.M., Newsom, J.T., Maciejewski, P.K., Rapp, S.R., Fasiczka, A., and Reynolds, C.F., Traumatic grief, depression and care giving in elderly spouses of the terminally ill. Omega: Journal of Death & Dying, 1997. 35(3): p. 261-279. Brintzen-hofeSzoc, K.M., Smith, E.D., and Zabora, J.R., Screening to predict complicated grief in spouses of cancer patients. A Multidisciplinary Journal of Cancer Care, 1999. 7(5): p. 233-239. Johnson, J.G., Vanderwerker, L.C., Bornstein, R.F., Zhang, B., and Prigerson, H.G., Development and validation of an instrument for the assessment of dependency among bereaved persons. Journal of Psychopathology and Behavioral Assessment, In Press. Raphael, B., Preventive intervention with the recently bereaved. Archives of General Psychiatry, 1977. 34(12): p. 1450-1454. Raphael, B., Mourning and the prevention of melancholia. British Journal of Medical Psychology, 1978. 51(4): p. 303-310. Parkes, C.M., Evaluation of a bereavement service. Journal of Preventive Psychiatry, 1981. 1(2): p. 179-188. Bromberg, M.H. and Higginson, I., Bereavement follow-up: What do palliative support teams actually do? Journal of Palliative Care, 1996. 12(1): p. 12-17. Bouton, B.L., The interdisciplinary bereavement team: Defining and directing appropriate bereavement care. The Hospice Journal, 1996. 11(4): p. 15-24. Parkes, C.M., Bereavement: Studies of grief in adult life. 3rd ed. 1996, London: Routledge. Jordan, J.R. and Neimeyer, R.A., Does grief counseling work? Death Studies, 2003. 27(9): p. 765 - 786. Schut, H., Stroebe, M.S., van den Bout, J., and Terheggen, M., The efficacy of bereavement interventions: Determining who benefits, in Handbook of Bereavement Research: Consequences, Coping, and Care, M.S. Stroebe and R.O. Hansson, Editors. 2001, American Psychological Association Press: Washington, DC. p. 705-737. Currow, D., Grief and loss: The needs of the whole population. Grief Matters: The Australian Journal of Grief and Bereavement, 2004. 7(3): p. 63-66. Parkes, C.M., Risk factors in bereavement: Implications for the prevention and treatment of pathologic grief. Psychiatric Annals, 1990. 20(6): p. 308-313.

114

155. 156.

157.

158.

159.

160.

161.

162.

163. 164. 165.

166. 167.

168. 169.

170. 171.

172. 173. 174.

175. 176. 177. 178.

