Patients’ preference for involvement in medical decision making:
A narrative review
Rebecca Say, Madeleine Murtagh, Richard Thomson*
School of Population and Health Sciences, Medical School, University of Newcastle, Framlington Place, Newcastle upon Tyne NE2 4HH, UK
Received 14 June 2004; received in revised form 31 January 2005; accepted 17 February 2005
Objective: This review aimed to clarify present knowledge about the factors which influence patients’ preference for involvement in medical
Methods: A thorough search of the literature was carried out to identify quantitative and qualitative studies investigating the factors which
influence patients’ preference for involvement in decision making. All studies were rigorously critically appraised.
Results: Patients’ preferences are influenced by: demographic variables (with younger, better educated patients and women being quite
consistently found to prefer a more active role in decision making), their experience of illness and medical care, their diagnosis and health
status, the type of decision they need to make, the amount of knowledge they have acquired about their condition, their attitude towards
involvement, and the interactions and relationships they experience with health professionals. Their preferences are likely to develop over
time as they gain experience and may change at different stages of their illness.
Conclusion: While patients’ preferences for involvement in decision making are variable and the process of developing them likely to be
which are consistent across studies.
Practice implications: By identifying the factors which might influence patients’ preference for involvement, health professionals may be
more sensitive to individual patients’ preferences and provide better patient-centred care.
# 2005 Elsevier Ireland Ltd. All rights reserved.
Keywords: Shared decision making; Patient participation
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Method. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1. Patients’ preferences for involvement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.Factors affecting patient preference: quantitative research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.1.Demographic variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.2.Experience of illness and medical care . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.3.Health status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.4.The type of decision. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.5.Information preference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.6.Limitations of these studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.Patient involvement in decision making: qualitative research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.1.Limitations of qualitative studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Patient Education and Counseling 60 (2006) 102–114
* Corresponding author. Tel.: +44 191 222 8760; fax: +44 191 222 8211.
E-mail address: firstname.lastname@example.org (R. Thomson).
0738-3991/$ – see front matter # 2005 Elsevier Ireland Ltd. All rights reserved.
4.Discussion and conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1.Limitations of this review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2.Implications for research. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3. Practice implications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Increasing emphasis is being placed on involving patients
inmaking decisions about their care. Thisisthe result of a
shifted towards a high prevalence of chronic disease there is,
for an increasing number of conditions, no overall best
themselves are in the best position to evaluate the trade-offs
between the benefits and risks of alternative treatments [3,4].
Patient expectations about their role in choice and
decision making have been influenced by living in a
media such as the Internet has become the norm for many.
Social movements, such as the women’s movement, have
emphasised the importance of autonomy and have actively
challenged medical authority .
Furthermore, highly publicised scandals and widely
reported concerns of under-funding have eroded patient
confidence in the NHS and the medical profession leaving
patients seeking more information and involvement in their
care [3,6]. Involving patients also helps meet demands for
accountability as health professionals can be more open
about decision making [3,6].
There are a number of roles for doctor and patient in
medical decision making. These have been conceptualised
as a spectrum with paternalism at one end, at the extreme of
which the doctor makes decisions on behalf of the passive
patient based on clinical expertise and without considering
the patient’s preferences. At the other end is informed
decision making, in which the doctor fully informs the
patient, detailing all treatment options and their implica-
tions, transferring technical expertise so that the patient can
make a decision alone, based on his or her own preferences
(active involvement) [5,7–9]. Shared decision making, in
which doctor and patient exchange information, both
detailing their treatment preferences, deliberating and then
deciding the treatment together, is in the middle of the
spectrum . An alternative model is that of an agency
relationship in which the patient fully informs the doctor of
his or her preferences and then delegates responsibility for
decision making to the doctor, who ideally would then make
an identical decision to that the patient would make, if the
patient had the clinical expertise of the doctor .
Whilst research has consistently shown that doctors
underestimate the amount of information that patients want
, it is less clear how much patients actually want to be
involved in making decisions about their treatment and what
influences their preference for involvement. This literature
review focuses on what influences patients’ desire to be
knowledge and identify further research opportunities.
Searches were carried out using Medline, Web of
Science, PsychINFO, CINAHL, The Cochrane Library
and HMIC (key words consumer participation, patient
participation, decision making, patient preferences, shared
decision making, patient involvement in decision making)
covering the years 1975–2003. The references of identified
articles, the indexes of journals from which articles were
retrieved, important texts about patient involvement in
decision making and key reviews were also searched.
Quantitative studies were included if they were primary
research investigating the factors affecting patient prefer-
ence for involvement in decision making published in
English. As this is an exploratory area of research all
identified studies were included but were rigorously
critically appraised by at least two of the authors to identify
methodological limitations and potential biases in order to
assess validity and reliability which are referred to
throughout this paper and summarised in Section 3.2.6.
Qualitative studies that focused on broader questions
were included if influence on decision making was a clear
focus of at least one aspect of analysis. Qualitative studies
were assessed in terms of the validity criteria for the
paradigm employed in the research, where that was clear in
the paper. Coherence of the research question or rationale
with the methods of data collection and analysis was
examined. Analysis was scrutinised for the appropriate
application of reliability and verification strategies. This
appraisal is summarised in Section 3.3.1.
Thirty three articles were identified; 25 quantitative
studies, seven qualitative studies and one mixed method
study (see Tables 1 and 2).
