International Journal of Social

Psychiatry
http://isp.sagepub.com

Measures of the Therapeutic Relationship in Severe Psychotic Illness: A

Comparison of Two Scales
Rob Bale, Jocelyn Catty, Hilary Watt, Nan Greenwood and Tom Burns
International Journal of Social Psychiatry 2006; 52; 256
DOI: 10.1177/0020764006067195

The online version of this article can be found at:

http://isp.sagepub.com/cgi/content/abstract/52/3/256

Published by:

http://www.sagepublications.com

Additional services and information for International Journal of Social Psychiatry can be found at:

Email Alerts: http://isp.sagepub.com/cgi/alerts

Subscriptions: http://isp.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.co.uk/journalsPermissions.nav

Citations http://isp.sagepub.com/cgi/content/refs/52/3/256

Downloaded from http://isp.sagepub.com by Pablo Fossa on October 29, 2008

 MEASURES OF THE THERAPEUTIC RELATIONSHIP IN SEVERE
PSYCHOTIC ILLNESS: A COMPARISON OF TWO SCALES

ROB BALE, JOCELYN CATTY, HILARY WATT, NAN GREENWOOD

& TOM BURNS

ABSTRACT
Background: A durable therapeutic relationship is central to mental health prac-
tice. The Working Alliance Inventory (WAI) and the Helping Alliance Question-
naire (HAQ) are established instruments for measuring such a relationship.
Aims: The project aimed to test the correlation between the two scales for
patients with severe psychotic illness treated in an Assertive Community Treat-
ment (ACT) team.
Methods: Ninety-one patients of an ACT team and their key-workers were
recruited to complete the measures.
Results: Seventy-one patients (78%) completed the scales, and key-workers
completed scales for every eligible patient. Both groups rated the relationship
positively. There was a strong and significant correlation between the patient
version of the WAI and the HAQ. There were significant but much weaker correla-
tions between the patient-rated WAI and HAQ and the key-worker WAI.
Conclusion: The patient version of the WAI and the HAQ seem to measure the
patients view of the relationship equivalently. The HAQ is simpler and easier to
administer than the WAI.

INTRODUCTION

The relationship between patients and clinicians has been subject to considerable research in
psychotherapy but much less has been written about it in day-to-day psychiatric practice with
patients with severe and enduring mental illness. In this setting, dierent terms, including
engagement, the therapeutic or working relationship and the therapeutic or working alli-
ance, are often used interchangeably, although each was originally coined to reect a distinct
concept.

The therapeutic alliance

The concept of the therapeutic alliance is central to psychoanalysis and psychoanalytic psy-
chotherapy. It originates in Freuds early writing, where he argued that, The rst aim of the
treatment consists in attaching [the patient] to the treatment and the person of the physician
(Freud, 1913). The early psychoanalytic conception of the alliance was closely bound up with
the transference (Sandler et al., 1973). Freud did not distinguish the two but saw the alliance
International Journal of Social Psychiatry. Copyright & 2006 Sage Publications (London, Thousand Oaks and
New Delhi) www.sagepublications.com Vol 52(3): 256266. DOI: 10.1177/0020764006067195

Downloaded from http://isp.sagepub.com by Pablo Fossa on October 29, 2008

 BALE ET AL.: MEASURES OF THE THERAPEUTIC RELATIONSHIP 257

as a form of positive transference, while Zetzel (1956), who coined the term, maintained this
view. Greenson, however, distinguished the transference from what he called the working
alliance, dening the latter as the relatively nonneurotic, rational rapport which the patient
has with his analyst and the patients capacity to work purposefully in the treatment situa-
tion (Greenson, 1967).
The psychoanalytic model of the therapeutic alliance is not the only one, however. It bears
some relationship to Rogers three core conditions for person-centred therapy (Rogers et al.,
1967), while Bordins pantheoretical formulation of the alliance draws on both psycho-
dynamic and cognitive behavioural concepts (Bordin, 1994). Research into the care of indi-
viduals with severe mental illness has generally neglected the therapeutic alliance despite an
expanding literature on treatment compliance and recognition that the doctorpatient rela-
tionship is a strong determinant of such compliance. The alliance has gradually become of
increasing interest in psychiatric services research, however, although the implications of
using the concept in a psychiatric context have been explored only recently (Catty, 2004;
McCabe & Priebe, 2004; McGuire et al., 2001).

