Patient Education and Counseling 90 (2013) 233238

Contents lists available at SciVerse ScienceDirect

Patient Education and Counseling

journal homepage: www.elsevier.com/locate/pateducou

Communication Study

Working alliance in communication skills training for oncology clinicians:

A controlled trial
Claudia Meystre a, Celine Bourquin a, Jean-Nicolas Despland a, Friedrich Stiefel b, Yves de Roten a,*
a
b

Department of Psychiatry, Lausanne University Hospital, CH-1011 CHUV-Lausanne, Switzerland

Service of Liaison Psychiatry, Lausanne University Hospital, CH-1011 CHUV-Lausanne, Switzerland

A R T I C L E I N F O

A B S T R A C T

Article history:
Received 12 March 2012
Received in revised form 16 October 2012
Accepted 19 October 2012

Objective: The aim of this study was to evaluate the impact of communication skills training (CST) on
working alliance and to identify specic communicational elements related to working alliance.
Methods: Pre- and post-training simulated patient interviews (6-month interval) of oncology physicians
and nurses (N = 56) who beneted from CST were compared to two simulated patient interviews with a
6-month interval of oncology physicians and nurses (N = 57) who did not benet from CST. The patient
clinician interaction was analyzed by means of the Roter Interaction Analysis System (RIAS). Alliance was
measured by the Working Alliance Inventory Short Revised Form.
Results: While working alliance did not improve with CST, generalized linear mixed effect models
demonstrated that the quality of verbal communication was related to alliance. Positive talk and
psychosocial counseling fostered alliance whereas negative talk, biomedical information and patients
questions diminished alliance.
Conclusion: Patientclinician alliance is related to specic verbal communication behaviors.
Practice implications: Working alliance is a key element of patientphysician communication which
deserves further investigation as a new marker and efcacy criterion of CST outcome.
2012 Elsevier Ireland Ltd. All rights reserved.

Keywords:
Oncology
Alliance
Communication skills training
RIAS

1. Introduction
1.1. Communication in cancer care
Communication has been recognized as a key element in
medical and especially in cancer care, affecting both clinicians and
patients [1,2]. Ineffective communication may result in patients
confusion [3,4], psychological distress and difculties in expressing feelings and understanding information [4,5], and contribute
to clinicians stress, lack of job satisfaction and emotional burnout
[6,7].
1.2. Communication skills training
These observations have led to the development and implementation of communication skills training (CST) for oncology
clinicians [8]. Such training aims to increase the clinicians skills to
respond to the patients needs, to foster their relationship and to
facilitate the exchange of information.

* Corresponding author at: Institut Universitaire de Psychotherapie, Avenue de

Morges 10, CH-1004 Lausanne, Switzerland. Tel.: +41 21 314 20 54.
E-mail address: Yves.DeRoten@chuv.ch (Y. de Roten).
0738-3991/$ see front matter 2012 Elsevier Ireland Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.pec.2012.10.013

There is increasing evidence, that CST improves patient

physician communication [913]. For example, Falloweld et al.
[14] found that clinicians attending a 3-day CST use more focused
and open questions, fewer leading questions, express more
empathy, interrupt patients less often and provide more appropriate responses than those without training. It has also been
demonstrated that CST enhances reassurance, recognition of
emotions and understanding of the patient [1517]; in addition,
psychosocial issues and concerns are more often addressed [18
20], the interview is prolonged [21], verbal dominance is reduced
[17] and patients participate more actively [16,18], in particular by
asking more questions about their diagnosis [20]. However, CST
has been criticized for overlooking the relational and affective
dimensions of the patient encounter [2224].
1.3. Alliance
In healthcare, the importance of the relational factor
described as the connectional dimension of medical care [25]
has been recognized in the patient-centeredness theoretical and
empirical literature. Patient-centered care promotes a physician
patient relationship in which psychosocial issues and the patients
illness experience are addressed and patient and physician share
decisional responsibility [26]. Working alliance, broadly dened as
the development of common agreed tasks and goals and a strong