3.1. Patients’ preferences for involvement
Quantitative studies reported wide variation in the
proportion of patients who preferred to be involved in
making medical decisions (Table 1). Variation in preference
for involvement may be due to a number of methodological
R. Say et al./Patient Education and Counseling 60 (2006) 102–114103
R. Say et al./Patient Education and Counseling 60 (2006) 102–114 104
Summary of quantitative studies
Reference Method and instrument
used to measure decision
in decision making
Degner scale 
diagnosed < 6
1012 breast cancer
response rate 84%
Patients: active 12%,
shared 29%, passive
active 64%, shared
27%; passive 9%
Degner et al. Structured interviews;
Degner scale 
>6 months from
Beaver et al.  Structured interviews;
Degner scale 
150 breast cancer
< 4 weeks; 200
17 Stage I and 18
Stage II breast cancer
2–6 months, Canada
active 20%; shared
28%; passive 52%;
benign group: active
23.5%; shared 45.5%;
Age; social class
Hack et al.  Structured interviews;
Degner scale 
coping; health value;
of chronic illness;
marital status; income;
Survey; cross sectional;
baseline assessment of
medical outcomes study;
five-point Likert scale
measuring opinion on:
‘I prefer to leave
my medical care
up to my doctor’
2197 patients with
congestive heart failure
or depression, USA;
response rate 82%
severity of chronic
Ende et al. Questionnaire; cross
sectional; API 
312 general and
Mean score was
33.2 (S.D. 12.6)
Catalan et al. Questionnaire; cross
sectional; API 
60 gay and bisexual
HIV positive men, UK
Gibson et al. Questionnaire; cross
sectional; API 
85 patients buying
acute severe asthma,
212 moderate to
managed at two
scored 51 and
group scored 52
Age; severity of
Adams et al.  Postal survey;
modified API 
37% subjects indicated
a preference for greater
input than their physicians.
were stable over
of the impact of
asthma on their
lives; greater self
efficacy in management;
perception of physicians’
propensity to involve
them in decision
of a written plan
adverse effects of
active coping strategies;
making style; cost
concerns that caused
delays in seeking care;
R. Say et al./Patient Education and Counseling 60 (2006) 102–114 105
Table 1 (Continued)
ReferenceMethod and instrument
used to measure decision
in decision making
et al. 
API ; desire
459 health maintenance
by age, USA
API vignettes vs.
For API: age;
API ; health
opinion survey 
167 outpatients with
pain or mild
median HOS 0.36
With API: age;
severity of condition
in vignettes; for HOS:
age; sex; education;
annual family income;
Decision making or
severity of illness;
from API  and
the health opinion
64.1% had low
scores on health
did not wish to
had medium scores
and 16.0% had
O’Dell et al.  Questionnaire;
160 men aged
in an RCT of
making in prostate
Strull et al. Self-administered
designed for this study
but to strongly
consider the patient’s
opinion; 19% shared
Cassileth et al.  Structured interviews;
for this study
256 cancer outpatients
and inpatients with
Age; sex; being
of involvement in
Age when controlling
Blanchard et al.  Observation and
et al. 
USA; response rate
representing 89 patients
69% stated they
would prefer to
Sutherland et al.  Questionnaire;
from health opinion
survey , Cassileth
 and Strull et al. 
R. Say et al./Patient Education and Counseling 60 (2006) 102–114 106
Table 1 (Continued)
Reference Method and instrument
used to measure decision
in decision making
Butow et al.  Questionnaires;
Sutherland et al. 
80 cancer outpatients,
rate 90.0% after first
change to their
changed with: first
visit; routine follow-up
or follow-up visit with
significant change to
Being a patient;
Sutherland et al. 
outpatients; 53 people
53 surgical outpatients
conditions; 36 people
rates: 91.2, 70.9, 88.3
29.7%; shared 24.3%;
passive 46%; surgical
patients: active 23.0%;
shared 25.0%; passive
active 21.4%; shared
42.9%; passive 35.7%
Sex; group; age
Wong et al.  Questionnaire; cross
and interviews with
101 prostate cancer
Stewart et al.  Questionnaire; cross
105 ovarian cancer
care treatment and
14.3%; shared 62.9%;
passive 22.9%; treatment:
shared 59.6%; passive
23.1%; after: autonomous
15.2%; shared 61.9%;
If condition was
perceived as very
serious or if
McKinstry Structured interview;
approaches in video
410 adult patients
from five general
response rate 65%
Age; smoking status;
social class; scenario;
perception of own
Age; smoking status;
social class; scenario
Mansell et al. Structured interview;
postal survey and
chart review; cross
Likert scale for who
should make decisions
relating to vignettes
physician alone –
Type of decision;
severity of heart
disease when making
decisions about future
Severity of heart
disease when making
decisions about future
Davis et al.  Questionnaire;
visual analogue scale
Mean VAS score ranged
from 3.3 to 5.1 for the
different triage groups
(0: physician decision
alone, 5: doctor and
patient equally and
10: patient decision
R. Say et al./Patient Education and Counseling 60 (2006) 102–114 107
Table 1 (Continued)
ReferenceMethod and instrument
used to measure decision
in decision making
Rosen et al.  Postal survey; cross
should be made the
doctor, shared (with
doctor having final
vote), shared (with
patient having final
vote) or patient
from 13 general
94% wanted some
Anell et al. Postal survey;
opinion on ‘If
there are alternative
therapies, my physician
should decide which
alternative to choose’
aged 61+ and
living with spouse
County of residence;
living in an urban area
Summary of qualitative studies
ReferenceData collection Participants and
Analysis Reliability/validityKey findings
et al. 
