Improving patients’ communication with doctors: a systematic review of intervention studies

Review

Improving patients’ communication with doctors: a systematicreview of intervention studies

Jane Harrington, Lorraine M. Noble*, Stanton P. NewmanDepartment of Psychiatry and Behavioural Sciences, Unit of Health Psychology, University College London, Wolfson Building,

48 Riding House Street, London W1N 8AA, UK

Received 12 April 2002; received in revised form 7 December 2002; accepted 27 December 2002

Abstract

A systematic review of the literature examined intervention studies designed to increase patients’ participation in medical consultations.

Twenty-five papers describing twenty studies met the inclusion criteria. About half of the intervention studies were randomised placebo

controlled experimental designs. The studies were predominantly conducted in outpatient or primary care settings, with 50% of the

interventions presented in a written form. Overall, half of the interventions resulted in increased patient participation, with slightly more

significant results found for bids for clarification than question-asking. However, of the 10 written interventions only two reported a significant

increase in question-asking. Patient satisfaction was the most commonly measured outcome, but few significant improvements were found.

However, there were significant improvements in other outcomes, including perceptions of control over health, preferences for an active role

in health care, recall of information, adherence to recommendations, attendance, and clinical outcomes. Few studies examined the links

between patient characteristics and the success of the interventions. Future research needs to establish which forms of intervention are most

effective and practical, for which groups of patients.

# 2003 Elsevier Science Ireland Ltd. All rights reserved.

Keywords: Doctor–patient communication; Intervention studies; Patient participation

1. Introduction

The importance of doctor–patient communication in the

process of health care has been established, with studies

demonstrating clear links between the quality of commu-

nication and patient satisfaction, adherence and clinical

outcomes [1–3]. Recognition of the importance of doc-

tor–patient communication has led to the formal teaching

of communication skills being considered as an essential

aspect of medical education, and this is now an integral

component of most medical curricula [4].

Whilst concerted efforts at behavioural change have been

directed at health care professionals, considering only the

doctor’s input into the consultation has the consequence of

neglecting half of the relationship. The patient’s contribution

to the consultation has also been found to be related to

outcomes. Patients who are more active during consultations

(in terms of asking questions, proffering information, and

expressing opinions) are more likely to understand the

treatment rationale and recommendations and there is a

growing body of evidence to suggest that more verbally

active patients may experience improved medical outcomes

[5–7]. Patients who are educated, have higher incomes, and

are older, have been found to receive more information [8],

which may indicate that different styles of communication

are more effective in eliciting information from doctors.

Alternatively, these patients may elicit more information

because of doctors’ responses to certain subgroups of

patients (e.g. positive stereotyping or subjective judgements

about the patient’s informational needs or requirements).

However, despite the evidence suggesting that active

participation is beneficial, in practice patients often con-

tribute little to the consultation apart from answering direct

questions. For example, patient question-asking has been

found to be the least frequent category of verbal behaviour in

medical visits [9].

A number of researchers have set out to redress the

balance, by intervening to improve the patients’ contribution

to their consultations. Six separate interventions to train

patients to communicate more effectively with their doctors

were reviewed by Anderson and Sharpe [10]. The evidence

suggested that interventions were beneficial, although there

Patient Education and Counseling 52 (2004) 7–16

* Corresponding author. Tel.: þ44-20-7679-9465;

fax: þ44-20-7679-9426.

E-mail address: [email protected] (L.M. Noble).

0738-3991/$ – see front matter # 2003 Elsevier Science Ireland Ltd. All rights reserved.

doi:10.1016/S0738-3991(03)00017-X

were too few studies to make reliable comparisons among

different forms of interventions.

This systematic review specifically examines intervention

studies directed at patients. It incorporates the six studies

that were subject of the earlier review and considers how the

evidence on this topic has accumulated to date.

This review examines intervention studies designed to

improve patients’ communication with their doctors. The

studies are considered in terms of: (a) the populations

investigated; (b) design; (c) the types of interventions used;

(d) the impact of the interventions on patients’ participation

during the consultation and on outcomes; and (e) the links

between patient characteristics and the success of the inter-

ventions.

2. Methods

2.1. Identification of studies

Literature searches were conducted to identify published

reports of studies using Medline (1966–2001), PsycLIT

(1966–2001), and BIDS Social Science Citation Index

(1981–2001). The key search terms were �patient-participa-

tion, �communication intervention, �communication-skills-

training, �patient involvement, �patient empowerment and�patient education. The search terms ‘patient-participation’

and ‘patient education’ were combined with the term ‘com-

munication’ to focus the search. Papers were included if they

described an intervention designed to improve patients’

communication with doctors in any setting, reported data

about the impact of the intervention on patients’ commu-

nication, and were published in English. The reference

sections of papers that met the inclusion criteria were also

searched, although this failed to yield any additional studies.