C. Meystre et al. / Patient Education and Counseling 90 (2013) 233238

234

therapeutic bond, is seen as a fundamental requirement, not

simply a useful addition to the medical model [27]. A sound
conceptual framework and growing empirical evidence conrm
the importance of working alliance in healthcare: moderate to
strong relationships were found between working alliance and
patients perceived utility or value of treatment, patients selfefcacy, treatment adherence, and satisfaction with community
care [2830].
Working alliance includes two elements: (1) at a fundamental
level, the patients ability to trust, hope and have faith in the
clinicians ability to help, and (2) different types of alliance
depending on the relevant therapeutic tasks and goals [31]. It
refers to a collaborative and/or a negotiated relationship characterized by the patients involvement in the treatment process.
Strong alliance implies that (i) the clinician and the patient dene
the medical problem together, (ii) they agree on the objectives and
(iii) jointly develop a treatment plan, (iv) for which they share
responsibility. To build and strengthen working alliance, communication behaviors that enable clinicians to reinforce patient
cooperation, like checking for his/her understanding, asking for
his/her opinion, approving his/her point of view or reecting his/
her feelings are important. But it depends on patients and
therapists characteristics. For example, if a patient is rather
anxious and somehow overwhelmed by the situation, it might not
be adequate to face him with open questions and to focus on his
emotions, but more benecial to give him the relevant information
concerning the current situation and the therapeutic strategy. It
also depends on the practitioners institutional role, related to
different tasks and goals. Salmon et al.s study on cancer surgeons
relationship [32] highlights the importance of the surgeons
expertise and character instead of emotional engagement for
building an authentic caring relationship.
Working alliance has mainly been developed and investigated
in psychotherapy. It has consistently been identied as the most
robust predictor of outcome across different therapeutic
approaches for a variety of patients [33]. Recent studies suggest
that both, the therapists skills and personal factors inuence the
working alliance with the patient; more specically, the quality of
communication skills [34,35], as well as the clinicians ability to
convey understanding of the patients phenomenological perspective [36,37] have been found to foster alliance. These results
highlight the need for a better understanding of the interactive
nature between technical (communicational skills) and relational
(interpersonal skills) dimensions of clinical relationship [38]. The
question of the validity of the concept of working alliance in the
eld of cancer care is an open question.
1.4. Objective and hypotheses
The aim of this study was to evaluate the impact of CST on
working alliance and to identify specic communicational
elements related to alliance for two professional groups, nurses
and physicians. Three hypotheses were tested:
Change in the working alliance. Alliance will be enhanced by CST;
alliance will be higher after training than in the control group.
Working alliance and verbal communication. The frequency of
communication behaviors directly related to working alliance
(more specically building a relationship and partnership
building, see Section 2.3.2) will be related to the level of
alliance.
Differences between physicians and nurses. As alliance is related
to the institutional role, no differences in the level of alliance
after training will be observed, but some specic communication behaviors (more biomedical communication for physician
and more psychosocial communication for nurses).

2. Methods
The sample study was based on a prepost controlled trial
including oncology clinicians who participated in CST (CST group)
and oncology clinicians who did not (CTRL group). The design and
the procedure of the study have been extensively described
elsewhere [22]. While a rst analysis of the sample focused on
clinicians defense mechanisms, a second grant allowed to
investigate alliance with the above mentioned aims.
2.1. Sample
One hundred and thirteen oncology physicians and nurses
participated in the study. For the physicians of the CST group, the
training was mandatory. For the CTRL group, oncology physicians
and nurses were recruited on a voluntary basis (these physicians
did not yet register for the CST). The group was formed after the
CST group and matched according to profession, age and gender.
The physicians and nurses of the CTRL group did not receive any
specic CST during their professional career.
In the CST group (N = 57), 43 participants (75.4%) were women,
30 (52.6%) physicians and 27 (47.4%) nurses, and mean age was
37.9 (SD = 7.2). In the CTRL group (N = 56), 35 participants (62.5%)
were women, 21 (37.5%) were physicians and 35 (62.5%) nurses,
and mean age was 39.4 (SD = 9.3). The sociodemographic variables,
such as years of professional experience, did not signicantly differ
between groups. The main physicians and nurses characteristics
(gender, age and years of professional experiences) are listed in
Table 1.
2.2. Study design
The CST consisted of a 2-day course, during which participants
are trained by means of case-history discussions, structured role
play and video analyses of simulated patient interviews, followed
by 46 individual supervisions over the next 6 months and another
half-day training session [1,2,11,22,3942].
The CST used in this study is part of a national CST, which has
been implemented over a decade ago [8]; meanwhile it has become
mandatory [42] and it has been demonstrated that it enhances
patient-centered communication [43,44] and it improves clinicians emotional regulation [1,2,22]. Specic features of alliance
were not part of the training, but the CST focuses on empathy,
careful listening, openness to the patients expression, shared
decision making and interest in the patients subjective and
psychosocial experiences [42]. These elements are related to
interpersonal skills, postulated to be key factors of alliance
building.
Each participant conducted two 15-min video-taped interviews
with simulated patients before and at the end of the CST. Short
written instructions were provided to the clinicians and actors
(simulated patients) prior to the interviews, specifying the type of
cancer, the age of the patient, the type of treatment and the
objectives of the interview. Participants of the CTRL group also
conducted two 15-min video-taped interviews, separated by a
6-month interval, with the same simulated patients, the same

Table 1
Characteristics of the clinicians.
Physicians

Nurses

Age
Gender (% women)
Experience (years)

