Journal of Psychosomatic Res

What do general practice patients want when they present medically

unexplained symptoms, and why do their doctors feel pressurized?

Peter Salmona,T, Adele Ringa,b, Christopher F. Dowrickb, Gerry M. Humphrisc

aDivision of Clinical Psychology, University of Liverpool, Whelan Building, Brownlow Hill, L69 3GB Liverpool, United KingdombDivision of Primary Care, University of Liverpool, Whelan Building, Brownlow Hill, L69 3GB Liverpool, United Kingdom

cBute Medical School, Fife, KY16 9TS St. Andrews, United Kingdom

Received 12 December 2004; received in revised form 28 February 2005; accepted 15 March 2005

Abstract

Objective: We tested predictions that patients with medically

unexplained symptoms (MUS) want more emotional support and

explanation from their general practitioners (GPs) than do other

patients, and that doctors find them more controlling because of this.

Design: Thirty-five doctors participated in a cross-sectional compar-

ison of case-matched groups. Three hundred fifty-seven patients

attending consecutively with MUS were matched for doctor and time

of attendance with 357 attending with explained symptoms. Patients

self-reported the extent to which they wanted somatic intervention,

emotional support, explanation and reassurance. Doctors rated their

0022-3999/05/$ – see front matter D 2005 Elsevier Inc. All rights reserved.

doi:10.1016/j.jpsychores.2005.03.004

T Corresponding author. Tel.: +44 151 794 5531; fax: +44 151 794

5537.

E-mail address: psalmon@liv.ac.uk (P. Salmon).

perception of patients’ influence on the consultation. Predictions

were tested by multilevel analyses. Results: Patients with MUS

sought more emotional support than did others, but no more

explanation and reassurance or somatic intervention. A minority of

doctors experienced them as exerting more influence than others.

The experience of patient influence was related to the patients’ desire

for support. Conclusions: Future research should examine why GPs

provide disproportionate levels of somatic intervention to patients

who seek, instead, greater levels of emotional support.

D 2005 Elsevier Inc. All rights reserved.

Keywords: Medically unexplained symptoms; Somatisation; General practitioner; Intentions

Introduction reflecting how doctors experience them [12,13], doctors’

Patients with unexplained physical symptoms are com-

mon in primary care, accounting for around 10–20% of

patients who attend general practitioners (GPs; [1,2]). They

receive disproportionate levels of symptomatic investigation

and treatment, which is largely ineffective and sometimes

iatrogenic [3,4]. The use of these interventions has been

widely attributed to patients’ belief that symptoms are caused

by physical disease, their consequent insistence on somatic

intervention and their rejection of psychological help [5,6].

Consistent with this reasoning, many GPs and other doctors

are dissatisfied with these consultations and feel that they are

controlled by the patient [6–11].

Although patients with medically unexplained symptoms

(MUS) have long been labelled as bheartsinkQ or bdifficult,Q

experience is unreliable as evidence of how the patients do

present. When interviewed, patients with persistent MUS

described seeking support or convincing explanation rather

than medical treatment [14]. When directly observed in

consultations with GPs, transparent pressure for symptoma-

tic intervention was rare [15]. Almost all patients indicated

psychological needs—for convincing explanation or emo-

tional support [16,17]. These recent qualitative findings are

incompatible with the influential assumption that patients

with MUS demand somatic intervention. Instead, they

suggest a very different hypothesis: that MUS patients

differ from others because they seek more emotional support

and explanation. The first aim of the present study was to

test this hypothesis.

The view that patients with MUS influence the con-

sultation more than do others has arisen from studies in

which GPs and other doctors have been interviewed about

their difficulties with such patients. Interviews about

challenging subjects often produce justifications rather than

earch 59 (2005) 255–262

P. Salmon et al. / Journal of Psychosomatic Research 59 (2005) 255–262256

valid descriptions or explanations. Therefore, it is unknown

whether doctors’ sense of patient influence arises as an

immediate response to consultation or only later, in the

context of interviews, as a way of justifying the somatic

treatment that they give these patients. Therefore, our

second aim was to test, across a large sample of consecutive

patients, whether GPs feel more influenced by patients with

MUS than by others.

The third aim was to test a hypothesis about why this

should happen. Rather than the generally accepted, but never

verified, assumption that patients pressurise GPs by seeking

symptomatic interventions that the doctor does not want to

give, we tested the view, suggested by qualitative findings

[16,17], that GPs feel more influenced by MUS patients

because of the patients’ desire for emotional support.