2.2. Overview of studies identified

Twenty-five papers describing twenty separate studies

met the inclusion criteria [11–35] (Table 1). The discrepancy

was due to investigators reporting findings from the same

samples in different papers. To avoid confusion, papers

using the same data set have been considered as a single

study. Two-thirds of the studies were published within the

last decade, indicating an increasing recognition of the

importance of patient participation in the consultation.

The studies were conducted in four countries, with the

majority from the USA.

Table 1

Studies of interventions to improve patients’ communication with doctors

References Country Population N Sample characteristics Design Control

groupsAgea Gender

(female, %)

[11], [12] USA Outpatient—medical clinic 200 Median ¼ 50 Mainly female RCT Placebo

[13] USA Simulated patients—hypertension 36 Median ¼ 28 100 RCT Placebo

[14] UK Primary care 131, 127 Not reported Not reported Quasi Placebo

[15], [16], [17] USA Outpatient—peptic ulcer, diabetes,

hypertension, breast cancer

252 47–55 (all groups) 58 RCT

and Quasi

Placebo

[18] USA Outpatient—hypertension 150 58 0 RCT Placebo

[19] USA Family medicine clinic 67 35 81 RCT Placebo

[20] USA Outpatient—obstetrics and gynaecology 53, 49 39 (study 2) 100 RCT Placebo

[21] USA General paediatric practice 141 8 43 RCT Placebo

[22] Australia Outpatient and

inpatient—medical oncology

142 51 84 RCT Placebo

[23] UK Primary care 80 Not reported Not reported UC UC

[24] UK Outpatient—medical oncology 95 44 59 RCT True

[25] USA Outpatient—breast cancer 60 59 100 RCT None

[26], [27] UK Primary care 120 43 61 RCT Placebo

[28] USA Primary care 201 46 54 RCT Placebo

[29] USA Primary care 20 37 70 Quasi True

[30] Australia Outpatient—medical oncology 60 53 52 RCT True

[31] UK Outpatient—dermatology,

gynaecology, orthopaedics

1077 Not reported 76 RCT True

[32], [33] USA Family practice clinic/private offices 150 45 71 RCT True

[34] USA Community sites 314 77 83 RCT True

[35] The Netherlands Outpatients—cardiology 103 64 19 RCT Placebo

Papers reporting findings from the same patient samples are grouped together; N: the number of patients providing data for the study; RCT: randomised

controlled trial; Quasi: quasi-experimental; UC: unclear; Placebo: placebo control group; True: non-intervention group; None: no control group.a Mean age reported unless otherwise specified.

8 J. Harrington et al. / Patient Education and Counseling 52 (2004) 7–16

3. Results

3.1. Populations studied

The settings were primarily outpatient and primary care

clinics (Table 1), with one study conducted in the commu-

nity [34], and one study considering inpatients [22]. No

studies considered mental health services, despite evidence

that patients attending these services often have unspoken

requests [36]. One study used healthy volunteers as simu-

lated patients [13], which is a procedure of questionable

validity.

The sample sizes varied considerably, with a range from

20 to 1077 (Table 1). The majority of studies were conducted

with adults, with a mean or median age of 28–77 years. Just

one of the interventions was directed at children [21]. In

most of the studies the sample was mainly or all female

(Table 1).

Although most of the studies described where the inter-

vention took place, many failed to explain the process of

care, for example, the sequence of events after the patient’s

arrival at the clinic, or the range of health professionals

providing care. Additionally, few reported whether patients

attended for more than one consultation, or described

patients’ previous experience with the service, despite the

possibility that familiarity and pre-established interaction

patterns may influence patients’ contribution to the consul-

tation. Only half of the studies reported the timing of the

intervention in relation to patients’ previous experience with

the service [15,19,22,24,25,28,30–32,35]. In six further

studies, previous experience was implied [11,14,23,26,

29,34].

3.2. Design

Over half of the studies (11/20) used a randomised

controlled experimental design, comparing one or more

experimental conditions with an attention placebo control

(Table 1). In three of these [11,21,28], the doctors or clinic

sessions were randomised, rather than individual patients.

Six studies using a randomised controlled trial design

failed to include an attention placebo control condition,

and three used a quasi-experimental design, both of which

reduce the confidence with which improvement in

patients’ communication can be attributed to the inter-

vention.

All but two of the studies [23,24] assessed the impact of

the intervention on the process of communication and out-

comes of the consultation. Three studies examined the

influence of patient characteristics on responses to the

interventions [13,18,26], considering a small range of vari-

ables.

Fifteen of the twenty studies used only post-intervention

measurement. Only four studies examined baseline and

post-intervention measures to enable changes due to the

intervention to be examined [15,24,25,32].

3.3. Overview of the interventions

3.3.1. Aims of the interventions

All of the interventions were designed to increase patient

participation during the consultation. The specific beha-

viours most frequently encouraged were question-asking,

raising concerns, and requesting clarification or checking

understanding (Table 2). In two studies, no specific beha-

viours were targeted.

Most of the interventions were delivered immediately

before the patients’ appointments (Table 2). In five studies,

interventions were delivered prior to the day of the appoint-

ment to allow patients more time to familiarise themselves

with the content of the intervention.