SD

SD

39.0
82.0
14.7

8.5

38.4
53.0
10.7

8.0

8.0

8.3

C. Meystre et al. / Patient Education and Counseling 90 (2013) 233238

scenarios and the same instructions. The three actors and ve

scenarios were equally distributed in both groups. Participants
worked in various hospitals of the French part of Switzerland.
2.3. Instruments
2.3.1. Working alliance
Alliance was evaluated by the Working Alliance Inventory
Short Revised Form (WAI-SR) [45], observer version. The WAI-SR is
one of the most widely used instruments to measure working
alliance; it consists of 12 items rated on a 5-point Likert scale
which assesses three aspects of collaborative features: (i) patients
agreement on the tasks of the treatment, (ii) patients agreement
on the goals of the treatment and (iii) quality of the interpersonal
bond between patient and clinician.
Reliability and validity of the WAI-SR have been demonstrated
in psychotherapy research [45,46]. For this study, reliability testing
was based on 30 cases (15%) and revealed to be good with a mean
ICC(2,1) = .722 (range = .345.898). Pearson correlation between
the two instruments was very high (r = .906).
2.3.2. Verbal communication
Verbal communication was evaluated by means of the Roter
Interaction Analysis System (RIAS) [47]. The RIAS is a widely used
instrument, specically developed to assess physicianpatient
interaction and to describe the socio-emotional and task-oriented
elements of medical communication [48,49]. Each clinicians and
patients verbal utterance dened as the smallest distinguishable
speech segment to which a classication may be assigned is
attributed to one of 39 mutually exclusive codes, which are
aggregated into 12 interaction process categories and organized
into four medical interview functions [50]:
i. Gathering data: Open- and closed-ended questions.
ii. Patient education and counseling: Biomedical and psychosocial
information-giving, biomedical and psychosocial counseling.
iii. Building a relationship: Social talk, positive talk (agreement,
approval, laughter or jokes), negative talk (criticism or
disagreement), emotional responsiveness (reassurance, empathy, concern or legitimatization).
iv. Activating and partnering: Partnership building (asking for
patient opinion or understanding, checking for understanding,
paraphrasing, suggesting partnership), orientation.
According to the conceptual framework of working alliance and
the patient-centered approach, we hypothesized that the two
relational functions building a relationship and partnership
building were the communicational behaviors directly related
to the working alliance.
The reliability and validity of the RIAS are well established
[49,51,52]. In our study, inter-rater reliability, based on 20 cases
(9%), revealed to be excellent with a mean intraclass coefcient
ICC(2,1) = .952 (range = .916.994).
2.4. Coding procedure
The rst three authors were trained to use the WAI-SR and the
RIAS. Reliability checks were done with the trainers. The alliance
and verbal communication codings were done independently and
the two instruments were coded by different coders for each case.
2.5. Data analyses
Change in the working alliance (prepost) and the effect of
training (CST vs CTRL) were studied using general linear models
with repeated measures.

235

To study the relation between working alliance and verbal

communication, generalized linear mixed effect models were used
[53], considering alliance as the response variable, the RIAS
interaction process categories and three controlled variables (sex,
profession and actor) as predictors, with a random effect on the
subject level. The set of variable that best tted the model was
identied by means of the Akaike Information Criterion [54], which
is constructed using the evaluated likelihood at its maximum.
Consequently, variables were identied which are not signicantly
inuential on the response variable, but may help other variables
to explain a best proportion of the observed response; chosen by
the AIC, they remained in the nal model. These statistical models
are simple and exible enough to identify and interpret relatively
complex relations.
Differences between physicians and nurses were studied using
independent sample t-tests. Change in the working alliance (pre
post) was studied using general linear models with repeated
measures.
All analyses were computed using SPSS 15.0 and R 2.11.1.
3. Results
3.1. Change in the working alliance
For the CST group, the level of alliance was M = 46.50
(SD = 10.96) at time 1 and M = 49.96 (SD = 10.57) at time 2; for
the CTRL group, it was respectively M = 51.69 (SD = 9.10) and
M = 50.68 (SD = 9.45). General linear model showed no signicant
effect of time (F(1, 91) = .888, p = .349) and time  training group
(F(1, 91) = 1.191, p = .278).
There was an effect of the training group (F(1, 91) = 4.256,
p = .042). Alliance was higher for the control group at time 1
(t(94) = 2.541, p = .013) and there was no difference at time 2
(t(97) = .359, p = .720).
Finally, there was no difference in alliance between physicians
and nurses neither at time 1 (t(94) = 1.417, p = .160) nor at time
2(t(97) = .261, p = .795).
3.2. Working alliance and verbal communication
Table 2 shows the correlations between the two dependent
variables. First part of the Table 3 shows the maximum likelihood
estimates of parameters in the model by taking into account the
clinicians and patients utterances concurrently. The coefcients
Table 2
Pearson correlations between interaction process categories and working alliance.
Variables

Alliance
Pre (N = 96)

Interaction process categories (RIAS)

I. Gathering data
Open questions
Closed Question
II. Patient education and counseling
Psychosocial counseling
Biomedical counseling
Psychosocial information
Biomedical information
III. Building a relationship
Positive talk
Negative talk
Social talk
Emotional responsiveness
IV. Activating and partnering
Partnership building
*
**

p < .05.
p < .01.