conducted on ward
18 male and 16 female
audit trail; reflexivity
Decision making promoted
by good health, access to
sufficient time since diagnosis,
education, good interactive
relationships with nurses
and physicians, encouragement
by nurses and physicians to
participate. Barriers to decision
making: poor health, lack of
age, time and resource pressures
on staff, high staff turnover,
poor interactive relationships
Active patients prefer a
participative role for the
sense of control. Barriers to
participation: mental infirmity,
lack of education, difficulty
accepting diagnosis. Preference
for active or passive role involved
faith in the physician, will to live,
general coping style. Preference
for information dependent on:
educational status, personality,
coping style, illness severity
Patients concerned about ‘doing
what is right’, pleasing the nurses.
Patient trust and confidence based
on ‘putting yourself in their hands’.
Some patients do not want the
responsibility of involvement
et al. 
17 Stage I and 18
Stage II breast
factors as well as reflecting real variations. For example,
different instruments may be measuring different constructs
or have differing reliability or validity.
Decision making preference was measured in a variety of
ways although many of these measures were informed by or
derived from each other (see method column in Table 1
where the instruments used are listed and cross-referenced).
For example, six studies reference the Degner Scale ,
namely, Degner and Sloan , Degner et al. , Beaver
, and six reference the Autonomy Preference Index ,
i.e., Ende et al. , Catalan et al. , Gibson et al. ,
R. Say et al./Patient Education and Counseling 60 (2006) 102–114 108
Table 2 (Continued)
ReferenceData collectionParticipants and
AnalysisReliability/validity Key findings
over 12 months;
22 patients with a
diagnosis of Type 1
diabetes (min 15
years). Nominated for
selection by clinicians.
Setting not described
Engagement in decision making is
a developmental process enhanced
by information, development of
personal expertise; relationship with
health care professional. Progression
from passive to active roles requires a
shift in levels of responsibility and
understanding of the role of health
professionals. Traditional compliance
models inhibit patient involvement in
decision making. Patients preferring
self-care methods did not want loss
of control. Patients preferring may
want a break from responsibility in
some circumstances: bereavement,
sickness, family crisis.
Social class, education and age or
generation continue to shape the
medical encounter. Gender and
ethnicity not found to be influential.
Professionals wanted to be treated
as an equal, a ‘fellow professional’.
Rationality and autonomy privileged
in a consumerist society, therefore
dependency viewed as weakness
Inhibiters of preference for involvement:
lack of understanding of the potential
role in participation; lack of knowledge
and experience of potential consequences;
fear of making a ‘wrong’ decision/blame
or responsibility for consequences; trust
in doctors. Enhancement of preference for
decision making: desire for responsibility
over own body; desire to have views/values
considered; recognise that the doctor is
fallible. Participation not a meaningful
option for those who have no knowledge
or expectation of anything but a passive
role. Participation is a process – requires
support to identify preferences and to
anticipate how these can change over time
Women wanted some form of shared
decision but felt they lacked the expertise,
knowledge and clinical experience. Choice
can only be made where there are
meaningful options. Participation
promoted a sense of control. Women
viewed treatment decision as highly value
laden (right or wrong) and in a shared
decision making process a dilemma occurs
in confusion about how to allocate blame
should a decision be wrong. Some patients
had high preference for information but
not for participation in decision making
60 lay people living
in Sydney, Australia
et al. 
40 women 1 year
after first treatment for
Stage I or II breast
reliability – double
review of interpretations
by two researchers
(one experienced in
et al. 
20 women with early
stage breast cancer
case load of two
Audit trail, independent
review of interpretation
by experienced qualitative
of representative quotes
Adams et al. , Thompson et al.  and Nease and
Brooks . Hence there may not have been so much
variation between questionnaire designs as appeared.
However, exact methods were rarely explicit; when various
instruments were referenced it was difficult to assess exactly
how they were used, particularly as results were presented in
a variety of ways which were not always directly
comparable between studies. For example, proportions of
people with a particular preference defined in categories
such as ‘passive’, ‘shared’ and ‘active’ are difficult to
compare with mean scores on an ordinal scale (see summary
of preferred roles in decision making column in Table 1).
Alternatively, differences between the differing groups
may be explained by the patients involved in the studies or
the actual decisions involved, since studies involved
different nationalities or people with different experience
of illness or health care. For example, perhaps the
preferences of men with HIV are influenced differently
from those of women with breast cancer.
3.2. Factors affecting patient preference:
A variety of socio-demographic variables and variables
relating to health care experience, health status, the type of
decision and information preference have been found to be
associated with decision making preference (Table 1).
3.2.1. Demographic variables
Twenty two studies investigated the association between
age and decision making preference; 17 found an associa-
tion, all describing that younger patients preferred a more
active role in medical decision making than older patients
(see Table 1). Fiveout of eight studies identified that sexwas
associated with preference, all finding that women were
more likely than men to prefer a more active role. Sixteen
outof20studiesdescribedanassociation between education
and preference, all finding that people with higher education
preferred more active involvement.