Measures of the therapeutic alliance

Several scales have been developed to measure the therapeutic alliance (Horvath et al., 1993),
mostly within the context of psychotherapy and often only reported in single studies. Of these
the Working Alliance Inventory (WAI) (Horvath & Greenberg, 1989), based on Bordins
pantheoretical model (Bordin, 1994), has been the most extensively used. The WAI is a
36-item self-completed measure using a seven-point Likert scale; it is divided into three sub-
scales measuring goals, bonds and tasks, seen as components of the therapeutic alliance.
Patient and therapist versions are both available, each giving a total score ranging from
36 to 252.
The WAI has been extensively validated in psychotherapy (Horvath & Greenberg, 1989)
and has also been validated for use with patients with severe mental illness in psychiatric set-
tings (Ralph & Clary, 1992; Stylianos & Goering, 1989). Stylianos and Goering tested its
validity and reliability with 50 people with chronic psychotic disorders living in the com-
munity. They found no evidence of its discriminant validity with this group, who did not dis-
tinguish between the sub-scales of goals, tasks and bonds. Gehrs and Goering (1994) also
used the measure for patients with schizophrenia in rehabilitation services. They found the
patient-rated and professional-rated versions to be highly correlated, which they argued
was evidence of the measures construct validity with this population. Neale and Rosenheck
(1995) and Gehrs and Goering (1994) also found associations between alliance ratings and
outcome, particularly with the professional-rated version of the WAI.
The Helping Alliance Questionnaire (Priebe & Grutyers, 1993) was developed to assess the
quality of the therapeutic alliance between patients with severe and long-term mental illness
and their key-workers. It comprises ve questions. The rst four are self-rated on 100 mm
visual analogue scales; the last is categorical.

Engagement in community care

The term engagement is also dened in more than one way in the literature. At its most
basic, it is used to indicate that patients are in contact with a service. It is also used to reect

Downloaded from http://isp.sagepub.com by Pablo Fossa on October 29, 2008

258 INTERNATIONAL JOURNAL OF SOCIAL PSYCHIATRY 52(3)

this contact being regular, rather than intermittent. A less clear denition incorporates the
ideas of contact with the service and the individual agreeing with and participating in the
treatment. Engagement is also used to describe a process that occurs in the development
of the relationship between patient and clinician, and to indicate, specically, the develop-
ment of a trusting relationship. Thus Dixon et al. (1995) dene the engagement phase in
a case management service as involving developing a trusting relationship between treatment
team and the patient. At this point, the concepts of engagement and therapeutic alliance
move closer together.
The need for mental health services to maintain contact with patients with severe and
enduring mental illness has been highlighted in the National Service Framework for
Mental Health (Department of Health, 1999). Early community mental health initiatives in
the deinstitutionalisation period emphasised the central importance of coordination of care
and proactive engagement of patients (Intagliata, 1982). These early case management
approaches rapidly evolved into more clinically focused services such as Assertive Commu-
nity Treatment (ACT) (Stein & Test 1980), in which establishing an enduring relationship
is currently seen as an essential prerequisite for the provision of eective treatments and reha-
bilitative interventions. Ironically, however, Stein and Test originally dened the therapeutic
relationship as having a secondary role, because of their emphasis on avoiding pathologically
dependent relationships.
The importance of continued engagement was recently demonstrated when Appleby and
colleagues (1999) found that, of patients known to mental health services who committed
suicide, 28% had lost contact with those services. There is also signicant concern expressed
in the UK that patients who lose contact with services may pose a risk to the public (Ritchie,
1994). ACT and intensive case management (ICM) have been adopted and endorsed speci-
cally in order to reduce the risk of patients dropping out of care (Department of Health,
1999). A key aim for these services is to engage the patient and to develop an ongoing rela-
tionship in order to deliver eective treatments (Kent & Burns, 1996).

AIMS AND OBJECTIVES

The primary aim of the present study was to compare the WAI and the HAQ; in particular, to
determine whether there was any correlation between the HAQ and the patient version of the
WAI, demonstrating the level of construct validity of the newer HAQ against the more estab-
lished scale. The overall objective of the study was thus to determine whether the shorter and
simpler HAQ measures the relationship equivalently to the WAI.
Further aims were to test the level of association between the patient and key-worker
versions of the WAI in this patient group, and to test the internal validity of the HAQ (the
degree of correlation between its items and their correlation with the total score).
Our hypotheses were that: the patient and key-worker versions of the WAI would be sig-
nicantly associated; the HAQ and the patient version of the WAI would be signicantly
associated; and the HAQ and the key-worker version of the WAI would be signicantly
associated.