Post (N = 99)

.108
.021

.023
.066

.297**
.019
.130
.121

.036
.210*
.148
.022

.108
.382**
.013
.123

.230*
.287**
.087
.005

.090

.076

C. Meystre et al. / Patient Education and Counseling 90 (2013) 233238

236

Table 3
Linear mixed-effects models of the interaction process categories explaining alliance.
Variables in equation

Interaction
Coeff

Positive talk
Negative talk
Biomed info
Psychosoc couns
Sex
Profession
Actors

Variables in equation
SE

Clinicians
Coeff

Variables in equation
SE

Patients

.091
.714
.060
.161

.031
.155
.023
.064

.004
.000
.010
.014

Negative talk
Biomed info
Psychosoc couns

1.667
.054
.201

1.133
.025
.069

.145
.037
.004

1.264
1.352
4.415

1.519
1.587
1.842

.407
.396
.019

Sex
Profession
Actor

.606
1.012
7.111

1.599
1.722
2.582

.705
.558
.007

Coeff
Positive talk
Negative talk

Question
Sex
Profession
Actors

SE

.124
.747

.046
.164

.008
.000

.409
1.949
.544
6.408

.139
1.527
1.524
2.133

.004
.205
.722
.003

Note. Coeff, coefcient; interaction, clinicians and patients are indicate the three separate models for individual contributions.

represent the estimated parameters of the linear model considering the working alliance as the response and the interaction
process categories as explanatory variables; the model is adjusted
for other variables.
To study the individual contributions, separate models for the
clinicians and patients utterances were tted (Table 3).
Table 4 displays pre- and post-training interaction process
categories for the CST and CTRL groups. Before training,
CST group showed less negative talk (t(76) = 3.358, p = .001)
and less psychosocial counseling (t(85) = 2.606, p = .011). After
training, CST group showed less social talk (t(95) = 2.705,
p = .008), less emotional responsiveness (t(100) = 2.445,
p = .016) and less orientation (t(111) = 2.222, p = .028) than
the control group.
3.3. Differences between physicians and nurses
Concerning change in working alliance, General linear model
showed no signicant effect of time (F(1, 41) = 1.573, p = .217),
time  professional group (F(1, 41) = .124, p = .726) and professional group (F(1, 41) = .842, p = .364).
Before training, physicians used more open-ended questions
(t(111) = 2.895, p = .005) and provided more biomedical information (t(111) = 3.231, p = .002) than nurses, whereas nurses
used more closed-ended questions (t(111) = 3.053, p = .003). After
training, physicians provided more biomedical information
(t(111) = 5.125, p = .000), and nurses used more closed-ended
questions (t(111) = 3.006, p = .003) and provided more psychosocial information (t(105) = 4.439, p = .000).

4. Discussion and conclusion

4.1. Discussion
Our hypotheses were only partially conrmed. CST did not
improve working alliance, but the level of alliance was related to
specic verbal communicational behaviors.
We showed no signicant improvement in alliance after CST.
The fact that the alliance was different before training (CST group
with a lower alliance) and similar after training (due to a non
signicant evolution of alliance for both groups) makes a
comparison too hazardous to test the rst hypothesis (does CST
improve alliance?). We assume that this initial difference in
alliance may be at least partially explained by selection bias: while
physicians and nurses were recruited on a voluntary basis for the
CTRL group, CST was mandatory for physicians and nurses in the
CST group. Volunteering for a study which implicates video-taped
clinicianpatient interviews indicates an interest in communicational issues and probably a certain condence in own interpersonal skills, which seems to be justied by these results.
The result that alliance did not improve after training reects
that CST did not focus on the complexity of the clinicianpatient
relationship and failed to improve clinicians working alliance with
the (simulated) patient [1,2].
With regard to specic communicational elements, we
observed relevant links between alliance and some interaction
process categories of the RIAS. The more positive talk and
psychosocial counseling in the interview, the higher the alliance
was rated. Thus, positive talk and utterances expressing agreement