Method

Participants and procedure

After approval by local ethics committees, 50 GPs from

11 practices in Liverpool and South Cheshire were

approached to take part in the study; 42 (84%) agreed

(22 males, 20 females), with 5–42 years medical experience.

Practice size ranged from 1 to 10 GPs (mean=4.5) and from

2087 to 13,116 patients (mean=7564). Six practices were

urban, four were suburban and one was rural. Jarman

deprivation scores ranged from �11 to 56 (mean 21.27).

Consecutive patients attending participating doctors on

consecutive days were approached by the researcher (AR) on

arrival and asked for written consent to take part in a study

that involved audiorecording their consultation (analyses of

recordings are not reported here). From 303 surgeries

provided by 35 (19 male, 16 female) GPs, 4424 patients

were available for inclusion in the study. Of these, 976 were

excluded: 542 (55.5%) because they were less than 16 years

old; 294 (30%) because they had been recruited previously;

64 (7%) because of communication difficulties or extreme

distress; 32 (3%) because they had declined previously; and

14 (1%) who were attending for a third party. Thirty (3%)

were excluded by GPs without stating a reason, while 69

other patients consulted before the researcher could seek

consent. Of the 3379 asked to participate, 2728 consented

(81%). GPs completed checklists on 2707 of these. From this

pool of participants, those patients with MUS and a

comparison group with physically explained symptoms were

selected for analysis.

MUS patients

There are no agreed research diagnostic criteria for

primary care patients with unexplained symptoms, and the

use of standardised instruments can be too restrictive [18].

The present study concerns difficulties that patients present

for doctors. Therefore, as reported previously [15], we used

a procedure based on that described by Peveler et al. [1] to

identify patients that, in the doctor’s opinion, have

unexplained symptoms. Immediately after each consulta-

tion, the doctor completed a checklist to indicate whether

the consultation (i) involved a presentation of a physical

symptom that (ii) could not entirely be explained by a

recognizable physical disease. Consultations satisfying

these two criteria were regarded as concerning MUS.

Although it is possible that some symptoms identified as

dunexplainedT might prove to have a pathological cause, our

procedure ensures a patient group that is defined by their

clinicians’ belief that such a cause is likely to be absent.

Control group

Comparison patients were drawn from those whom the

GPs recorded as presenting with physical symptoms and for

whom they did not designate the symptoms as unexplained.

A subsample of these patients was selected by a semi-

random procedure to ensure a group that was matched with

the MUS sample for place and time of recruitment. For each

MUS patient, the records of consultations with consenting

patients were used to identify the patient with explained

symptoms that consulted next or previously (chosen at

random) in that surgery session. Where no suitable patient

was available (31% of matches), the search continued to the

next or previous surgery.

Data collection

Before consultation, each patient completed the 22-item

version of the Patient Request Form [20]. This provides

scores for three major types of help that patients seek from

their GPs, i.e., the patients’ intentions: medical investigation

and treatment; explanation and reassurance; and emotional

support. Raw scores on these scales vary from 0 to 14, 18

and 16, respectively, but are skewed. Scores were tricho-

tomised to provide normally distributed variables: scores of

0 and 1 were coded 0; scores at the maximum and

maximum–1 were coded 2; and intermediate scores were

coded 1. The resulting scores were standardised to a mean

of 0 and standard deviation of 1. Patients’ age and patient

and GP gender were noted as potential control variables.

Immediately after consultation with each consenting

patient, the GP rated dwho, in your opinion, influenced the

consultation mostT using a five-point Likert scale [19]:

definitely me (+2); somewhat me (+1); equal (0); somewhat

patient (�1); definitely patient (�2).

Data analysis

Groups were compared on age by t test, on gender ratio by

v2 and on GPs’ rating of influence by Mantel–Haenszel v2.

There are two levels of variability in the data. Because

each GP sees several patients, scores can vary between

patients and between groups of patients seen by the same

GP. Therefore, in multilevel analyses, we distinguished

variability at the patient and GP levels [21]. For each of the

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

Emotionalsupport

Explanation &reassurance

Investigation andtreatment

Sta

nd

ard

ised

sco

reMUS Control

Fig. 1. Mean standardised (sample mean=0; variance=1) scores on each

intention scale for MUS and control patients.