Downloaded from http://isp.sagepub.com by Pablo Fossa on October 29, 2008

 BALE ET AL.: MEASURES OF THE THERAPEUTIC RELATIONSHIP 259

METHODS

Sample
The ACT team from which the sample was drawn was evaluated as part of a multi-centre trial
of ICM in the UK, the UK700 study (Burns et al., 1999). After the initial research period of
two years, the team became an established part of the local adult mental health services. The
team consisted of 10 case managers (key-workers) with a maximum caseload of 12 patients
each. The inclusion criteria were having a diagnosis of psychosis, being between 18 and
65 years of age, having had at least two admissions to hospital, the last one within the pre-
ceding two years, and not having a primary organic or substance misuse diagnosis. All
patients who had been cared for by the team for more than three months were eligible, 69
of them being from the original UK700 sample. Patients were invited to take part in the
study by a clinician-researcher (RB) and interviewed after giving written, informed consent.

Data collection
Demographics, diagnostic data and psychiatric history were obtained from face-to-face inter-
view and patient notes. Baseline data from the UK700 study were used where appropriate.
Demographic data were collected using the UK700 demographic schedule (Burns et al.,
1999). A Research Diagnostic Criteria diagnosis was obtained using the OPCRIT system
(McGun et al., 1991). Ethnicity was obtained by face-to-face interview or from case
notes for patients who refused interview.
The patient-rated WAI and HAQ were completed in face-to-face interviews with patients.
Key-workers were asked to complete the key-worker version for each of their patients
included in the study. In this study an adapted version of the WAI was used, with the term
key-worker substituted for therapist. If patients were unable or unwilling to read the
measure, each question was read to them along with the choice of answers.

Data analysis
Data were analysed using SPSS (Version 8). Items that are scored negatively were entered
into SPSS with the polarity reversed. Spearmans rank correlation coecients were used to
test for associations. The sum score of the HAQ was obtained according to the methods
described in the original paper (Priebe & Grutyers, 1993), with the nal question (how do
you feel after a session with your key-worker?) scored as zero for unchanged/worse and
50 for better.

RESULTS

Of 91 eligible patients, 71 (78%) agreed to all or part of the interview. For the 20 patients who
refused interview, the demographic schedule and OPCRIT were completed from notes, and
key-workers completed the key-worker version of the WAI.

Patient characteristics
All patients had a diagnosis of a psychosis (60% having schizophrenia) and had been pre-
viously admitted at least twice, once within the two years prior to acceptance by the ACT

Downloaded from http://isp.sagepub.com by Pablo Fossa on October 29, 2008

260 INTERNATIONAL JOURNAL OF SOCIAL PSYCHIATRY 52(3)

Table 1
Interviewees and refusers: demographics and diagnoses

Interviewees Refusers p value

(n 71) (n 20)

Male 38 (54%) 15 (75%) 0.121

Median age (years) 42 39 0.442
Single 38 (54%) 16 (80%) 0.303
Inpatient 14 (20%) 4 (20%) 1.01
Legally detained 9 (13%) 2 (10%) 1.0 1
Ethnicity White 32 (45%) 3 (15%) 0.181
Black African 16 (23%) 7 (35%)
Black Caribbean 8 (11%) 2 (10%)
Other 14 (20%) 3 (15%)
Missing 1 (1%) 5 (25%)
Diagnosis Major depression 3 (4%) 1 (5%) 0.224
Bipolar aective disorder 3 (4%) 0 (0%)
Schizophrenia 45 (63%) 10 (50%)
Schizoaective 17 (24%) 7 (35%)
Other 3 (4%) 2 (10%)

1. Fishers Exact Test (2-sided); 2. Mann-Whitney U test (2-tailed); 3. Pearson
-squared test.

team. More of the patients who refused interview were single and male, and fewer were white,
but none of these dierences was statistically signicant (Table 1).