Table 4
Interaction process categories for CST and CTRL groups.
CST

CTRL
Post

Pre

Social talk
Positive talk
Negative talk
Emotional resp
Part build
Orientation
Open Q
Closed Q
Biomed info
Psychosoc info
Biomed couns
Psychosoc couns

SD

SD

1.04
49.75
6.72
13.07
13.49
3.54
11.74
11.53
66.58
33.89
3.40
10.54

1.4
24.7
8.2
8.6
6.3
3.1
6.4
9.5
26.9
16.9
3.4
7.3

1.39
50.18
2.39
13.07
14.79
3.30
12.37
11.54
72.21
33.86
3.40
13.23

1.6
27.0
3.0
8.0
9.3
3.0
7.9
8.2
33.6
22.1
3.4
9.1

1.2
.1
3.6**
.0
1.1
.5
.5
.0
1.2
.0
.0
1.8

Pre

.23
.02
.70
.00
.16
.08
.09
.00
.18
.00
.00
0.33

Post

SD

SD

1.39
58.70
2.77
12.52
12.45
3.04
11.73
11.48
78.29
37.52
3.13
15.88

1.9
29.9
3.5
7.1
6.4
2.4
5.8
7.3
39.7
23.6
2.8
13.5

.70
49.00
3.34
9.89
13.41
2.20
9.82
12.27
65.07
36.14
2.80
13.16

1.0
27.0
3.7
5.6
7.5
2.2
6.1
8.9
35.9
19.3
4.0
10.6

2.4*
2.0
.9
2.4*
1.0
2.1*
1.8
.8
2.35*
.4
.5
1.6

.45
.34
.16
.41
.14
.36
.32
.10
.35
.06
.10
.22

Note: t, within group t-test; Emotional resp, emotional responsiveness; Part build, partnership building; Open Q, open questions; Closed Q, closed questions; Biomed info,
biomedical information; Psychosoc info, psychosocial information; Biomed couns, biomedical counseling; Psychosoc couns, psychosocial counseling.
*
p < .05.
**
p < .01.

C. Meystre et al. / Patient Education and Counseling 90 (2013) 233238

and approval, and statements and orientation related to lifestyle or

psychosocial issues contributed to alliance building. This illustrates that the clinicians task is not limited to biomedical
information exchange, but also includes agreement and issues,
such as repercussions on the family or work, emotional distress
and concerns or general advice.
The more biomedical information the clinicians provided, the
lower the alliance was rated; it is well known that medical
information exchange can be used to avoid psychosocial and
emotional aspects of disease [22], which hampers alliance building.
Negative talk criticism and disapproval was also related to
decreased alliance; this category belongs to the relational function
Building a relationship conrming its close links to alliance.
However, while criticism may express disagreement, it may also
serve to negotiate and reach an agreement, since mutual
afrmation is part of the concept of alliance [55]. We therefore
hypothesize that negative talk can only have a positive impact on
alliance when not surpassing a critical cut-off.
Patients questions were also negatively correlated with
alliance. Below is an example of an interview between an oncology
clinician and a (simulated) patient with breast cancer, which
illustrates what could be called the biomedical trap and why
patients questions have diminished alliance:
Patient:
Clinician:

Patient:
Clinician:
Patient:
Clinician:
Patient:
Clinician:
Patient:
Clinician:

[What is the percentage that Ill get over it completely?]

There is a very high percentage that you will recover about
three-quarters a very good probability that you will completely
get over it.
[And then I will have to have chemotherapy?]
Thats what we do after taking off the nodule to ght against the
cancer.
[But why chemotherapy if there arent any metastases?]
Its a security measure to be sure there isnt a little cell we
cant see with our medical analysis.
[But chemotherapy weakens you a lot?]
Yes it does chemotherapy is a heavy treatment with side effects
but[What are the side effects?]
[. . .]

This exchange illustrates that the patient is anxious about

prognosis and treatment and asks compulsively for more
information. The clinician answers by providing only biomedical
information which again triggers new questions by the patient
without addressing the emotions, leading to reassure the patient.
Finally, the results show that the actors had also an impact on
the working alliance. Of the three actors, one had a negative impact
on alliance, even if there is no signicant difference in mean
alliance between actors. Two different types of actors could be
distinguished: actors adapting their role according to the
physicians and nurses relational and communicational skills;
and actors who do not fully perceive the clinicians, acting
somehow rigidly. This difference in acting might have introduced
a bias that was not controllable. Those ndings illustrate the
interactive nature of the alliance building process where clinicians
and patients play an active role, and question results of studies
which did not take into account this variable.
After training, communication skills change more for the CTRL
groups less social talk, less emotional responsiveness, less
orientation and less biomedical information than for the CST
group less negative talk.
Literature showed small to moderate effect of CST on
communication behaviors in oncology with a large heterogeneity
between the studies [56]. Differences between the studies in the
initial percentages of communication behaviors have to be taken
into account. For instance, Langewitz et al.s [57] found 11.6% of
positive talk and 3.8% of closed questions before training in
comparison to respectively 22.1% and 5.1% in our study.