3.3.2. Modes of presentation

Written interventions were most common, followed by

face-to-face coaching and videotape (Table 2). A few studies

used a combination of modes. In two studies, the interven-

tion was simultaneously directed at other participants in the

consultation: either doctors [28] or doctors and parents [21].

Four studies compared the impact of different interventions,

either using the same mode with different content [13,14,18]

or comparing different modes of presentation [25].

3.3.3. Duration of interventions

The duration of the interventions varied considerably,

from 10 to 25 min for the video interventions to a 2 h

combined face-to-face and written intervention (Table 2).

In several studies, patients were allowed as much time as

they wanted with the intervention. However, nearly half of

the studies failed to report the duration of the interventions.

Both the different types of interventions and their duration

have practical implications, if the intention is that they will

be used in routine practice.

3.3.4. Content of the interventions

The written interventions tended to be presented in book-

let or checklist form, sometimes with instructions to write

down any questions that the patient may have had. The face-

to-face interventions were generally based on coaching

patients in question-asking techniques and the video inter-

ventions were based on modelling techniques. Modelling

was also employed in one of the face-to face interventions

[34] (Table 2).

3.4. Effects on the process of communication

Pendleton [37] distinguished between three types of

variables in doctor–patient communication research: ante-

cedents, process and outcome. The twenty studies consid-

ered in this review examined the effects of the interventions

on the process and outcome of communication. Process

variables include measures of the verbal and non-verbal

interactions between the doctor and the patient. Outcome

variables can be immediate, intermediate, and long-term

J. Harrington et al. / Patient Education and Counseling 52 (2004) 7–16 9

Table 2

Overview of intervention studies

References Behaviours

encouraged

Mode of

presentation

Time of

intervention

delivery

Duration

(min)

Brief description of

experimental intervention

[11], [12] Question-asking Face-to-face Immediately

pre-appointment

10 To work through a question-asking protocol, identifying questions patients may have

concerning five main aspects of illness (aetiology, duration, severity, prevention) and treatment.

[13] Question-asking,

clarification

Video,

face-to-face

Immediately

pre-appointment

20–25, Not reported Model expresses confusion and asks for clarification during a partially confusing presentation

on symptoms, causes and consequences of hypertension. All patients participated in a live

teaching session (on dietary advice); 50% were told to ask questions if confused during the

session.


checking, understanding

Written Immediately

pre-appointment

Not reported One group received permission to ask questions and the other received guidance on how to

check understanding.

[15], [16], [17] Question-asking,

decision-making,

overcoming barriers

Face-to-face Immediately

pre-appointment

20 Using a treatment algorithm as a guide, patients read their medical records and were coached to

ask questions and negotiate medical decisions with their doctors.


clarification, raising

concerns

Video Immediately

pre-appointment

14 Video 1. Model expresses confusion and asks for clarification during a partially confusing

presentation on definitions, control, complications and medication for hypertension. Video 2.

Model discloses problems and concerns about hypertension, including difficulties with the

regimen.

[19] Question-asking Written Immediately

pre-appointment

Self-determined Booklet educating patients to recognise information needs and verbalise requests for

information. Based on 33 questions covering five areas, state of body, reasons for state of

body, treatment, reasons for treatment, and impact on life.


obtaining information

Written Immediately

pre-appointment

Not reported Patients provided with a list of possible concerns, with instructions to write down at least three

questions to ask the doctor and to take the list to the appointment. Note from the doctor to

patient giving express permission to ask questions.


raising concerns,

decision-making,

active role

Video Immediately

pre-appointment

10 (child), 10

(parent), 15 (doctor)

Children: featured a child modelling effective communication. Parent: emphasised the

importance of doctor–patient communication and fostering the child’s active involvement.

Doctor: presented research evidence on health consequences of doctor–patient communication

and demonstrated building rapport, checking children’s understanding of medical information.

[22] Question-asking Written Immediately

pre-appointment

Not reported Prompt sheet with 11 items designed to encourage patient active participation in the

consultation. Patients encouraged to write down specific questions and to take sheet into

consultation.

[23] Raising concerns,

seeking information

Written Immediately

pre-appointment

Not reported Instructions encouraging patients to think about reason for attending, problems experienced,

worries, etc. Patient instructed to tell the doctor all these things clearly, early on, and to ask for

more information on points of interest.

[24] No specific behaviours Audiotape Up to one month

pre-appointment

Length of audiotape Patients provided with an audiotape of their previous consultation (first consultation, primary

bad news, i.e. diagnosis, or secondary bad news i.e. unsuccessful initial treatment), to take

home and listen to as many times as they wished.


raising concerns,

expressing opinions

Multimedia,

written

Immediately

pre-appointment

35–40, 15–20 Interactive programme consisting of text, graphics, audio narration, music, and video clips.

Programme divided into four sections, introduction, understanding the problem, treatment

options and the experiences of other women. Brochure, based on the above, conveying the

same message.