Working Alliance Inventory

Seventy patients completed the patient version of the WAI. The mean total score was 198
(range 115 to 252, SD 38). Key-worker-rated WAIs were completed for 91 patients, with
a mean total score of 198 (115252, SD 28). There were no statistically signicant dier-
ences between key-worker ratings of the WAI for patients who were interviewed and those
who refused, on any of the sub-scales or the sum scores.
In each version of the WAI (patient- or key-worker-rated), the sub-scales (bonds, goals and
tasks) correlated highly with each other (Spearmans rank correlations of greater than 0.8,
p < 0:001) and with the total scores (Spearmans rank correlations greater than 0.9,
p < 0:001).
Table 2 shows the associations between the two versions of the WAI and their sub-scales.
The sum scores of the patient and key-worker versions were signicantly correlated, but the
strength of the association was much lower (0.315, p 0:008). Correlations between the sub-
scales of the patient-rated WAI and sub-scales of the key-worker-rated WAI were low but
signicant in each case, with the exception of the correlations of patient-rated goals with
key-worker-rated bonds and the key-worker-rated tasks.

Helping Alliance Questionnaire

Mean scores for the rst four questions were all greater than 72 out of a possible maximum of
100 (Table 3). Forty-six patients (65%) felt better after a session with their key-worker,
whereas 25 (35%) felt unchanged or worse. The four questions that were continuous variables
were moderately correlated with each other, with coecients ranging from 0.3 to 0.52. Each

Downloaded from http://isp.sagepub.com by Pablo Fossa on October 29, 2008

 BALE ET AL.: MEASURES OF THE THERAPEUTIC RELATIONSHIP 261

Table 2
Working Alliance Inventory: correlations between key-worker and patient sum and sub-scale scores

Spearmans rank correlation Key-worker Key-worker Key-worker Key-worker

coecients (p values) WAI sum WAI bonds WAI goals WAI tasks

Patient WAI Sum 0.315 0.245 0.322 0.274

(0.008) (0.041) (0.007) (0.022)
Patient WAI bonds 0.366 0.322 0.343 0.333
(0.002) (0.006) (0.004) (0.005)
Patient WAI goals 0.252 0.174 0.273 0.216
(0.035) (0.149) (0.022) (0.073)
Patient WAI tasks 0.324 0.251 0.337 0.284
(0.006) (0.036) (0.004) (0.017)

question demonstrated a correlation of over 0.64 with the sum score (Table 4). These values
suggest that the scale is internally consistent.
As Table 5 shows, the sum score of the HAQ was highly and signicantly correlated with
the sum score of the patient version of the WAI (correlation coecient 0.76, p < 0:001). All
individual questions of the HAQ were correlated with the sum score of the patient version of
the WAI, with coecients of at least 0.4. Correlations between the HAQ and the key-worker
version of the WAI were much lower, with a coecient of 0.32 for the sum scores and between
0.1 and 0.3 for individual questions.

DISCUSSION

ICM and ACT are expensive forms of care. If engagement is to be adequately targeted and
achieved by these service models, there is a clinical need to identify those patients who have

Table 3
Helping Alliance Questionnaire scores1

Mean (SD)

HAQ1 (do you feel understood by your key-worker?) 72.2 (30.6)2

HAQ2 (do you feel criticised by your key-worker?) 80.6 (26.9)2
(Polarity reversed)
HAQ3 (how much is your key-worker committed to 80.0 (24.54)2
and involved in your treatment?)
HAQ4 (is the treatment you are currently receiving 72.1 (33.82)2
right for you?)
Sum HAQ 337.4 (97.9)3

1. Questions 1 to 4 (continuous scale) only; 2. Maximum score 100 (i.e. most

positive) scored on a visual analogue scale; 3. Maximum score 450. Includes
Q5 scored as 0 unchanged/worse, 50 better.

Downloaded from http://isp.sagepub.com by Pablo Fossa on October 29, 2008

262 INTERNATIONAL JOURNAL OF SOCIAL PSYCHIATRY 52(3)

Table 4
Associations between individual HAQ questions1

Spearmans rank correlation HAQ1 HAQ2 HAQ3 HAQ4

coecients (p values)

HAQ2 (<0.459
(<0.001)
HAQ3 (<0.518 (<0.522
(<0.001) (<0.001)
HAQ4 (<0.311 (<0.437 (<0.488
(<0.001) (<0.001) (<0.001)
Sum HAQ (<0.722 (<0.722 (<0.697 (<0.644
(<0.001) (<0.001) (<0.001) (<0.001)