237

Percentages before training partially reect the clinicians initial

level of communication skills, and good communication skills
before training may involve less important changes. The fact that
patients were simulated may also partially explain the effect sizes.
Change is a variable difcult to apprehend and specic analyses
would be necessary to better explain this issue.
Finally, concerning our question about the differences in
communication between nurses and physicians, we observed that
physicians used more biomedical information both before and
after training, and that nurses provided more psychosocial
information after training. It has to be mentioned that the
quotient between medical and psychosocial talk is not addressed
during CST. It raises the issue of the role that nurses and
oncologists are willing to take when they care for oncology patient
in a multi-professional team and it has some implication for the
design of CST [57].
Some limitations of the study have to be mentioned. The CST
and the CTRL groups were not comparable, probably due to
selection bias; thus, the difference of alliance before training has
hampered the investigation of the evolution of alliance after CST. It
is possible that some participants of the CST group were less
motivated to improve their communication skills, while volunteers
of the CTRL group might have had an intrinsic interest in
communication. Another limitation concerns the simulated
patients; using actors has some advantages in research, particularly in terms of comparison between groups and time, but they do
not represent the variety of patients encountered in a daily clinical
practice. Moreover, the interview situation was a rst encounter,
which may explain the relative low alliance scores and it is well
known that alliance needs a time to develop [58].
Further research may address these limitations by studying real
patients, who have already build an alliance with the clinician,
especially in a context that may stress the alliance like breaking
bad news. To control for patient characteristics, each clinician
should see several patients. Alliance would be rated by both the
patient and the clinician and outcome measures should be
included to look for the allianceoutcome relationship. Concerning
training, it would be interesting to investigate the efcacy of
adding a specic module focused on the therapeutic alliance to CST
to see if alliance could be improved during CST.
4.2. Conclusion
The results underline the role of specic communication
behaviors and their impact on alliance building. Further studies
are needed to conrm these ndings, especially with real patients,
who the clinicians already know, to reect the clinical reality.
4.3. Implications for practice
Relational dimensions of the patient encounter and more
specically alliance are a key area of research as they have a direct
and high impact on communication and therefore on clinical
practice. The importance of interpersonal elements of communication for alliance building, such as positive talk and psychosocial
counseling, was demonstrated in our study. Since CST did not
improve alliance, their current form probably does not focus
enough on relational aspects of communication. Relational
elements of communication and alliance and their effects on
patients and clinicians still have to be further investigated in order
to improve clinical training and practice.
Author note
This study was supported by Oncosuisse, Grant OCS 01595-082004 and OCS 02338-02-2009.