P. Salmon et al. / Journal of Psychosomatic Research 59 (2005) 255–262 257

three intention scores, a mixed-effect model with random

intercepts and random slopes was fitted. GPs were

distinguished as a random factor, and, to avoid the

restrictive assumption that MUS patients should affect all

GPs similarly, the influence of clinical group (MUS vs.

comparison group) was modeled as an effect that varied

randomly between GPs (i.e., random slopes). As control

variables, patients’ age and gender were included as fixed

effects. Initially, GPs’ gender was included as a potential

predictor of all effects. Because we had no a priori

hypothesis about its influence, we tested its significance

for each intention by a multiparameter test (which simulta-

neously tests the significance of all relevant coefficients). In

every case, GP gender was nonsignificant and the analysis

was repeated excluding it. Results are shown for the latter

analyses. In analysing each intention, the remaining two

intentions were included as control variables with random

slopes. Coefficients for the GP level were used to test the

significance of the effects of clinical group and patient

gender and age. For each analysis, variance components

were examined to indicate the relative magnitude and the

significance of variability between GPs in the mean

(intercept) and in the influence of clinical group (slope).

Doctors’ rating of who influenced the consultation was

analysed similarly. An initial analysis included clinical group

and patient age and gender as the patient-level predictors. The

second examined whether differences in the influence that

GPs perceived from MUS and comparison patients could be

explained by differences in what the patients sought. There-

fore, in separate analyses, each intention was included as a

predictor, with slope varying randomly at the GP level.

In all analyses, age and intentions were centred at the

grand mean. Analyses were by HLM 5.05 and SPSS 11.0

for Windows. The criterion for significance was Pb.05.

Results

Participants

Samples

Of 2707 consenting patients for whom GPs completed

checklists, 466 (17%) were identified as having MUS. Of

Table 1

Patient characteristics

MUS (n=357) Controls (n=357)

Male 131 (37%) 142 (40%)

Age: mean, median (range) 45.4, 43 (16–85) 49.0, 48 (16–88)

Ratings of influence over

consultation (n)

346 306

Definitely patient 12 (4%) 5 (2%)

Somewhat patient 53 (15%) 33 (11%)

Equal 176 (51%) 150 (49%)

Somewhat doctor 88 (25%) 89 (29%)

Definitely doctor 17 (5%) 29 (10%)

these, 357 (77%) provided questionnaire data. The median

number of clinics observed for each GP was 10 (range:

1–14); the median number of MUS patients recruited from

each GP was 9.0 (range: 1–25); and the median ratio of MUS

patients per clinic was 1 (range: 0.17–2.50). Out of those

patients who were not identified as having MUS, GPs

identified 1644 (73%) patients as having (explained) physical

symptoms. Of these, 1261 (76%) consented and provided

data. A subsample of 357 patients was selected as described

above. MUS patients were slightly younger than controls

(t=2.71, Pb.01), but did not differ in gender (Table 1).

GPs failed to complete 62 ratings of influence (9%),

mostly from patients with explained symptoms (Table 1);

analyses of this variable used only those patients for whom

it was completed. Patients for whom ratings were provided

did not differ in age, gender or any intention score from

those for whom ratings were missed. Apart from the GPs’

rating, 1.2% of data were missing. For multivariate analysis,

these were interpolated by an expectation-maximization

algorithm implemented in SPSS 11.

Intentions

Mean intention scores are shown in Fig. 1 and analyses

are summarized in Table 2. MUS patients were more likely

than those with physically explained symptoms to seek

emotional support. Although Fig. 1 suggested that they

sought more explanation and reassurance, this difference

was insignificant. They were no more likely than controls to

seek medical investigation and treatment.

Table 2

Results of multilevel analyses of patients’ intentions

Emotional

support

Explanation

and

reassurance

Investigation

and

treatment

Variance

components

GP .042TTT .019 .041TTGP�MUS .063 .008 .025

Within-GP

error

.703 .673 .700

Coefficients Intercept a �.204 .059 .015

Female

gender

.063 �.111 �.03

Age �.004TT �.006TTT .007TTTMUS .350T,TT .015 .004

For each analysis, the other two intention scores were used as covariates

(see text) but are not shown here.a Value for male patients of mean age with physically explained

symptoms.

T Pb.05.

TT Pb.01.

TTT Pb.001.