[26], [27] Raising concerns Written Immediately

pre-appointment

Not reported Leaflet asking patients to identify their problems, think about the causes, treatment, etc. and

provided space for the patient to write down their ideas and then voice these concerns during

the consultation and to ask questions about diagnosis.

[28] Raising concerns Written Immediately

pre-appointment

Not reported Patients completed a 25 item questionnaire containing five categories of concerns, desire for

medical information, psychosocial assistance, therapeutic listening, general health advice and

biomedical treatments.

10

J.H

arrin

gto

net

al./P

atien

tE

du

catio

na

nd

Co

un

seling

52

(20

04

)7

–1

6


information-seeking

and verifying

Face-to-face Immediately

pre-appointment

Not reported Patients presented with information aimed at improving their understanding of the typical

communication goals relevant to the three parts of the examination (history-taking,

examination, conclusion). It was stressed that patients should become the information seeker

and doctor information provider.

[30] Question-asking Written,

face-to-face

Immediately

pre-appointment

Not reported,

not reported

Prompt sheet, based on 17 frequently asked questions by patients. Coaching based on

discussion about the importance of question-asking, generating questions, comparison of

questions with the prompt sheet and adding new questions. Exploration of benefits and barriers

to question asking, and question-asking rehearsed.

[31] Question-asking Written 2 weeks,

pre-appointment

Self-determined A help card and an information letter to help patients prepare and prioritise questions to ask in

consultation covering patients’ condition (7 items), tests and treatment (12 items) other (4

items)

[32], [33] Information

seeking and

verifying,

information-provision

Written 2–3 days,

pre-appointment

Self-determined 14 page training booklet formatted like a workbook with examples and space for notes.

Prompted patients to write questions on a variety of topics and to list the details of their

symptoms and related medical history. Written summary of training booklet.

[34] Active role Face-to-face,

written

Up to 3 months

pre-appointment

2 h Group programme including modelling of both undesirable and desirable behaviours, role-

playing exercises, and discussion of potentially negative effects of passive involvement.

Provision of cue cards with a list of desirable active behaviours and a booklet for recording

reasons for visit, medications, and questions for the doctor, with room for recording the

answers.

[35] No specific behaviours Written 1 week,

pre-appointment

Self-determined Frequently asked question checklist. 49 items covering 10 topics: nature of disease, physical

state, medication, risk factors and lifestyle, treatment, examination, disease related problems in

the family, psychosocial problems, practical matters, other problems.

J.H

arrin

gto

net

al./P

atien

tE

du

catio

na

nd

Co

un

seling

52

(20

04

)7

–1

61

1

[37]. Immediate outcomes include patient satisfaction and

recall, intermediate outcome variables include adherence to

recommendations, and long-term outcomes include changes

in health status or lifestyle.

3.4.1. Measures of the process of communication

In the majority of studies, the process of communication

was measured using interaction analysis, usually from

audiotaped recordings [11,13–15,18,19,24–26,28,30,32],

with two studies using video [21,29]. Six of these used

previously reported interaction analysis systems [11,15,

18,21,22,24], with the remainder using newly devised sys-

tems, but since 1995 none of the studies have employed

reliable and valid methods. The use of different systems

made comparisons of results across studies difficult, parti-

cularly when the reliability of the measures had not been

established. Furthermore, studies using new measures were

less likely to examine inter-rater reliability than studies

using previously reported methods, which is a serious

methodological flaw. Researchers coding the interactions

between doctors and patients were reported to be blinded to

group allocation in only six of the studies [15,18,19,

21,26,32], leaving the remaining studies open to the possi-

bility of bias.

The interaction analysis systems were used to examine a

wide range of communicative behaviours, measuring fre-

quency, duration and ratios (e.g. patient to doctor beha-

viours). The most commonly measured variable was patient

question-asking. However, there was otherwise little con-

sistency in the measures used.

In several of the studies [20,23,31,34,35], the process

of communication was assessed by obtaining patients’ or

doctors’ perceptions of the consultation, and this was often

the only measure of the process of communication. None of

the studies using perceptions as a measure of process

attempted to establish the reliability or validity of the criteria

used. The sole use of subjective perceptions as an index of

the process of communication is problematic, as subjective

perceptions and behavioural measures have been found to

correlate poorly [25,38].

Nearly half of the studies measured only one or two

categories of communicative behaviour, thus failing to take

the opportunity to make use of the richness of the data

available. Five studies measured 10 or more categories of

communicative behaviour [11,15,21,22,29].

3.4.2. The impact of the interventions on the process

of communication

Although the studies used different measures of patient

participation, overall it was clear that the interventions had

the effect of encouraging patients to be more active in their

consultations (Table 3). Out of 16 studies that examined

variables related to patient participation, 10 reported a

significant increase and five reported a non-significant

increase. Broken down by mode of presentation, all but

one of the six face-to-face interventions reported a signifi-

cant increase in overall participation [11,13,15,29,30], and

all of the video interventions reported significant increases

[13,18,21]. However, of the 10 written interventions, only

two reported a significant increase [14,32].