Four out of 10 studies found that marital status was
associated with decision making preference. One showed
that unmarried people were more likely than married people
to prefer to participate, but this was a univariate analysis
. Another study found that separated or divorced people
had a stronger preference for involvement than married
people but weaker than people who had never been married
, whilst another found that married women were more
likely to prefer an active or shared role, but this was not an
independent predictor when analysed by multivariate
methods . Confounding by variables such as sex and
age may occur; for example, Arora and McHorney found
that while marital status had a significant univariate
association with decision making preference it was not
significant in a multivariate model . Therefore, the
association between marital status and decision making
preference is unclear.
Class, occupation, and income have all been associated
with variation in preference for involvement in medical
classes to prefer a more passive role [14,17,22,25,33]. Only
one study found no association . Again confounding is an
issue; for example, Arora and McHorney found that income
failed to reach significance in a multivariate model .
Ethnicity was associated with decision making prefer-
ence in one out of seven studies, with white patients more
likely to prefer to be involved than black patients using
multivariate analysis .
3.2.2. Experience of illness and medical care
Various studies found that experience of illness and
medical care may affect patients’ preferences for involve-
ment in decision making. Some studies found that
experience of being a patient was associated with a
decreased preference for involvement. Stiggelbout and
Kiebert compared cancer and surgical patients with people
accompanying them to clinics using hypothetical vignettes
and found that while there was no difference in preference
between the two patient groups, companions preferred a
more active role than patients . They suggest that this
might be a feature of being a patient per se, which is exerting
a stronger influence than the type of decision and wonder if
this is related to the sick role, rather than experience of care
itself. Degner and Sloan compared newly diagnosed cancer
patients to householders and found that while 64% of the
householders said they would prefer to play an active role in
treatment decision making if they were to develop cancer,
only 12% of newly diagnosed cancer patients preferred an
active role . Thompson et al. found that hypertensive
patients were less likely than non-hypertensive patients to
indicate a preference for involvement in a vignette about
However, the results of some studies suggest that
patients’ preference may change as their experience of
their illness develops. Butow et al. found that cancer patients
attending their first consultation were more likely to seek
greater involvement in decision making that those attending
a follow-up . In contrast, Degner et al. found that breast
cancer patients who had been diagnosed for less than 6
months were less likely to prefer an active role than those
who had been diagnosed for more than 6 months  and
Mansell et al. found that patients with prior experience of
severe heart disease wanted more involvement in decisions
about future myocardial infarction than those who had had
no or mild-to-moderate heart disease . Butow et al.
found that patients were more likely to change over time
towards a greater desire for involvement if they were having
a routine follow-up than if their follow-up involved
discussing a significant change to their condition .
It is difficult to disentangle experience of care from
experience of ill health in these studies. However, other
studies have shown that experiences of interactions with
health professionals may also affect patients’ desire to be
R. Say et al./Patient Education and Counseling 60 (2006) 102–114 109
involved in decision making. Adams et al. found that the
greater the patient’s perception of their own physician’s
propensity to involve them in decision making the more
activea role the patient preferred . McKinstry found that
generally patients described their own doctor as having the
same style as they preferred .
Blanchard et al. identified physician behaviours that were
associated with cancer patients having a greater desire to
participate: using the patient’s first name; attempting to
establish privacy for physical examination; examining the
patient on their trunk; and discussing test results with the
patient . Those patients who preferred to be involved
were also more likely to perceive that the doctor: discussed
future tests or treatment; discussed test or treatment results;
gave good news; discussed discharge; discussed side effects
of treatments; and performed a physical examination.
3.2.3. Health status
Research has also shown that diagnosis and health status
may affect patients’ preference. Some studies have shown
that patients with severe illness may be less likely to prefer
involvement than patients who are less ill. Arora and
McHorney, using data from 2197 patients with chronic
diseases, found that patients with severe diabetes or heart
diseasewere less likely than patients with mild hypertension
to prefer an active role . Catalan et al. found that
asymptomatic HIV patients were more likely to prefer an
active role than symptomatic patients . Similarly
Blanchard et al. found that patients who were bed-ridden
notbed-ridden .Beaver et al.compared women who had
recently been diagnosed with breast cancer to those with
benign breast disease and found that the benign group were
more likely to want a shared or active role, whereas recently
diagnosed breast cancer patients were more likely to prefer a
passive role .
In contrast, Stewart et al. found that women of all ages
with ovarian cancer were more likely to prefer shared
decision making if their condition was very serious, or if
they had metastases, than if their condition were less serious
3.2.4. The type of decision
By comparing responses to different hypothetical clinical
situations studies have also demonstrated that patients’
preferences may be affected by the type of decision they are
making. Participants have been found to prefer active roles
in decisions about minor illness, such as upper respiratory
tract infections [17,22], but to prefer passive roles in
decisions about physical problems, such as a sprained calf
, more serious illnesses such as myocardial infarction
[17,22] or during severe exacerbations of a condition
[19,20]. Patients were also found to prefer an active role in
behavioural decisions, such as giving up smoking , and
in making major decisions such as whether or not to have a
coronary artery bypass graft .