238

C. Meystre et al. / Patient Education and Counseling 90 (2013) 233238

The authors do not have any interests that might be interpreted

as inuencing the research.
The authors wish to express their gratitude to Anne Weber for
her conscientious reading and editing.
References
[1] Stiefel F, Favre N, Despland JN. Communication skills training in oncology: it
works! In: Stiefel F, editor. Communication in cancer care. Berlin: Springer
Verlag; 2006. p. 1139.
[2] Stiefel F, Razavi D. Informing about diagnosis, relapse and progression of
disease-communication with the terminally III cancer patient. In: Stiefel F,
editor. Communication in cancer care. Berlin: Springer Verlag; 2006 . p.
3746.
[3] Lamont EB, Christakis NA. Prognostic disclosure to patients with cancer near
the end of life. Ann Intern Med 2001;134:1096105.
[4] Lerman C, Daly M, Walsh WP, Resch N, Seay J, Barsevick A, et al. Communication between patients with breast cancer and health care providers: determinants and implications. Cancer 1993;72:261220.
[5] Maguire P, Faulkner A, Booth K, Elliott C, Hillier V. Helping cancer patients
disclose their concerns. Eur J Cancer 1996;32A:7881.
[6] Falloweld L. Can we improve the professional and personal fullment of
doctors in cancer medicine? Brit J Cancer 1995;71:11323.
[7] Ramirez AJ, Graham J, Richards MA, Cull A, Gregory WM, Leaning MS, et al.
Burnout and psychiatric disorders among cancer clinicians. Brit J Cancer
1995;71:12639.
[8] Kiss A. Communication skills training in oncology: a position paper. Ann Oncol
1999;10:899901.
[9] Baile W, Kudelka AP, Beale EA, Glober GA, Myers EG, Greisinger AJ, et al.
Communication skills training in oncology: description and preliminary outcomes of workshops on breaking bad news and managing patient reactions to
illness. Cancer 1999;86:88797.
[10] Falloweld L, Lipkin M, Hall A. Teaching senior oncologists communication
skills: results from phase I of a comprehensive longitudinal program in the
United Kingdom. J Clin Oncol 1998;16:19618.
[11] Favre N, Despland JN, de Roten Y, Drapeau M, Bernard M, Stiefel F. Psychodynamic aspects of communication skill training for clinician in oncology: a pilot
study. Support Care Cancer 2007;15:3337.
[12] Levinson W, Roter D. The effects of two continuing medical education programs on communication skills of practicing primary care physicians. J Gen
Intern Med 1993;8:31824.
[13] Perle M, Maguire P, Heaven C. The development of a training model to improve
health professionals skills, self-efcacy and outcome expectancies when
communicating with cancer patients. Soc Sci Med 1997;44:23140.
[14] Falloweld L, Jenkins V, Farewell V, Saul J, Duffy A, Eves R. Efcacy of a cancer
research UK communication skills training model for oncologists: a randomised controlled trial. Lancet 2002;359:6506.
[15] Hart CN, Drotar D, Gori A, Lewin L. Enhancing parentprovider communication
in ambulatory pediatric practice. Patient Educ Couns 2006;63:3846.
[16] De Ridder DT, Theunissen NC, van Dulmen SM. Does training general practitioners to elicit patients illness representations and action plans inuence
their communication as a whole? Patient Educ Couns 2007;66:32736.
[17] Roter D, Larson S, Shinitzky H, Chernoff R, Serwint GA, Wissow L. Use of an
innovative video feedback technique to enhance communication skills training. Med Educ 2004;38:14557.
[18] Lamiani G, Furey A. Teaching nurses how to teach: an evaluation of a workshop
on patient education. Patient Educ Couns 2009;75:2703.
[19] Pieterse AH, van Dulmen AM, Beemer FA, Ausems MG, Bensing JM. Tailoring
communication in cancer genetic counseling through individual video-supported feedback: a controlled pretestposttest design. Patient Educ Couns
2006;60:32635.
[20] Timmermans LM, van der Maazen RW, van Spaendonck KP, Leer JW, Kraaimaat
FW. Enhancing patient participation by training radiation oncologists. Patient
Educ Couns 2006;63:5563.
[21] Schouten BC, Meeuwesen L, Harmsen HA. The impact of an intervention in
intercultural communication on doctorpatient interaction in The
Netherlands. Patient Educ Couns 2005;58:28895.
[22] Bernard M, de Roten Y, Despland JN, Stiefel F. Communication skills training
and clinicians defenses in oncology: an exploratory, controlled study. Psychooncology 2010;19:20915.
[23] Falloweld L, Jenkins V. Communicating sad, bad, and difcult news in
medicine. Lancet 2004;363:3129.
[24] Maguire P, Pitceathly C. Key communication skills and how to acquire them.
Brit Med J 2002;325:697700.
[25] Suchman AI, Matthews DA. What makes the doctorpatient relationship
therapeutic? Exploring the connectional dimension of medical care. Ann
Intern Med 1988;108:12530.
[26] Mead N, Bower P. Patient-centredness: a conceptual framework and review of
the empirical literature. Soc Sci Med 2000;51:1087110.
[27] Mead N, Bower P. Patient-centered consultations and outcomes in primary
care: a review of the literature. Patient Educ Couns 2002;48:5161.
[28] Fuertes JN, Mislowack A, Bennet J, Paul L, Gilbert TC, Fontan G, et al. The
physicianpatient working alliance. Patient Educ Couns 2007;66:2936.