GP

Influ

ence

rat

ing

2.0

1.0

0.0

-1.0

-2.0

Control MUS

MainlyGP

Mainlypatient

Fig. 2. Variability between GPs in their experience of being influenced by

MUS and control patients. Each line shows, for one GP, the difference in

influence ratings between MUS and control patients.

P. Salmon et al. / Journal of Psychosomatic Research 59 (2005) 255–262258

Age influenced each intention. Older patients sought less

support and less information and reassurance, but more

investigation and treatment than did younger ones. Patient

gender was not significant.

Patient influence

In both groups, GPs rated most consultations as influ-

enced equally by doctor and patient or mainly by the doctor

(Table 1). Although MUS consultations were more likely

than control consultations to be experienced as influenced by

the patient (Mantel–Haenszel v2=9.68, df=1, Pb.01), thiseffect was small. The significant coefficient for MUS in the

Table 3

Results of multilevel analyses of GPs’ ratings of who influenced the consultation

Analysis 1

Variance

components

GP .156TTTGP�MUS .103TTExplanation and reassurance

Emotional support

Investigation and treatment

Within-GP error .602

Coefficients Intercept a .434

Female gender �.082

Age �.002

MUS �.188TExplanation and reassurance

Emotional support

Investigation and treatment

Negative ratings indicate patient influence.a Value for male patients with physically explained symptoms, of mean age

intention.

T Pb.05.

TT Pb.01

TTT Pb.001.

first multilevel analysis confirmed this difference, but the

significant variance component associated with it showed

that GPs differed considerably, and significantly, in the

extent to which MUS patients were experienced as more

influencing than others (Table 3; Fig. 2). Indeed, a few GPs

felt less influenced by them.

The subsequent analyses included patients’ intentions

(Table 3). The significant negative coefficient of emotional

support confirmed that GPs’ perception of patients’ influ-

ence was associated with patients’ desire for more emotional

most

Analyses controlling for each intention

Controlling for

explanation and

reassurance

Controlling

for emotional

support

Controlling fo

investigation

and treatment

.152TTT .164TTT .161TTT

.094TT .104TT .108TT

.005

.002

.002

.598 .557

.428 .407 .432

�.069 �.077 �.083

�.002 �.002 �.001

�.202T �.148 �.179T.041

�.073TT�.027

, and, in analyses controlling for each intention, with mean score on tha

r

t

P. Salmon et al. / Journal of Psychosomatic Research 59 (2005) 255–262 259

support. Controlling for the desire for support, the effect of

MUS was reduced and no longer significant, although the

significant variance component indicates that GPs still

differed in the effect of MUS on their perception of patient

influence. The other intentions were not associated with this

perception and did not account for the greater influence

associated with MUS.

Discussion

Current views of what patients with MUS want when

they consult a GP have been shaped by what doctors say

when interviewed about the difficulties that they have with

these patients. In this study, we asked patients directly. The

results negate the influential view that patients with MUS

pressurise GPs by seeking more somatic intervention than

do other patients. Instead, they are consistent with hypoth-

eses, derived from recent observational studies, that they

seek more emotional support.

MUS patients wanted more emotional support than did

patients whose symptoms their doctors felt able to explain

but they did not want more somatic intervention. We found

no significant evidence for our prediction that MUS patients

want more explanation than others do. However, the

questionnaire that we used conflates convincing explanation

with simple reassurance, which patients with MUS do not

generally seek [16]. These findings contradict the belief of

many GPs and researchers that patients with MUS deny

psychological needs. They are, however, consistent with

previous observations that emotionally distressed patients

who present somatically in primary care readily acknowl-

edge psychological needs when asked [22] and that MUS

patients generally provide GPs with cues to psychological

needs [17]. That MUS patients receive more somatic

intervention than do others can therefore no longer be

attributed to their desire for it. As we have proposed

previously, it seems that GPs provide somatic responses that

patients do not seek. The present study cannot indicate why

this happens. GPs often misperceive what patients seek

[23,24], and their treatment decisions are more closely

related to their perception of patients’ wishes than to

patients’ wishes [25,26]. Therefore, GPs might treat MUS

patients somatically because they think that patients want

this. However, using the same measure of patients’

intentions as we used in this study, we previously found

that the desire for support was the one intention that GPs

could detect more accurately than chance [24]. Alterna-

tively, evidence that many GPs feel hostile to MUS patients,

consider their symptoms to be invalid [7], do not want to

provide psychological support to them [9] and actively

disregard their psychological cues [17] suggest that somatic

responses might be a way of avoiding emotional engage-

ment with the patient.