Factors relating to separate components of the decision
making process may also affect patients’ preference. Deber
suggested that medical choices consist of ‘problem solving’
and‘decisionmaking’elements .Problem solvingrefers
to the ‘search for the single correct solution to a problem’,
which requires expertise, whereas decision making refers to
choosing between alternative treatments, which requires the
contribution of patients’ values and preferences . Deber
et al. designed a questionnaire which aimed to separate the
problem solving and decision making elements and found
that patients preferred to be involved in decision making
elements, but preferred the doctor to perform the problem
solving tasks . Thompson et al. also found that patients
preferred to be involved in decisions that did not require
medical knowledge more than in those that did .
3.2.5. Information preference
Preferenceformoreinformation hasbeen associated with
preference for a more active role in decision making
[18,26,28]. This might be associated with having an active
coping style which was also found to be associated with
preference for involvement [16,20].
3.2.6. Limitations of these studies
With the exception of one study , all quantitative
studies were based on single cross sectional study design.
Causation cannot therefore be reliably assigned to observed
associations. For example, it is unclear whether patients
prefer the same role in decision making as they experience
with their doctor because clinician style affects their
preference or because they chose doctors with their
preferred style. Butow et al.  only followed patients
longitudinally between two consecutive consultations, so
had limited ability to demonstrate changes in preference
over time and the course of an illness.
Many studies had relatively small sample sizes, and
random sampling was rare leading to potential selection
bias, particularly when convenience samples were used.
Some samples were too homogenous to identify particular
associations. For example seven studies had single sex
samples and even more had uneven sex distributions with
insufficient numbers of either men or women to have power
to find a significant association between decision making
preference and sex. Response rates ranged from 43.2  to
91.2%  so poorer response rates may also have
introduced bias (see Table 1).
Studies undertaken with specific groups of patients, in
particular contexts may have limited generalisability. For
example, the preferences of patients with cancer may not be
shared by patients with non-malignant conditions and in-
patients may have different preferences from out-patients.
Results of studies performed in other countries may not be
relevant to a UK population.
There are also limitations associated with measuring
different variables. For example there is no entirely
satisfactory way of measuring decision making preference.
R. Say et al./Patient Education and Counseling 60 (2006) 102–114 110
Instruments requiring patients to choose between a small
number of statements may not reflect the complexity of
decision making preference. The use of hypothetical
vignettes and indeed theoretical preferences may not relate
to patients’ preferences in real consultations.
Often little detailed information was provided about how
specific variables were measured and different measurement
techniques may have led to variation in results. This may be
particularly important for more subjective variables such as
be questionable, particularly hard to measure personality
variables such as optimism measured by Wong et al. .
Only a few studies extended beyond univariate analysis,
thus there are considerable limitations in the interpretation
of the results in the face of potential confounders.
Furthermore, studies that have used multivariate analysis
to identify independent predictors of preference have
consistently been able to explain only a small part of the
observed variation by the identified predictors; for example
socio-demographic and diseasevariables have been found to
accountfor as little as 6.9% of thevariance in preference .
This is likely in part to reflect the problems of incorporating
some of the more complex factors demonstrated by
qualitative studies (see below) into quantitative surveys,
but also a failure to incorporate a full enough range of
variables into individual studies.
3.3. Patient involvement in decision making:
Qualitative research is useful for exploring patients’
attitudes, beliefs and preferences to develop a more in-depth
understanding than is possible in quantitative research .
Qualitative studies demonstrated the complexity and
making and the importance of process and relationships in
decision making (Table 2).
A number of factors were found to affect patients’
preferences for involvement in decision making, many of
which were similar to, and elaborated on, those identified in
the quantitative studies. Sainio et al. developed a list of
‘preconditions for patient participation’ from focused
interviews with cancer patients. These consisted of
promoting factors (good health; adequate information;
ability to confront situations; desire to participate; activity,
consisting of personality, asking questions and acquiring
information; courage; self-interest; and availability of a
supportperson) and restrictingfactors(poor physical/mental
health leading to fatigue and an inability to receive
information; ignorance; fear of future, reality or authorities;
increasing age; and lack offaith in their own influence) .
Other studies attempted to develop a greater under-
standing of factors which might be pre-requisites for
involvement, such as having knowledge and experience of
the illness; understanding the role patients could play in
decision making; their attitude towards decision making;
their attitude towards their illness, and the interactions they
had with health professionals.
Charles et al. found that, while most participants wanted
some form of shared decision making, many felt that
decisions ‘required expertise, knowledge and clinical
experience that they did not have’ . This would be
consistent with quantitative findings that patients prefer
more involvement in hypothetical decisions where less
medical knowledge is required.
Understanding information was also important. Patients
and to make choices when they have no experience of the
potential consequences . This suggests that the ability to
access and understand medical information may affect
patients’ preferred role in decision making and provides a
possible explanation for why those patients who prefer more
information also prefer an active role in decision making.
Patients may prefer different roles in decision making at
different stages of their illness. Thorne and Paterson found
that illness trajectory was very important in developing
expertise in making decisions about diabetes self-care .
Theyidentified that, in order to become an expert in diabetes
self-care management, it was important to develop ‘an
intricate and sophisticated ability’ to respond to one’s own
body and adapt medical advice accordingly and therefore
patients expected different levels of support from health
professionals as they moved through different stages of
responsibility for managing their condition. This dynamic
relationship promoting involvement in treatment decision
making is reflected in recent health policy . Hack et al.
found that difficulty in accepting a diagnosis of cancer was
one reason women gave for not actively participating in
decision making . This offers a possible explanation for
why patients’ preference for involvement may increase with
experience of an illness.