Question-asking was the most frequently targeted specific

behaviour, with equal numbers of studies reporting signifi-

cant increases and non-significant trends (Table 3). The

range of question-asking was low in many studies, with a

Table 3

Impact of interventions on the process of communication

References Overall patient

participation

Patient question

asking

Patient

clarification

Consultation

length

Expressed

affect

Doctor encourages

patient participation

[11], [12] Sig " Sig " ¼ ¼ Sig " Sig "[13] Sig " Sig "[14] Sig " Sig "[15], [16], [17] Sig " NS " ¼ Sig "[18] Sig " Sig "[19] NS " NS "[20] NS " NS " ¼[21] Sig " Sig "[22] ¼ NS "[23]

[24] Sig " ¼ Sig "[25]

[26], [27] NS " NS " Sig "[28] Sig "[29] Sig " Sig " Sig " ¼[30] Sig " Sig "[31]

[32], [33] Sig " Sig " NS "[34] NS "[35] NS " ¼

(Sig "): significant increase; (Sig #): significant decrease; (¼): no change; (NS): non-significant change; blank: not assessed.


range of 0–7% questions per consultation reported for

control groups and 0–9% for intervention groups. Ques-

tion-asking appeared to represent very little of the patient’s

verbal contribution to the consultation, three studies report-

ing the proportion gave a range of 0.07–8.7% [11,22,24].

Differences among the studies may have been related to the

criteria used for question-asking. For example, one study

[29] that included direct and indirect questions, found that

patients asked more questions than those reported in other

studies. Significant increases in requests for clarification

were more consistently reported (Table 3), raising the

possibility that patients perceived it as less threatening to

increase their contribution based on the information that the

doctor was providing, rather than to ask spontaneous ques-

tions.

Six studies compared the effects of different interven-

tions. One found that instructing patients to ask questions

most effective, followed by a modelling video [13], and

another found that guidance in checking understanding was

more effective than giving permission to ask questions

[14]. Other findings indicated that modelling of requesting

clarification was more effective than modelling of raising

concerns [18]. A further study [30] found that while a

question prompt sheet resulted in increased question-asking

behaviour, adding face-to-face coaching did not encourage

further question-asking. Two studies found no differences

between their interventions, one comparing two similar

written interventions [20] and one comparing written and

multi-media presentations of the same information [25].

Anderson and Sharpe [10] suggested that future interven-

tion studies should be expected to extend assessments

beyond question-asking behaviour, but there is little

evidence that investigators have broadened the range of

communicative behaviours under consideration and ques-

tion-asking in its various guises still predominates.

Increasing patient participation might be expected to

significantly lengthen the consultation, an outcome that

would not generally be welcomed by doctors. However,

the evidence suggested that this concern may be unwar-

ranted. In five of the seven studies that considered this,

increasing patient participation did not result in an increase

in consultation length. Where differences in consultation

length between the groups were observed, in one study [26]

the difference was relatively small (1.2 min), but in the other

[28] it was considerably longer (6 min) (Table 3).

Although the aim of the interventions was to increase

patient participation, there was also evidence that the con-

sultation could be modified in other ways. In one study,

where the intervention consisted of giving patients an audio-

tape of their consultation [24], a significant reduction was

found in patients’ subsequent requests for information that

had been given during that consultation.

Two studies considered the impact of the intervention on

expressions of affect during the consultation. In one study

[11], patients who had received the intervention were rated

by independent observers as more anxious and angry, and

their doctors as more matter-of-fact, angry, and less sympa-

thetic. This suggested that increased patient participation

was uncomfortable for both parties, raising a concern that

although the intervention may have been effective, the

consequences were detrimental. In a further study [15,17]

the intervention resulted in significantly increased displays

of affect from doctors and patients, although the only detail

provided about this was that this included both positive and

negative emotions. Despite these interesting findings from

these earlier studies, expressed emotion has not been con-

sidered in any of the more recent investigations.

The impact of the interventions on doctors’ behaviour was

examined in three further studies, two of which found

significant increases in the encouragement doctors gave to

patients to ask questions (Table 3), and the third finding that

experimental group consultations were more likely to be

rated by the doctor as a ‘good consultation’ [23].

3.5. Effects on outcomes of the consultation

Immediate outcomes were the most commonly consid-

ered, with patient satisfaction the single most frequently

examined outcome. Most of the studies used newly-devised

measures of immediate outcomes or modified versions of

existing scales, which raises questions about validity and

reliability, and makes comparisons among studies proble-

matic.

Intermediate and long-term outcomes were considered in

only four studies [11,15,26,32], presumably due to the

resource implications of following up patients over time.

Attendance was considered in three studies and one study

measured adherence to both medications and behavioural

treatments (Table 4). Perhaps the most important long-term

outcome variable, disease control, was considered in only

one study. The lack of emphasis on outcomes that have a

clear practical importance in health care is a serious short-

coming of work in this field.