We had no a priori hypotheses about the effect of patients’

age on their intentions. Therefore, the findings that older

patients sought less support, less explanation and reassurance

and more somatic investigation and treatment are suggestive

rather than definitive. Age or cohort effects are possible.

Older patients might have more concerns about their health

and, therefore, be more inclined to seek investigation and

referral. Alternatively, older people might have learned a

more treatment-focused relationship with their doctors.

Studies in which doctors have been interviewed about

MUS have suggested that they feel that such patients control

them, largely by demanding physical intervention [7,8,11].

Such findings should be interpreted cautiously because

interviewees’ accounts of challenging issues often justify

their behaviour rather than explain it. Our findings provide

the first empirical confirmation that GPs do experience

MUS patients as exerting more influence than others do.

However, the effect was very small when averaged across

doctors because, while a few experienced MUS patients as

considerably more controlling, many experienced no differ-

ence. Moreover, the perception that patients influenced the

consultation did not arise from their desire for somatic

intervention. In general, GPs’ experience of patients’

influence was related to how much emotional support

patients wanted, and their greater desire for such support

could account statistically for the slightly greater influence

perceived from MUS patients on average. However, the

wide variability between GPs in whether they felt more

influenced by MUS patients could not be so explained.

This study has limitations. Although we ensured diver-

sity between practices, GPs were recruited from one

geographical region. Moreover, we necessarily recruited

them from a population that was interested in training and

research and, perhaps, therefore, better able to manage MUS

patients than other GPs. A better way of measuring patients’

intentions is needed, which quantifies desire for the kind of

convincing explanation that qualitative work has suggested

that patients seek for their MUS. Similarly, the measurement

of GPs’ experience of patient influence should be elabo-

rated. In the context of MUS, doctors have described overt

patient influence as challenging and, inasmuch as it leads to

unnecessary somatic intervention, damaging. However,

being open to patient influence is valued in current concepts

of patient centeredness. Future studies should measure not

just the perception of patient influence, but also its

emotional impact. The generalisability of our findings is

restricted to MUS, as operationalized in our case-definition

procedure, and more evidence is needed about variability

within the patients identified by this procedure, e.g., in

relation to persistence of symptoms or the presence of

emotional distress, as well about relationships between this

and other ways of identifying patients. Finally, our GPs

failed to complete more ratings of influence for patients with

explained than unexplained symptoms, although whether

they completed the ratings was unrelated to any other

variable being studied and might reflect their greater

vigilance concerning MUS patients because these were the

focus of the research project.

P. Salmon et al. / Journal of Psychosomatic Research 59 (2005) 255–262260

Therefore, our findings need to be replicated and

extended using different methodologies. However, they

already also indicate new research questions. In particular,

we need to investigate why, in this population, some GPs

felt so controlled by MUS patients. That is, the explanation

for why MUS patients can be experienced in this way

should be sought in doctors rather than, as presently, in

patients. Doctors might simply not regard patients’ com-

plaints as legitimate demands on their time, or they might

regard themselves as unable to meet these demands. They

might have personal and consulting characteristics that

influence their attitudes to MUS. In particular, GPs vary in

the extent to which they espouse patient-centred models of

care [27] and take roles of friend and healer versus

technician or detective [28].

There is growing awareness of the role of medical

services in shaping MUS [4,29–31]. Our findings concur

with previous qualitative reports that patients with MUS

who attend seeking psychological responses receive somatic

ones. Such responses are, in turn, likely to shape the

patients’ beliefs and consulting behaviour, encouraging

future somatic presentation and dependency. On this

analysis, somatisation is a product of consultation rather

than psychopathology, and the key to understanding why

this group of patients receive disproportionate somatic

intervention lies in understanding why doctors respond

somatically to patients who seek emotional support.

Acknowledgments

The study was funded by the UK Medical Research

Council Grant G9900294. We are grateful for the enthusi-

astic cooperation of the participating GPs.

References

[1] Peveler R, Kilkenny L, Kinmoth AL. Medically unexplained

physical symptoms in primary care: a comparison of self-report

screening questionnaires and clinical opinion. J Psychosom Res

1997;42:245–52.

[2] Weijden van der T, Velsen van M, Dinant G-J, Hasselt van CM, Grol

R. Unexplained complaints in general practice: prevalence, patients’

expectations, and professionals’ test-ordering behavior. Med Decis

Making 2003;23:226–31.