1. Questions 1 to 4 (continuous scale) only.

Table 5
Associations between HAQ items and the WAI1

Spearmans rank correlation Patient Patient Patient Patient

coecients (p values) WAI bonds WAI goals WAI tasks WAI sum

HAQ1 (<0.546 (<0.579 (<0.706 (<0.657

(<0.001) (<0.001) (<0.001) (<0.001)
HAQ2 (<0.453 (<0.483 (<0.550 (<0.539
(<0.001) (<0.001) (<0.001) (<0.001)
HAQ3 (<0.564 (<0.586 (<0.611 (<0.620
(<0.001) (<0.001) (<0.001) (<0.001)
HAQ4 (<0.410 (<0.431 (<0.463 (<0.455
(<0.001) (<0.001) (<0.001) (<0.001)
HAQ Sum (<0.676 (<0.622 (<0.797 (<0.758
(<0.001) (<0.001) (<0.001) (<0.001)

Key-worker Key-worker Key-worker Key-worker

WAI bonds WAI goals WAI tasks WAI sum

HAQ1 (0.257 (0.294 (0.281 (0.312

(0.03) (0.013) (0.018) (0.008)
HAQ2 (0.272 (0.297 (0.293 (0.315
(0.022) (0.012) (0.013) (0.008)
HAQ3 (0.128 (0.120 (0.104 (0.117
(0.288) (0.319) (0.389) (0.333)
HAQ4 (0.129 (0.135 (0.136 (0.150
(0.282) (0.261) (0.258) (0.211)
HAQ Sum (0.287 (0.302 (0.293 (0.324
(0.015) (0.0) (0.013) (0.006)

1. Questions 1 to 4 (continuous scale) only.

Downloaded from http://isp.sagepub.com by Pablo Fossa on October 29, 2008

 BALE ET AL.: MEASURES OF THE THERAPEUTIC RELATIONSHIP 263

not fully engaged with the services and who are likely to need increased input. Whether
ratings of the therapeutic relationship may be related to rates of retention of patients in
the service is beyond the scope of the present paper, but it is clearly an important question
for the future.
While engagement and the therapeutic relationship are not synonymous, however, they are
clearly related, and there is thus a pressing need for a simple measure of the therapeutic rela-
tionship for use in clinical practice. Only one of the two instruments identied as being used
within community care of the severely mentally ill, the WAI, currently has a key-worker
version. Patients who use ACT and ICM services are likely to be reluctant to complete
long questionnaires, which gives a clinician-rated scale much to recommend it particularly
as it is the clinician who will act on its ndings. Conversely, the HAQ is short, though not
currently available in a clinician-rated version, but its validity has not been established as
extensively as that of the WAI.

Working Alliance Inventory

The results of this study suggest that both patients and key-workers rate the relationship
highly, with a mean score of 198 each out of a maximum of 252. Despite the identical
means, however, the patient and key-worker versions had only a weak correlation. Predicting
an individual patient score from a key-worker score cannot therefore be done condently.
This suggests that the two versions of the scale may not be measuring the same concepts in
this group. Questions may only supercially mirror each other and may have dierent mean-
ings for patients and key-workers. This is in contrast with previous ndings not only with
psychotherapy patients but also with schizophrenia patients in rehabilitation services
(Gehrs & Goering, 1994).
Examination of the three sub-scales (goals, bonds and tasks) within each version (patient
and key-worker) demonstrates such high levels of correlation between them that it is ques-
tionable whether they can be regarded as separate components. This bears out Ralph and
Clarys (1992) and Gehrs and Goerings (1994) ndings, which similarly failed to support
the original conceptualisation of the working alliance as having three distinct dimensions
in this population.

Helping Alliance Questionnaire

The HAQ is a much briefer and simpler questionnaire but currently obtains only the patients
views. We found moderate correlations between each question of the HAQ. This suggests
that the questions are related but do measure slightly dierent items. All correlations were
higher than the 0.20 suggested (Streiner & Norman, 1995) as the minimum to demonstrate
the internal consistency of the scale. Overall the HAQ ndings indicate that patients all
rated the relationship highly.
The HAQ was correlated strongly both overall and in each individual question with the
patient version of the WAI. The strongest association was the global one, between total
scores, and this is consistent with reviews of measures of the relationship in psychotherapy,
where the dierent measures have been found to be strongly associated globally, but less asso-
ciated in their sub-scales (Horvath et al., 1993). Its correlation with the key-worker version of

Downloaded from http://isp.sagepub.com by Pablo Fossa on October 29, 2008

264 INTERNATIONAL JOURNAL OF SOCIAL PSYCHIATRY 52(3)

the WAI was lower, however, and not all the sub-scales were signicant. This suggests that
the HAQ provides as good a measure as the WAI of the patients view of the relationship
but that patients and key-workers views are not strongly related. Conversely, this implies
that the key-workers view of the relationship cannot be used safely as an indicator of the
patients assessment. The low correlation between the key-worker and patient versions of
the WAI provides further evidence for this.