3.5.1. The impact of the interventions on outcomes

of the consultation

The studies generally found no significant increase in

patient satisfaction due to the interventions, with only two

exceptions (Table 4). Overall, high levels of satisfaction

were reported, with ceiling effects specifically reported in

three studies [20,28,34]. The high frequency of ceiling

effects and the insensitivity of measures of satisfaction

has been previously noted as an issue limiting findings on

this dimension [39,40]. Two studies considering doctor

satisfaction found no significant change [20,28].

Significant improvements were found in the accuracy of

patients’ recall of the consultation in three of the five studies

which examined this, although none of the five studies

considering patients’ knowledge of their illness found sig-

nificant changes (Table 4). This discrepancy might be due to

the fact that measures of recall emphasise treatment recom-

mendations, which may be more salient to patients, but also


may reflect differences in the way information about diag-

nosis and treatment are given in the consultation.

The aim of many of the interventions was to encourage

patients to be involved in their health care. Significant

increases in patients’ perceptions of control over their health

and preferences for an active role in their health care were

found in the four studies that considered this.

Only one of the five studies considering patients’ percep-

tions of their own health or adjustment to illness found a

significant improvement (Table 4). However, these percep-

tions were assessed soon after the consultation, and it could

be argued that these constructs would be more appropriately

measured as long-term outcomes.

Despite the evidence about increases in observable affect

from the two studies mentioned above, patient and doctor

self-reports about the consultation were more positive.

Patient anxiety was the most commonly measured mood

state, and decreases in anxiety due to the intervention were

noted in three out of five studies (Table 4). Furthermore, no

effects on doctors’ satisfaction with, or perceptions of, the

consultation were found in any of the studies that considered

this [20,26,28].

Although longer-term outcomes were rarely examined,

results were encouraging. Significant improvements in

attendance, adherence and disease control were found for

three separate samples of patients with chronic conditions

[11,15–17,27,33].

3.6. The influence of patient variables on the impact

of the interventions

If interventions are to be used routinely to improve patients’

communication with their doctors, it is important to determine

not only which interventions are successful and how they

affect outcomes, but also whether certain groups of patients

may benefit. Only three studies considered this issue.

Increases in question-asking were found to be higher for

patients from social classes I and II, and increases in the

length of the consultation were particularly evident in patients

who were younger, male, and from social classes I and II [26].

Two studies considered assertiveness, but their generalizabil-

ity was limited by their sampling techniques: in one study,

healthy volunteers were used to represent simulated patients

[13] and in another, the sample consisted only of male patients

[18]. Little can therefore be concluded, except that it remains a

possibility that patient characteristics can affect responses to

an intervention.

3.7. Quality of reporting and methodological issues

The studies reviewed clearly demonstrated that non-sig-

nificant findings are common in this area of research. In a

few studies, however, it was difficult to determine the results

of significance testing due to ambiguous reporting. Four

particular problems were evident: describing non-significant

Table 4

Impact of the interventions on patient outcomes

References Satisfaction Knowledge Recall Response to illness Mood Adherence Attendance Disease control

[11], [12] Sig # Sig " locus of control Sig " short term

[13] ¼[14] Sig "[15], [16], [17] ¼ ¼ or Sig # Sig " preference for

active role, Sig "perceived health

Significant

improvement

hypertension

and diabetes

[18] ¼ ¼[19] NS "[20] NS " ¼ Sig " locus of control NS # anxiety

[21] Sig " Sig " Sig " preference for

active role

¼ anxiety

[22] ¼ ¼ NS adjustment

[23]

[24]

[25] ¼ NS optimism

[26], [27] ¼ ¼ perceived health ¼ long-term

[28] ¼ ¼ perceived health Sig # anxiety,

Sig # depression

[29] Sig "[30] ¼ NS adjustment ¼ anxiety

[31] Sig "[32], [33] Sig " Sig "[34] ¼[35] NS " NS " Sig # anxiety

(Sig "): significant increase; (Sig #): significant decrease; (¼): no change; (NS): non-significant change; blank: not assessed.


trends as ‘marginally significant’ or ‘approaching signifi-

cance’, implying in the summary of the study that the

findings were significant, misusing the convention of aster-

isking significant findings in a table of results, and failing to

report the results of significance testing for all relevant

comparisons [11,18,20,26,28].

Several studies provided information about the relation-

ships between input, process and outcome variables in the

consultation, but failed to clearly report on the impact of the

intervention [12,13,16–19,22,24–26].

Few of the studies examined changes in measures from

baseline to post-intervention and post-intervention, with

most relying on a post-test control group design. This design

is not ideal as it fails to provide evidence that any differences

between experimental and control groups are a direct con-

sequence of the intervention, because the assumption that

both groups have equal baseline levels of performance is not

tested. In the context of a communication intervention the

collection of baseline data for certain measures, for example,

anxiety, knowledge and health status as such measures

would provide supporting information regarding the efficacy

of the intervention.