[29] Rubin HR, Gandek B, Rogers WH, Kosinski M, McHorney CA, Ware JE. Patients
ratings of outpatient visits in different practice setting. Results from the
Medical Outcomes Study. J Amer Med Assoc 1993;270:83540.
[30] Van der Feltz-Cornelis CM, Van Oppen P, Van Marwijk HW, De Beurs E, Van
Dyck R. A patientdoctor relationship questionnaire (PDRQ-9) in primary care:
development and psychometric evaluation. Gen Hosp Psychiatry 2004;26:
11520.
[31] Safran JD, Muran JC. Negotiating the therapeutic alliance. New York: Guilford;
2000.
[32] Salmon P, Mendick N, Young B. Integrative qualitative communication analysis of consultation and patient and practitioner perspectives: toward a theory
of authentic caring in clinical relationships. Patient Educ Couns 2011;82:
44854.
[33] Horvath AO, Del Re AC, Fluckiger C, Symonds D. The alliance in individual
psychotherapy. Psychother Theor Res 2011;48:916.
[34] Kolden GG. Change in early sessions of dynamic therapy: universal process
and the generic model of psychotherapy. J Consult Clin Psychol 1996;64:
48996.
[35] Priebe S, Gruyters T. The role of helping alliance in psychiatric community
care: a prospective study. J Nerv Ment Dis 1993;181:5527.
[36] Castonguay LG, Goldfried MR. Psychotherapy integration: an idea whose time
has come. Appl Prev Psychol 1994;3:15972.
[37] Diamond GM, Liddle HA, Houge A, Dakof GA. Alliance-building interventions
with adolescents in family therapy: a process study. Psychotherapy
1999;36:35568.
[38] De Roten Y. Ce qui agit effectivement en psychotherapie: facteurs communs ou
agents speciques What really works in psychotherapy: common or specic
factors Bull Psychol 2006;59:58590.
[39] Stiefel F. Kommunikationstraining fur Onkologinnen und Onkologen in der
Schweiz Communication training for oncologists in Switzerland Psychosom
Konsiliarpsychiatr 2007;1:2725.
[40] Stiefel F, Rousselle I, Despland JN, Guex P. Ameliorer les competences communicationnelles: experience clinique et evaluation scientique Improving
communication skills: clinical experience and scientic evaluation Rev Med
Suisse 2006;2:3902.
[41] Stiefel F, Barth J, Bensing J, Falloweld L, Jost L, Razavi D, et al. Communication
skills training in oncology: a position paper based on a consensus meeting
among European experts in 2009. Ann Oncol 2009;21:2047.
[42] Stiefel F, Bernhard J, Bianchi G, Dietrich L, Hurny CH, Kiss A, et al. The Swiss
model. In: Kissane D, Bultz B, Butow P, Finlay I, editors. Handbook of
communication in oncology and palliative care. Oxford: Oxford University
Press; 2010. p. 6428.
[43] Langewitz W, Conen D, Nubling M, Weber H. Communication matters
decits in hospital care from the patients perspective. Psychother Psych
Med 2002;52:34854.
[44] Langewitz W, Szirt L, Nubling M, Weber H. Evaluation of the Swiss cancer
league communication skills program for oncologists and oncology nurses.
Final report of the research project; 2006.
[45] Hatcher RL, Gillaspy JA. Development and validation of a revised short version
of the Working Alliance Inventory. Psychother Res 2006;16:1225.
[46] Munder T, Wilmers F, Leonhart R, Linster HW, Barth J. Working Alliance
Inventory Short Revised (WAI-SR): psychometric properties in outpatients
and inpatients. Clin Psychol Psychother 2010;17:2319.
[47] Roter DL. Roter Interaction Analysis System (RIAS): coding manual. Unpublished manual, Baltimore: Department of Health Policy and Management,
Johns Hopkins School of Public Health; 1999.
[48] Roter DL, Larson S. The relationship between residents and attending physicians communication during primary care visits: an illustrative use of the
Roter Interaction Analysis System. Health Commun 2001;13:3348.
[49] Roter D, Larson S. The Roter Interaction Analysis System (RIAS): utility and
exibility for analysis of medical interactions. Patient Educ Couns 2002;46:
24351.
[50] Cohen-Cole S. The medical interview: the three function approach. St. Louis,
MO: Mosby; 1991.
[51] Ishikawa H, Takayama T, Yamazaki Y, Seki Y, Katsumata N, Aoki Y. The
interaction between physician and patient communication behaviors in Japanese cancer consultations and the inuence of personal and consultation
characteristics. Patient Educ Couns 2002;46:27785.
[52] Ong LM, Visser MR, Kruyver IP, Bensing JM, Van den Brink-Muinen A,
Stouthard JM, et al. The Roter Interaction Analysis System (RIAS) in oncological
consultations: psychometric properties. Psychooncology 1998;7:387401.
[53] Singer JD, Willett JB. Applied longitudinal data analysis: modeling change and
event occurrence. Oxford: Oxford University Press; 2003.
[54] Burnham KP, Anderson DR. Multimodel inference: understanding AIC and BIC
in model selection. Sociol Method Res 2004;33:261304.
[55] Saunders SM. Examining the relationship between therapeutic bond subscales
and the three phases of treatment outcome. Psychother Theor Res 2000;37:
20618.
[56] Barth J, Lannen P. Efcacy of communication skills training courses in oncology: a systematic review and meta-analysis. Ann Oncol 2011;22:103040.
[57] Langewitz W, Heydrich L, Nuebling M, Szirt L, Weber H, Grossman P. Swiss
Cancer League communication skills training programme for oncology nurses:
an evaluation. J Adv Nurs 2010;66:226677.
[58] de Roten Y, Fischer M, Drapeau M, Beretta V, Kramer U, Favre N, et al. Is one
assessment enough? Patterns of early alliance development and outcome.
Clin Psychol Psychother 2004;11:32431.