Limitations
This study had a number of limitations. The scales were administered only once to the patient
group at dierent points in their contact with the service. They had all been in contact for at
least three months but some had been in contact for much longer. There is also no clinical
gold standard of the therapeutic relationship against which to establish the criterion validity
of the scales, unlike, for instance, a clinical diagnosis of depression to validate depression
scales. Further work should include repeated measures of the relationship over time and
investigate the association of the therapeutic relationship with clinical outcome.
Patients who agreed to interview were representative of the total eligible group both demo-
graphically and diagnostically and in the level of therapeutic alliance their key-workers
regarded them as having. This should be treated with some caution, however, as the study
established that the key-worker ratings could not predict patient ratings strongly.
The present study does not represent a thorough validation of the HAQ, not covering, for
instance, test-retest reliability. It does, however, cover internal consistency and construct
validity against both the patient-rated and the key-worker-rated WAI.

Implications
This study conrms the feasibility of both scales to measure the relationship between patients
and key-workers in an ACT service. It demonstrates that patients in this group perceive the
relationship dierently from key-workers, so that one perception cannot be assumed from the
other. The practical advantages of the key-worker WAI may therefore be outweighed by its
low correlation with the patient-rated measures. Whether any ratings of the alliance may be
associated with clinical outcome demands further investigation. The HAQ, however, is brief
and easy to administer and provides an equally reliable measure of the patients perception of
the therapeutic relationship. Use of this measure is feasible in routine clinical practice and
may assist clinicians direct eorts at engaging patients, thereby improving contact with
this vulnerable group.

REFERENCES

APPLEBY, L., SHAW, J., AMOS, T., McDONNELL, R., HARRIS, C., McCANN, K., KIERNAN, K.,
DAVIES, S., BICKLEY, H. & PARSONS, R. (1999) Suicide within 12 months of contact with mental
health services: national clinical survey. BMJ, 318 (7193), 12351239.
BORDIN, E.S. (1994) Theory and research on the therapeutic working alliance: new directions. In The Work-
ing Alliance: Theory, Research, and Practice (eds A.O. Horvath & L.S. Greenberg). New York, NY: John
Wiley & Sons, pp. 1337.