In terms of statistical analysis, several studies performed

multiple correlations on the same data set [12,17,24–26],

thus increasing the risk of Type I errors. Only two studies

reported power analyses [26,30]. The absence of a power

calculation is problematic, as the reader cannot determine

whether non-significant results are a consequence of an

inadequate sample size, or whether significant results (for

example, correlations) are related to a very large sample

size. In several studies it was difficult to ascertain the

numbers of patients participating, as some investigators

emphasised the number recruited rather than the number

providing data [11,13,14,24] or failed to report exactly how

many patients comprised the experimental and control

groups [15,22]. It would appear that since the publication

of the Anderson and Sharpe review [10], some of the

criticisms the authors made regarding the quality of report-

ing and sample size still have not been adequately addressed.

4. Conclusions

The studies considered in this review generally demon-

strated that interventions directed at patients can be success-

ful in increasing patient participation, and that, in some

circumstances, this can be achieved without an increase in

consultation length. Where the findings about patient parti-

cipation were statistically non-significant, there were trends

in the expected direction. The overall increase in patient

participation was most frequently assessed by examining

patient question-asking or attempts to clarify information,

although it has been noted that question-asking often repre-

sents only a small proportion of the amount of overall talk. It

is possible that patient question-asking may be regarded by

patients and doctors as inappropriately assertive behaviour.

Encouraging patient participation appeared to lead to a

greater sense of control and preference for a more active role

in consultations in a number of studies, although the results

were less clear when considering satisfaction, knowledge,

and recall.

4.1. Further research

Several areas warrant further investigation. Firstly, it

needs to be determined whether responses to interventions

are moderated by patient characteristics, to identify which

patients would derive greatest benefit from a communica-

tion intervention. Secondly, although both difficult and

costly, the impact of interventions on longer-term clinical

outcomes should be assessed. Studies should also take

advantage of the richness of the data available from audio

or videotaped recordings, given that interventions may

improve patient participation or affect the consultation

in more ways than simply increasing patient question-

asking.

Future research also needs to address the basic flaws

evident in some of the studies reviewed here. In particular,

the results of the interventions should be presented clearly,

the numbers of participants in each group should be

reported, and patients’ prior experience with the service

should be described. The key methodological issues that

need to be addressed are the inclusion of the appropriate

control groups, appropriate sample size, proper use of

randomisation and blinding, and the use of valid and reliable

measures.

4.2. Practice implications

At present, partially due to the diversity of the interven-

tions, there is insufficient evidence to draw definitive con-

clusions about which forms of intervention are most

successful in which contexts, although there was some

evidence that the more intensive face-to-face interventions

were more effective than written interventions. This raises

the issue about what is practical compared to what is

effective, which is important if the ultimate aim is for

interventions to be used routinely. Both the financial and

other resource implications of developing and implementing

interventions need to be considered, and more emphasis

could be placed on comparing the effectiveness of different

modes in achieving a particular outcome.

Acknowledgements

This work was undertaken by Jane Harrington, Lorraine

Noble and Stanton Newman, who received funding from the

NHS London Regional Office, Research and Development

Programme. The views expressed in the publication are

those of the authors and not necessarily those of the NHS

or the Department of Health.


References

[1] Hall JA, Roter DL, Katz BA. Meta-analysis of correlates of provider

behavior in medical encounters. Med Care 1988;26:657–75.

[2] Roter DL, Hall JA, Katz NR. Patient–physician communication: a

descriptive summary of the literature. Patient Educ Counsel 1988;

12:99–119.

[3] Stewart MA. Effective physician–patient communication and health

outcomes: a review. Can Med Assoc J 1995;152:1423–33.

[4] Simpson M, Buckman R, Stewart M, Maguire P, Lipkin M, Novak D,

et al. Doctor–patient communication: the Toronto consensus

statement. Br Med J 1991;303:1385–7.

[5] Rost K, Carter W, Inui T. Introduction of information during the initial

medical visit: consequences for patient follow-through with physician

recommendations for medication. Soc Sci Med 1989; 28: 315–21.

[6] Roter DL, Hall JA. Doctors talking to patients/patients talking to

doctors. Westport (CT): Auburn; 1993.

[7] Street RL. Active patients as powerful communicators: the commu-

nicative foundation of participation in care. In: Robinson WP, Giles

H, editors. The new handbook of language and social psychology.

Chichester (UK): Wiley. p. 541–60.

[8] Street RJ. Information-giving in medical consultations: the influence

of patients’ communicative styles and personal characteristics. Soc

Sci Med 1991;32:541–8.

[9] Roter DL, Hall JA, Katz NR. Relations between physicians’

behaviors and analogue patients’ satisfaction, recall, and impres-

sions. Med Care 1987;25:437–51.

[10] Anderson LA, Sharpe PA. Improving patient and provider commu-

nication: a synthesis and review of communication interventions.

Patient Educ Counsel 1991;17:91–134.

[11] Roter DL. Patient participation in the patient-provider interaction:

the effects of patient question asking on the quality of interaction

satisfaction and compliance. Health Educ Monogr 1977;281–315.