Patients’ understanding of medical decision making
appeared to affect their preference for involvement. Kenny
etal., insemi-structured face to face interviews with patients
with early stage breast cancer, identified that the patient
must have a realistic understanding of the role they can play
in decision making and also that they should appreciate that
decisions are often based on value judgements as well as
medical knowledge .
‘right’ or ‘wrong’ outcomes may also affect their preference.
often discourages patients from participating [46,47]. Charles
allocate blame for negative outcomes of a shared decision
been made by the doctor alone . Recognition that doctors
are not infallible was suggested by some patients as a reason
for being involved , whereas having faith and trust in
doctors may be associated with preference for a passive role
[15,46,47]. Involvement in decision making has been
identified to give women a feeling of control [15,47].
R. Say et al./Patient Education and Counseling 60 (2006) 102–114 111
Patients’ attitude towards their illness may also affect
their preference. ParticipantsinHack etal.’sstudy suggested
thatthewill tolivewouldaffect preference.Kennyetal.
found that some participants believed it was important to
take responsibility for one’s own body . Therefore
patients’ attitudes towards decision making and their illness
may influence their preference for involvement.
Interactions with health professionals may also affect
patients’ preferences. Sainio et al. identified that a good
relationship with nurses and physicians; the patient taking
initiative, asking questions and an active approach;
provision of information; encouragement to participate;
and the intermediary role of the nurse in passing information
to physicians promoted involvement and that a poor
relationship with nurses and physicians who were ‘snappy
or arrogant’; incompatible personalities; objectification of
patients; lack of time leading to time pressure and short
appointments; inadequate information; and staff turnover
restricted involvement .
Clinicians’ decision making styles may also affect
preferences. The major theme of Waterworth and Luker
was of patients ‘toeing the line’ . They suggest that
patients may respond to staff promoting their involvement in
order to behave as they perceive staff want, rather than
because they truly want to be involved, hence patients were
‘reluctant collaborators’ .
Patients’ perceptions of themselves in relation to the
doctor may also be important. For example, Lupton
describes how a 59-year-old partner in a law firm described
his need to position himself as a ‘‘fellow professional’’ .
She also found that less well educated people of lower
socioeconomic status had greater levels of respect for
doctors. This provides a possible explanation for the
associations between better educated patients preferring a
more active role and patients of low social class preferring a
3.3.1. Limitations of qualitative studies
There were a number of limitations to the qualitative
studies. Some studies gave little detail about who the
participants were and how they were selected. For example,
Waterworth and Luker used a convenience sample of
patients from three mixed-sex medical wards but did not
state the numbers of men or women or what types of
condition theyhad had .Since the rationale for sampling
was not clear it is not possible to judge how the setting or
participants’ characteristics may have affected the results
Methods were frequently unclear, making it difficult to
analytic methods. For example, Lupton states that she
analysed for recurring discourses but no details were given
on how this was done . Therefore it is not always
possible to assess how systematic the analysis was.
Researchers were rarely explicit about the paradigm they
were working in and therefore it was not always possible to
assess the assumptions that they made and thereby the
validity of interpretations. Validity was rarely considered.
While some did involve analysis by more than one
researcher, few looked for disconfirming cases and two
studies did not include extracts of interview transcripts
[15,43]. Therefore it was difficult to assess the validity and
rigour of the analysis.
While some studies provided useful insights, others
provided little depth beyond that possible in quantitative
research. For example, Hack et al.  drew up lists of
factors which may affect patients’ preferences without
offering any explanation for why associations might exist.
This limits the usefulness of the research since little in-depth
understanding is developed.
4. Discussion and conclusion
In the context of socio-political change patient involve-
ment in decision making has become an important issue for
health professionals and an understanding of and respon-
siveness to individual patient’s preferences is important in
improving the quality of care provided.
Patients’ preferences for involvement in decision making
are clearly variable and the process of developing them is
likely to be highly complex. This review has identified a
number of influences on patients’ preference for involve-
ment in medical decision making, some of which are quite
consistent across studies. Quantitative studies demonstrate
that demographic factors affect patients’ preferences:
a more active role in decision making.
Experience of illness and medical care is also likely to
affect patients’ preferences.Involvement indecision making
may be a process and patients’ preferences are likely to
they are making is likely to affect their preference and this
may relate to the amount of knowledge they have or the
amount of knowledge that they perceive is necessary to
make the decision. Patients’ diagnosis and health status,
along with their attitude towards their illness is likely to be
important. This may reflect the social meaning of different
illnesses (for example, that cancer is life threatening) and
how well known the illness is to the patient. It may be more
difficult to involve patients in making decisions about
illnesses which they are unfamiliar with or which have well-
developed psycho-social constructs.
Participants’ attitude toward decision making, and the role
in it they can have, also affects their preference. Patients
identify a moral imperative in decision making: that there are
a sense of vulnerability and of what they could lose or a
perceived lack of knowledge and experience. Patients did not
appear to recognise that in many circumstances evidence
R. Say et al./Patient Education and Counseling 60 (2006) 102–114 112
importance of having someone to blame if the outcome of
treatment is unsatisfactory may also be significant and
contrasts with the sense of empowerment gained when being
involved in making one’s own decisions.
Participants’ relationship with their clinician and the
interactions they have had with health professionals are also
important. It is unclear whether clinicians’ decision making
style affects patients’ preferences or if patients seek out a
clinician whose style matches their preference. Training
health professionals ininvolvingpatients indecision making
may be useful in facilitating patient involvement.