Downloaded from http://isp.sagepub.com by Pablo Fossa on October 29, 2008

 BALE ET AL.: MEASURES OF THE THERAPEUTIC RELATIONSHIP 265

BURNS, T., CREED, F., FAHY, T. & WHITE, I. (1999) Intensive versus standard case management for
severe psychotic illness: a randomised trial. UK 700 Group. Lancet, 353 (9171), 21852189.
CATTY, J. (2004) The vehicle of success: theoretical and empirical perspectives on the therapeutic alliance
in psychotherapy and psychiatry. Psychology and Psychotherapy: Theory, Research and Practice, 77,
255272.
DEPARTMENT OF HEALTH (1999) Modern Standards and Service Models: National Service Framework for
Mental Health. London: Department of Health.
DIXON, L.B., KRAUSS, N., KERNAN, E., LEHMAN, A.F. & DEFORGE, B.R. (1995) Modifying the
PACT model to serve homeless persons with severe mental illness. Special Section: Assertive community
treatment. Psychiatric Services, 46, 684688.
FREUD, S. (1913) On beginning the treatment. In The Standard Edition of the Complete Psychological Works
of Sigmund Freud (J. Strachey). London: Hogarth Press, 1964, Volume 12.
GEHRS, M. & GOERING, P. (1994) The relationship between the working alliance and rehabilitation out-
comes of schizophrenia. Psychosocial Rehabilitation Journal, 18, 4354.
GREENSON, R. (1967) The Technique and Practice of Psycho-analysis. London: Hogarth Press.
HORVATH, A. & GREENBERG, L.S. (1989) Development and validation of the Working Alliance
Inventory. Journal of Counseling Psychology, 36(2), 223233.
HORVATH, A., GASTON, L. & LUBORSKY, L. (1993) The therapeutic alliance and its measures. In
Psychodynamic Treatment Research: A Handbook for Clinical Practice (eds N. Miller, L. Luborsky,
J. Barber &u. Docherty). New York, NY: Basic Books, pp. 247273.
INTAGLIATA, J. (1982) Improving the quality of community care for the chronically mentally disabled: the
role of case management. Schizophrenia Bulletin, 8(4), 655674.
KENT, A. & BURNS, T. (1996) Setting up an assertive community treatment service. Advances in Psychiatric
Treatment, 2, 143150.
McCABE, R. & PRIEBE, S (2004) The therapeutic relationship in the treatment of severe mental illness: a
review of methods and ndings. International Journal of Social Psychiatry, 50(2), 115128.
McGUIRE, R., McCABE, R. & PRIEBE, S. (2001) Theoretical frameworks for understanding and investigat-
ing the therapeutic relationship in psychiatry. Social Psychiatry & Psychiatric Epidemiology, 36, 557564.
McGUFFIN, P., FARMER, A. & HARVEY, I. (1991) A polydiagnostic application of operational criteria
in studies of psychotic illness: development and reliability of the OPCRIT system. Archives of General
Psychiatry, 48(8), 764770.
NEALE, M.S. & ROSENHECK, R.A. (1995) Therapeutic alliance and outcome in a VA intensive case
management program. Psychiatric Services, 46, 719721.
PRIEBE, S. & GRUTYERS, T. (1993) The role of the helping alliance in psychiatric community care: a pro-
spective study. Journal of Nervous & Mental Disease, 181(9), 552557.
PRIEBE, S. & GRUTYERS, T. (1995) Patients assessment of treatment predicting outcome. Schizophrenia
Bulletin, 21, 8794.
RALPH, R. & CLARY, B. (1992) The Working Alliance Inventory: measuring the relationship between
patient and case manager. Paper presented at the 1992 National Conference on State Mental Health
Agency Services Research Conference, Baltimore, Maryland.
RITCHIE, J.H. (1994) The report of the enquiry into the care and treatment of Christopher Clunis presented to
the Chairman of the North East Thames and South East Thames Regional Health Authorities. London:
HMSO.
ROGERS, C.R., GENDIN, G.T., KIESLE, D.V. & TRUAX, L.B. (1967) The Therapeutic Relationship and its
Impact: A Study of Psychotherapy with Schizophrenics. Madison, WI: University of Wisconsin Press.
SANDLER, J., DARE, C. & HOLDER, A. (1973)The treatment alliance. In The Patient and the Analyst (eds
J. Sandler, C. Dare & A. Holder). London: Karnac Books, pp. 2736.
STEIN, L.I. & TEST, M.A. (1980) Alternative to mental hospital treatment. I. Conceptual model, treatment
program, and clinical evaluation. Archives of General Psychiatry, 37(4), 392397.
STREINER, D.L. & NORMAN, G.R. (1995) Health Measurement Scales: A Practical Guide to their Develop-
ment and Use (2nd edition). Oxford: Oxford University Press.
STYLIANOS, S.K. & GOERING, P.N. (1989) The Working Alliance Inventory: Reliability and Validity in
the Context of Rehabilitation for Clients with Chronic Psychotic Disorders. Paper presented at the
annual meeting of the Society for Psychotherapy Research, Toronto.
ZETZEL, E. (1956) Current concepts of transference. International Journal of Psychoanalysis, 39, 369376.

Downloaded from http://isp.sagepub.com by Pablo Fossa on October 29, 2008

266 INTERNATIONAL JOURNAL OF SOCIAL PSYCHIATRY 52(3)

Dr Rob Bale, Consultant Psychiatrist, City East Community Mental Health Team, Littlemore Mental Health Centre,
Sandford Road, Littlemore, Oxon, UK.
Dr Jocelyn Catty, Research Fellow, Department of Mental Health, St Georges Hospital Medical School, London,
UK.
Hilary Watt, Senior Statistician, Department of Mental Health, St Georges Hospital Medical School, London, UK.
Nan Greenwood, Department of Mental Health, St Georges Hospital Medical School, London, UK.
Professor Tom Burns, Department of Mental Health, St Georges Hospital Medical School, London, UK.
Correspondence to Dr Rob Bale, Consultant Psychiatrist, City East Community Mental Health Team, Littlemore
Mental Health Centre, Sandford Road, Littlemore, Oxon OX4 4XN, UK.
Email: Rob.Bale@obmh.nhs.uk

Downloaded from http://isp.sagepub.com by Pablo Fossa on October 29, 2008