[12] Roter DL. Patient question asking and physician–patient interaction.

Health Psychol 1984;3:395–409.

[13] Wallston BS, Wallston KA, DeVellis BM, McLendon E, Percy J.

Modification of question asking behaviour in high and low assertive

women through modeling and specific instructions. Soc Behav Pers

1979;6:195–204.

[14] Robinson EJ, Whitfield MJ. Improving the efficiency of patients’

comprehension monitoring: a way of increasing patients’ participa-

tion in general practice consultations. Soc Sci Med 1985;21:915–9.

[15] Greenfield S, Kaplan S, Ware JE. Expanding patient involvement in

care: effects on patient outcomes. Ann Int Med 1985;102:520–8.

[16] Greenfield S, Kaplan S, Ware JE, Yano EM, Frank HJL. Patients’

participation in medical care: effects on blood sugar and quality of

life in diabetes. J Gen Int Med 1988;3:448–57.

[17] Kaplan SH, Greenfield S, Ware JE. Assessing the effects of physician–

patient interactions on the outcomes of chronic disease. Med Care

1989;27:110–27.

[18] Anderson LA, DeVellis B, DeVellis RF. Effects of modeling on

patient communication, satisfaction, and knowledge. Med Care

1987;25:1044–56.

[19] Tabac ER. Encouraging patient question-asking: a clinical trial.


[20] Thompson SC, Nanni C. Patient-oriented interventions to improve

communication in a medical office visit. Health Psychol 1990;9:

390–404.

[21] Lewis CC, Pantell RH, Sharp L. Increasing patient knowledge,

satisfaction, and involvement: randomized trial of a communication

intervention. Pediatrics 1991;88:351–8.

[22] Butow PN, Dunn SM, Tattershall MHN, Jones QJ. Patient

participation in the cancer consultation: evaluation of a question

prompt sheet. Ann Oncol 1994;5:199–204.

[23] Fredrikson LG, Bull PE. Evaluation of a patient education leaflet

designed to improve communication in medical consultations.


[24] Ford S, Fallowfield L, Hall A, Lewis S. The influence of audiotapes

on patient participation in the cancer consultation. Eur J Cancer

1995;31A:2264–9.

[25] Street RL, Voight B, Geyer C, Manning T, Swanson GP. Increasing

patient involvement in choosing treatment for early breast cancer.

Cancer 1995;76:2275–85.

[26] McCann S, Weinman J. Empowering the patient in the consultation:

a pilot study. Patient Educ Counsel 1996;27:227–34.

[27] McCann S, Weinman J. Encouraging patient participation in general

practice consultations: effect on consultation length and content.

Psychol Health 1996;11:857–69.

[28] Hornberger J, Thom D, MaCurdy T. Effects of a self-administered

previsit questionnaire to enhance awareness of patients’ concerns in

primary care. J Gen Int Med 1997;12:597–606.

[29] McGee DS, Cegala DJ. Patient communication skills training for

improved communication competence in the primary care medical

consultation. J Appl Commun Res 1998;26:412–30.

[30] Brown R, Butow PN, Boyer MJ, Tattersall MHN. Promoting patient

participation in the cancer consultation: evaluation of a prompt

sheet and coaching in question-asking. Br J Cancer 1999; 80:

242–8.

[31] Fleissig A, Glasser B, Lloyd M. Encouraging out-patients to make

the most of their first hospital appointment to what extent can a

written prompt help patients get the information they want? Patient

Educ Counsel 1999;38:69–79.

[32] Cegala DJ, McClure L, Marinelli TM, Post DM. The effects of

communication skills training on patients’ participation during

medical interviews. Patient Educ Counsel 2000;41:209–22.

[33] Cegala DJ, Marinelli T, Post D. The effects of patient communication

skills training on compliance. Arch Fam Med 2000;9:57–64.

[34] Tennstedt SL. Empowering older patients to communicate more

effectively in the medical encounter. Clin Geriatric Med 2000;16:61–

70.

[35] Martinali J, Bolman C, Brug J, van den Borne B, Bar F. A checklist

to improve patient education in a cardiology outpatient setting.


[36] Noble LM, Douglas BC, Newman SP. What do patients want and do

we want to know? a review of patients’ requests of psychiatric

services. Acta Psychiatr Scand 1999;100:321–7.

[37] Pendleton D. Doctor–patient communication: a review. In: Pendleton

D, Hasler J. (Eds.), Doctor–patient Communication. London:

Academic Press; 1983.

[38] Street RL. Analyzing communication in medical consultations. Med

Care 1992;30:976–88.

[39] Balogh R, Simpson A, Bond S. Involving clients in clinical audits

of mental health services. Int J Qual Health Care 1995;7:343–

53.

[40] Calnan M. Towards a conceptual framework of lay evaluation of

health care. Soc Sci Med 1988;27:927–33.


Improving patients’ communication with doctors: a systematic review of intervention studies

Documents

Transcript of Improving patients’ communication with doctors: a systematic review of intervention studies