4.1. Limitations of this review
While every effort was made to search for articles
systematically and thoroughly, this is not a systematic
review. We retrieved all identified articles that were
available from a wide variety of journals and included both
quantitative and qualitative work. There is, however, a bias
towards work that is published in English.
Since all the studies investigated slightly different
variables in different groups of patients it was not possible
to undertake a combined quantitativeanalysis. Nevertheless,
this review provides an up to date discussion of what
influences patients’ preferences for involvement in decision
making and identifies areas where further work would be
4.2. Implications for research
studies have been demonstrated and the body of research in
this exploratory area is small. Therefore, further research
which overcomes these limitations would be useful in
developing a better understanding of patients’ preferences.
Further quantitative research, employing multivariate
methods and following patients over time, would be useful
in overcoming confounding, to develop a better under-
standing of causation and to better understand how patients’
preferences might change over time and with experience.
Further qualitative research would be particularly useful
at patient understandings of involvement in decision
patients’ preferences such as the effect of the doctor–patient
interaction in the process of decision making.
4.3. Practice implications
Involving patients in decision making is one of many
ways by which the quality of health care can be improved
. This is reflected in the association between patient
involvement in decision making and improvements in health
outcomes. Positive effects have been reported, in various
groups of patients, on a diverse range of outcomes [51–54].
Involving patients is also thought to improve concordance
[55,56] and increase patient satisfaction .
Nonetheless, preference for involvement in medical
decision making varies widely. Involvement may not be
acceptable or appropriate for everyone. By identifying the
factors which might affect patient preference for involve-
ment, health professionals may be more sensitive to
patient-centred care. This might include developing ways
to empower patients so that they may experience the benefits
associated with involvement.
 Bowling A, Ebrahim S. Measuring patients’ preferences for treatment
and perceptions of risk. Qual Saf Health Care 2001;10:2–8.
 Gillespie R, Florin D, Gillam S. Changing relationships: findings from
the patient involvement project. King’s Fund; 2002.
 Charles C, Gafni A, Whelan T. Decision-making in the physician-
patient encounter: revisiting the shared treatment decision-making
model. Soc Sci Med 1999;49:651–61.
 Holman H, Lorig K. Patients as partners in managing chronic disease.
Br Med J 2000;320:526–7.
 Charles C, Whelan T, Gafni A. What do we mean by partnership in
making decisions about treatment? Br Med J 1999;319:780–2.
 Coulter A. After Bristol: putting patients at the centre. Br Med J
 Elwyn G, Edwards A, Kinnersley P. Shared decision-making in
primary care: the neglected second half of the consultation. Br J
Gen Pract 1999;49:477–82.
 Robinson A, Thomson R. Variability in patient preferences for parti-
cipating in medical decision making: implication for the use of
decision support tools. Qual Saf Health Care 2001;10:34–8.
 Charles C, Gafni A, WhelanT. Shared decision-making in the medical
encounter: what does it mean? (or it takes at least two to Tango). Soc
Sci Med 1997;44:681–92.
 Gafni A, Charles C, Whelan T. The physician–patient encounter: the
physician as a perfect agent for the patient versus the informed
treatment decision-making model. Soc Sci Med 1998;47:347–54.
 Guadagnoli E, Ward P. Patient participation in decision making. Soc
Sci Med 1998;47:329–39.
 Degner LF, Sloan JA. Decision making during serious illness: what
J, Bilodeau B, Watson P, Mueller B. Information needs and decisional
preferences in women with breast cancer. J Am Med Assoc
 Beaver K, Luker KA, Owens RG, Leinster SJ, Degner LF, Sloan JA.
Treatment decision making in women newly diagnosed with breast
cancer. Cancer Nurs 1996;19:8–19.
 Hack TF, Degner LF, Dyck DG. Relationship between preferences for
decisional control and illness information among women with breast
cancer: a quantitative and qualitative analysis. Soc Sci Med
 Arora NK, McHorney CA. Patient preferences for medical decision
making: who really wants to participate. Med Care 2000;38:335–41.
 Ende J, Kazis L, Ash A, Moskowitz MA. Measuring patients’ desire
for autonomy: decision making and information-seeking preferences
among medical patients. J Gen Intern Med 1989;4:23–30.
 Catalan J, Brener N, Andrews H, Day A, Cullum S, Hooker M,
Gazzard B. Whose health is it? Views about decision-making and
R. Say et al./Patient Education and Counseling 60 (2006) 102–114 113
information-seeking from people with HIVinfection and their profes- Download full-text
sional carers. AIDS Care 1994;6:349–56.
 Gibson PG, Talbot PI, Tonneguzzi RC. Self-management, autonomy
and quality of life in asthma. Chest 1995;107:1003–8.
 Adams RJ, Smith BJ, Ruffin RE. Patient preferences for autonomy in
decision making in asthma management. Thorax 2001;56:126–32.
in medical decision-making: situational and demographic influences.
Patient Educ Couns 1993;22:133–40.
 Nease RF, Brooks WB. Patient desire for information and decision
making in health care decisions: the autonomy preference index and
health opinion survey. J Gen Intern Med 1995;10:593–600.
 Deber RB, KraetschmerN, Irvine J. What role do patientswish to play
in treatment decision making? Arch Intern Med 1996;156:1414–20.