Gamino, L.A., Sewell, K.W., and Easterling, L.W., Scott and White grief study - phase 2: Toward an adaptive model of grief. Death Studies, 2000. 24(7): p. 633-660. Prigerson, H., Mong, T.H., Reynolds, C.F., Begley, A., Houck, P.R., Bierhals, A.J., and Kupfer, D.J., Lifestyle regularity and activity level as protective factors against bereavement-related depression in late-life. Depression, 1995-1996. 3: p. 297-302. Kissane, D.W., Bloch, S., Dowe, D.L., Snyder, R.D., Onghena, P., McKenzie, D.P., and Wallace, C., The Melbourne family grief study, I: Perceptions of family functioning in bereavement. American Journal of Psychiatry, 1996. 153(5): p. 650-658. Kissane, D.W., Bloch, S., Onghena, P., McKenzie, D.P., Snyder, R.D., and Dowe, D.L., The Melbourne family grief study, II: Psychosocial morbidity and grief in bereaved families. American Journal of Psychiatry, 1996. 153(5): p. 659-666. Kissane, D.W., Bloch, S., McKenzie, M., McDowall, A.C., and Nitzan, R., Family grief therapy: A preliminary account of a new model to promote healthy family functioning during palliative care and bereavement. Psycho-Oncology, 1998. 7(1): p. 14-25. Kissane, D.W., McKenzie, M., McKenzie, D.P., Forbes, A., O'Neill, I., and Bloch, S., Psychosocial morbidity associated with patterns of family functioning in palliative care: Baseline data from the Family Focused Grief Therapy controlled trial. Palliative Medicine, 2003. 17: p. 527-537. Bradley, E., Prigerson, H., Carlson, M.D.A., Cherlin, E., Johnson-Hurzeler, R., and Kasl, S.V., Depression among surviving caregivers: Does length of hospice enrolment matter? American Journal of Psychiatry, 2004. 161(12): p. 2257-2262. Christakis, N.A. and Iwashyna, T.L., The health impact of health care on families: A match cohort study of hospice use by decendents and mortality outcomes in surviving, widowed spouses. Social Science & Medicine, 2003. 57(3): p. 465-475. Ferrell, B.R., Cancer deaths and bereavement outcomes. Home versus hospital. The American Journal of Hospice & Palliative Care, 1985. 2(4): p. 18-24. Kissane, D.W., Bloch, S., and McKenzie, D.P., Family coping and bereavement outcomes. Palliative Medicine, 1997. 11(3): p. 191-201. Kristjanson, L.J., Sloan, J.A., Dudgeon, D., and Adaskin, E., Family members' perceptions of palliative cancer care: Predictors of family functioning and family members' health. Journal of Palliative Care, 1996. 12(4): p. 10-20. Fakhoury, W., McCarthy, J., and Addington-Hall, J., Determinants of informal caregivers' satisfaction with services for dying cancer patients. Social Science & Medicine, 1996. 42(5): p. 721-731. Goodkin, K., Blaney, N.T., Feaster, D.J., Baldewicz, T., Burkhalter, J.E., and Leeds, B., A randomized controlled clinical trial of a bereavement support group intervention in human immunodeficiency virus type 1-seropositive and -seronegative homosexual men. Archives of General Psychiatry, 1999. 56(1): p. 52-59. van den Boom, F., AIDS, euthanasia and grief. AIDS Care, 1995. 7(2 Suppl): p. 5175-5185. Summers, J., Zisook, S., Atkinson, J.H., Sciolla, A., Whitehall, W., Brown, S., Patterson, T., and Grant, I., Psychiatric morbidity associated with acquired immune deficiency syndrome-related grief resolution. The Journal of Nervous and Mental Disease, 1995. 183(6): p. 384-389. Kelly, B.J. and Varghese, F.T., Assisted suicide and euthanasia: What about the clinical issues? Australian and New Zealand Journal of Psychiatry, 1996. 30(1): p. 3-8. Swarte, N.B., van der Lee, M.L., van der Bom, J.G., van der Bout, J., and Heintz, P.M., Effects of euthanasia on the bereaved family and friends: A cross sectional study. British Medical Journal, 2003. 327: p. 189-192. Jacobs, S., Mazure, C., and Prigerson, H., Diagnostic criteria for traumatic grief. Death Studies, 2000. 24(3): p. 185-199. Swarte, N.B. and Heintz, A.P., Euthanasia and physician-assisted suicide. Annals of Medicine, 1999. 31(6): p. 364-71. van der Mass, P.J., van der Wal, G., Haverkate, I., de Graff, C.I., Kester, J.G., Onwuteka-Philpsen, B.D., van der Heide, A., Bosma, J., and Willems, D.L., Euthanasia, physician-assisted suicide, and other medical practices involving end of life in the Netherlands. 1990-1995, 1996. 335: p. 1699-1705. Hansson, R. and Stroebe, M.S., Grief, older adulthood, in Encyclopaedia of Primary Prevention and Health Promotion, T.P. Gullotta and M. Bloom, Editors. 2003, Plenum: New York. p. 515-521. Hansson, R. and Carpenter, B.N., Relationships in old age: Coping with the challenge of transition. 1994, New York: Guilford Press. Cartensen, L.L., Gross, J.J., and Fung, H.H., The social context of emotional experience. Annual Review of Gerontology and Geriatrics, 1997. 17: p. 325-352. Wells, Y.D. and Kendig, H.L., Health and well-being of spouse caregivers and the widowed. The Gerontologist, 1997. 37(5): p. 666-674.

115

179.

180. 181. 182. 183. 184. 185.

186.

187. 188. 189. 190. 191. 192.

193. 194.

195. 196.

197.

198. 199.

200.

201. 202.

Szanto, K., Prigerson, H.G., Houck, P.R., Ehrenpreis, L., and Reynolds, C.F., Suicidal ideation in elderly bereaved: The role of complicated grief. Suicide & Life-Threatening Behavior, 1997. 27(2): p. 194-207. Macias, C., Jones, D., Harvey, J., Barreira, P., Harding, C., and Rodican, C., Bereavement in the context of serious mental illness. Psychiatric Services, 2004. 55(4): p. 421-426. Jones, D., Harvey, J., Giza, D., Rodican, C., Barreira, P., and Macias, C., Parental death in the lives of people with serious mental illness. Journal of Loss & Trauma, 2003. 8(4): p. 307-322. Piper, W.E., Ogrodniczuk, J.S., Azim, H.F., and Weideman, R., Prevalence of loss and complicated grief among psychiatric outpatients. Psychiatric Services, 2001. 52(8): p. 1069-1074. Johnson, J.G., First, M.B., Block, S., Vanderwerker, L.C., Bambauer, K.Z., Zhang, B., and Prigerson, H.G., Stigmatization and receptivity to mental health services among recently bereaved adults. In Press. Nakao, M., Kashiwagi, M., and Yano, E., Alexithymia and grief reactions in bereaved Japanese women. Death Studies, 2005. 29(5): p. 423-433. Prigerson, H., Ahmed, I., Silverman, G.K., Saxena, A.K., Maciejewski, P.K., Jacobs, S., Kasl, S.V., Aqeel, N., and Hamirani, M., Rates and risks of complicated grief among psychiatric clinic patients in Karachi, Pakistan. Death Studies, 2002. 26(10): p. 781-792. Momartin, S., Silove, D., Manicavasagar, V., and Steel, Z., Complicated grief in Bosnian refugees: Associations with posttraumatic stress disorder and depression. Comprehensive Psychiatry, 2004. 45(6): p. 475-482. Tatz, C., Aboriginal suicide is different: A portrait of life and self-destruction. 2001, Canberra: Aboriginal Studies Press. Swan, P., Grief and health: The Indigenous legacy. Grief Matters: The Australian Journal of Grief and Bereavement, 1998. 1(2): p. 9-11. Swan, P. and Raphael, B., Ways forward: National Aboriginal and Torres Strait Islander mental health policy national consultancy report. 1998, Commonwealth of Australia,: Canberra. Brave-Heart, M.Y.H., The historical trauma response among natives and its relationship with substance abuse. Journal of Psychoactive Drugs, 2003. 35(1): p. 7-13. Brave-Heart, M.Y.H. and DeBruyn, L.M., The American Indian Holocaust: Healing historical unresolved grief. American Indian and Alaska Native Mental Health Research, 1998. 8(2): p. 56-78. Leenaars, A.A., Complicated grief: Suicide among the Canadian Inuit, in Complicated Grieving and Bereavement: Understanding and Treating People Experiencing Loss, G.R. Cox, R.A. Bendiksen, and R.G. Stevenson, Editors. 2002, Baywood Publishing Company Inc: Amityville. p. 99-118. Wanganeen, R., Spiritual healing using loss and grief. Aboriginal and Islander Health Worker Journal, 2001. 25(5): p. 12-14. Zubrick, S.R., Silburn, S.R., Lawrence, D.M., Mitrou, F.G., Dalby, R.G., Blair, E.M., Griffin, J., Milroy, H., DeMaio, J.A., Cox, A., and Li, J., The Western Australian Aboriginal Child Health Survey: The Social and Emotional Wellbeing of Aboriginal Children and Young People. Volume 2. 2005, Curtin University of Technology,Telethon Institute for Child Health Research: Perth. p. 632. Sullivan, K., National Indigenous Palliaitve Care Needs Study, in The National Palliative Care Program. 2003, Commonwealth of Australia,: Canberra. p. 381. Zygmont, M., Prigerson, H.G., Houck, P.R., Miller, M.D., Shear, M.K., Jacobs, S., and Reynolds, C.F., A post hoc comparison of paroxeline and nortriptyline for symptoms of traumatic grief. The Journal of Clinical Psychiatry, 1998. 59(5): p. 241-245. Abouguendia, M., Joyce, A.S., Piper, W.E., and Ogrodniczuk, J.S., Alliance as a mediator of expectancy effects in short-term group psychotherapy. Group Dynamics: Theory, Research, and Practice, 2004. 8(1): p. 3-12. Kipnes, D.R., Piper, W.E., and Joyce, A.S., Cohesion and outcome in short-term psychodynamic groups for complicated grief. International Journal of Group Psychotherapy, 2002. 52(4): p. 483-509. McCallum, M., Piper, W.E., Ogrodniczuk, J.S., and Joyce, A.S., Early process and dropping out from short-term group therapy for complicated grief. Group Dynamics: Theory, Research, and Practice, 2002. 6(3): p. 243-254. McCallum, M., Piper, W.E., Ogrodniczuk, J.S., and Joyce, A.S., Relationships among psychological mindedness, alexithymia and outcome in four forms of short-term psychotherapy. Psychology and Psychotherapy, 2003. 76: p. 133-144. Ogrodniczuk, J.S., Piper, W.E., Joyce, A.S., McCallum, M., and Rosie, J.S., Social support as a predictor of response to group therapy for complicated grief. Psychiatry, 2002. 65(4): p. 346-357. Ogrodniczuk, J.S., Piper, W.E., McCallum, M., Joyce, A.S., and Rosie, J.S., Interpersonal predictors of group therapy outcome for complicated grief. International Journal of Group Psychotherapy, 2002. 52(4): p. 511-535.