 O’Dell KJ, Volk RJ, Cass AR, Spann SJ. Screening for prostate cancer
with prostate-specific antigen test: are patients making informed
decisions. J Fam Pract 1999;48:682–8.
decision making? J Am Med Assoc 1984;252:2990–4.
 Cassileth BR, Zupkis RV, Sutton-Smith K, March V. Information and
participation preferences among cancer patients. Ann Intern Med
 Blanchard CG, Labrecque MS, Ruckdeschel JC, Blanchard EB.
Information and decision-making preferences of hospitalized adult
cancer patients. Soc Sci Med 1988;27:1139–45.
 Sutherland HJ, Llewellyn-Thomas HA, Lockwood GA, Tritchler DL,
Till JE. Cancer patients: their desire for information and participation
in treatment decisions. J Roy Soc Med 1989;82:260–3.
 Butow PN, Maclean M, Dunn SM, Tattersall MH, Boyer MJ. The
dynamics of change: cancer patients’ preferences for information,
involvement and support. Ann Oncol 1997;8:857–63.
 Stiggelbout AM, Kiebert GM. A role for the sick role. Patient
preferences regarding information and participation in clinical deci-
sion making. Can Med Assoc J 1997;157:383–9.
 Wong F, Stewart DE, Dancey J, Meana M, McAndrews MP, Bunston
T, Cheung AM. Men with prostate cancer: influence of psychological
factors on informational needs and decision making. J Psychosom Res
 Stewart DE, Wong F, Cheung AM, Dancey J, Meana M, Cameron JI,
McAndrews MP, Bunston T, Murphy J, Rosen B. Information needs
and decisional preferences among women with ovarian cancer. Gyne-
col Oncol 2000;77:357–61.
 McKinstry B. Do patients wish to be involved in decision making in
the consultation? A cross sectional survey with video vignettes. Br
Med J 2000;321:867–71.
 Mansell D, Poses RM, Kazis L, Duefield CA. Clinical factors that
influence patients’ desire for participation in decisions about illness.
Arch Intern Med 2000;160:2991–6.
 Davis MA, Hoffman JR, Hsu J. Impact of patient acuity on preference
for information and autonomy in decision making. Acad Emerg Med
 Rosen P, Anell A, Hjortsberg C. Patient views on choice and parti-
cipation in primary health care. Health Policy 2001;55:121–8.
 Anell A, Rosen P, Hjortsberg C. Choice and participation in the health
services: a survey of preferences among Swedish residents. Health
 Krantz DS, Baum A, Wideman MV. Assessment of preferences for
self-treatment and information in health care. J Pers Soc Psychol
 Brody DS. The patient’s role in clinical decision-making. Ann Intern
 Onel E, Hamond C, Wasson JH, Berlin BB, Ely MG, Laudone VP,
Tarantino AE, Albertsen PC. Assessment of the feasibility and impact
of shared decision making in prostate cancer. Urology 1998;51:63–6.
 Deber RB. Physicians in health care management. 8. The patient–
physician partnership: decision making, problem solving and the
desire to participate. Can Med Assoc J 1994;151:423–7.
 Green J, Britten N. Qualitative research and evidence based medicine.
Br Med J 1998;316:1230–2.
 Sainio C, Lauri S, Eriksson E. Cancer patients’views and experiences
of participation in care and decision making. Nurs Ethics 2001;8:97–
management in chronic illness: insights from diabetes research.
Patient Educ Couns 2001;42:81–90.
 LuptonD. Consumerism,reflexivity ad the medicalencounter. Soc Sci
choice: lay constructions oftreatmentdecision-making amongwomen
with early-stage breast cancer. Sociol Health Illn 1998;20:71–95.
 Kenny P, Quine S, Shiell A, Cameron S. Participation in treatment
decision-making by women with early stage breast cancer. Health
 Waterworth S, Luker KA. Reluctant collaborators: do patients want to
be involved in decisions concerning care? J Adv Nurs 1990;15:
 Department of Health. The expert patient: a new approach to chronic
disease management for the 21st century. London: Department of
Health, 2001. p. 12.
 Elwyn G, Edwards A. Evidence-based patient choice. Oxford: Oxford
University Press, 2001.
 Kaplan S, Greenfield S, Ware JE. Assessing the effects of physician–
patient interactions on the outcomes of chronic disease. Med Care
 Greenfield S, Kaplan S, Ware Jr JE. Expanding patient involvement in
care. Ann Intern Med 1985;102:520–8.
 Greenfield S, Kaplan SH, Ware Jr JE, Yano EM, Frank HJ. Patients’
participation in medical care: effects on blood sugar control and
quality of life in diabetes. J Gen Intern Med 1988;3:448–57.
effects of being offered choice of surgery for breast cancer. Br Med J
 Haynes B, McKibbon KA, Kanani R. Systematic review of rando-
mised trials of interventions to assist patients to follow prescriptions
for medications. Lancet 1996;348:383–6.
 Royal Pharmaceutical Society of Great Britain. From compliance to
concordance:towards sharedgoals in medicinetaking.London:Royal
Pharmaceutical Society, 1997.
K, Payne S. Observational study of effect of patient centredness and
positive approach on outcomes in general practice consultations. Br
Med J 2001;323:908–11.
R. Say et al./Patient Education and Counseling 60 (2006) 102–114 114