116

203. 204.

205. 206. 207.

208.

209.

210.

211. 212. 213.

214.

215. 216. 217.

218. 219. 220.

Ogrodniczuk, J.S., Joyce, A.S., and Piper, W.E., Changes in perceived social support after group therapy for complicated grief. The Journal of Nervous and Mental Disease, 2003. 191(8): p. 524-530. Ogrodniczuk, J.S., Piper, W.E., Joyce, A.S., McCallum, M., and Rosie, J.S., NEO-five factor personality traits as predictors of response to two forms of group psychotherapy. International Journal of Group Psychotherapy, 2003. 53(4): p. 417-442. Ogrodniczuk, J.S. and Piper, W.E., The effect of group climate on outcome in two forms of short-term group therapy. Group Dynamics: Theory, Research, and Practice, 2003. 7(1): p. 64-76. Ogrodniczuk, J.S., Piper, W.E., and Joyce, A.S., Differences in men's and women's responses to shortterm group psychotherapy. Psychotherapy Research, 2004. 14(2): p. 231-243. Ogrodniczuk, J.S., Piper, W.E., and Joyce, A.S., The negative effect of alexithymia on the outcome of group therapy for complicated grief: What role might the therapist play? Comprehensive Psychiatry, 2005. 46(3): p. 206-213. Piper, W.E., McCallum, M., Joyce, A.S., Rosie, J.S., and Ogrodniczuk, J.S., Patient personality and time-limited group psychotherapy for complicated grief. International Journal of Group Psychotherapy, 2001. 51(4): p. 525-552. Piper, W.E., Ogrodniczuk, J.S., Joyce, A.S., McCallum, M., and Rosie, J.S., Relationships among affect, work, and outcome in group therapy for patients with complicated grief. American Journal of Psychotherapy, 2002. 56(3): p. 347-361. Piper, W.E., Ogrodniczuk, J.S., McCallum, M., Joyce, A.S., and Rosie, J.S., Expression of affect as a mediator of the relationship between quality of object relations and group therapy outcome for patients with complicated grief. Journal of Consulting and Clinical Psychology, 2003. 71(4): p. 664-671. Shear, K., Frank, E., Foa, E., Cherry, C., Reynolds, C.F., Vander Bilt, J., and Masters, S., Traumatic grief treatment: A pilot study. American Journal of Psychiatry, 2001. 158(9): p. 1506-1508. Shear, K., Frank, E., Houck, P.R., and Reynolds, C.F., Treatment of complicated grief: A randomized controlled trial. JAMA, 2005. 293(21): p. 2601-2608. Sprang, G., The use of Eye Movement Desensitization and Reprocessing (EMDR) in the treatment of traumatic stress and complicated mourning: Psychological and behavioral outcomes. Research on Social Work Practice, 2001. 11(3): p. 300-320. Layne, C.M., Pynoos, R.S., Saltzman, W.R., Arslanagic, B., Black, M., Savjak, N., Popovic, T., Durakovic, E., Music, M., Campara, N., Djapo, N., and Houston, R., Trauma/grief-focused group psychotherapy: School-based postwar intervention with traumatized Bosnian adolescents. Group Dynamics: Theory, Research, and Practice, 2001. 5(4): p. 277-290. Kato, P., M. and Mann, T., A synthesis of psychological interventions for the bereaved. Clinical Psychology Review, 1999. 19(3): p. 275-296. Jacobs, S. and Prigerson, H., Psychotherapy of traumatic grief: A review of evidence for psychotherapeutic treatments. Death Studies, 2000. 24(6): p. 479-495. Stroebe, M., Hansson, R.O., Stroebe, W., and Schut, H., eds. Handbook of bereavement research: Consequences, coping, and care. Handbook of bereavement research: Consequences, coping, and care. 2001, American Psychological Association Press: Washington, DC. Neria, Y. and Litz, B., Bereavement by traumatic means: The complex synergy of trauma and grief. Journal of Loss & Trauma, 2004. 9(1): p. 73-87. Sanders, C.M., Comparison of adult bereavement in the death of a spouse, child, and parent. Omega: Journal of Death and Dying, 1979. 10: p. 303-322. Green, B.L., Traumatic loss: Conceptual and empirical links between trauma and bereavement. Journal of Personal and Interpersonal Loss, 2000. 5: p. 1-17.

117