[3] Barsky AJ, Ettner SL, Horsky J, Bates DW. Resource utilization of

patients with hypochondriacal health anxiety and somatization. Med

Care 2001;39:705–15.

[4] Stanley IM, Peters S, Salmon P. A primary care perspective on

prevailing assumptions about persistent medically unexplained phys-

ical symptoms. Int J Psychiatry Med 2002;32:125–40.

[5] Goldberg D, Bridges K. Somatic presentations of psychiatric illness in

primary care settings. J Psychosom Res 1988;32:137–44.

[6] Reid S, Whooley D, Crayford T, Hotopf M. Medically unexplained

symptoms: general practitioners’ attitudes towards their cause and

management. Fam Pract 2001;18:519–23.

[7] Wileman L, May C, Chew-Graham CA. Medically unexplained

symptoms and the problem of power in the primary care consultation:

a qualitative study. J Fam Pract 2002;19:178–82.

[8] Chew-Graham C, May C. Chronic low back pain in general practice:

the challenge of the consultation. J Fam Pract 1999;16:46–9.

[9] Garcia-Campayo J, Sanz-Carrillo C, Yoldi-Elcid A, Lopez-Aylon R,

Monton C. Management of somatisers in primary care: are family

doctors motivated? Aust N Z J Psychiatry 1998;32:528–33.

[10] Hartz AJ, Noyes R, Bentler SE, Damiano PC, Willard JC, Momany

ET. Unexplained symptoms in primary care: perspectives of doctors

and patients. Gen Hosp Psychiatry 2000;22:144–52.

[11] Steinmetz D, Tabenkin H. The difficult patient as perceived by family

physicians. Fam Pract 2001;18:495–500.

[12] Clements WM, Haddy R, Backstrom D. Managing the difficult

patient. J Fam Pract 1980;10:1079–83.

[13] Mathers NJ, Jones N, Hannay D. Heartsink patients: a study of their

general practitioners. Br J Gen Pract 1995;45:293–6.

[14] Peters S, Stanley I, Rose M, Salmon P. Patients with medically

unexplained symptoms: sources of patients’ authority and implications

for demands on medical care. Soc Sci Med 1998;46:559–65.

[15] Ring A, Dowrick C, Humphris G, Salmon P. Do patients with

unexplained physical symptoms pressure GPs for somatic treatment?

A qualitative study. BMJ 2004;328:1057–60.

[16] Dowrick CF, Ring A, Humphris GM, Salmon P. Normalisation of

unexplained symptoms by general practitioners: a functional typology.

Br J Gen Pract 2004;54:165–70.

[17] Salmon P, Dowrick CF, Ring A, Humphris GM. Voiced but unheard

agendas: qualitative analysis of the psychosocial cues that patients

with unexplained symptoms present to general practitioners. Br J Gen

Pract 2004;54:171–6.

[18] Schilte AF, Portegijs PJM, Blankenstein AH, Knottnerus JA.

Somatisation in primary care: clinical judgement and standard-

ised measurement compared. Soc Psychiatr Epidemiol 2000;35:

276–82.

[19] Salmon P, Marchant-Haycox S. Surgery in the absence of physical

pathology: relationship of patients’ presentation to gynecologists’

decision for hysterectomy. J Psychosom Res 2000;49:119–24.

[20] Valori R, Woloshynowych M, Bellenger N, Aluvihare V, Salmon

P. The patients requests form: a way of measuring what patients

want from their general practitioner. J Psychosom Res 1996;40:

87–94.

[21] Raudenbush SW, Bryk AS. Hierarchical linear models: applications

and data analysis methods. 2nd ed. Thousand Oaks7 Sage, 2002.

[22] Kirmayer LJ, Robbins JM. Patients who somatize in primary care:

a longitudinal study of cognitive and social characteristics. Psychol

Med 1996;26:937–51.

[23] Britten N, Stevenson FA, Barry CA, Barber N, Bradley CP.

Misunderstandings in prescribing decisions in general practice:

qualitative study. BMJ 2000;320:484–8.

[24] Salmon P, Sharma N, Valori R, Bellenger N. Patients’ intentions in

primary care: relationship to physical and psychological symptoms

and their perception by general practitioners. Soc Sci Med 1994;38:

585–92.

[25] Cockburn J, Pit S. Prescribing behaviour in clinical practice: patients’

expectations and doctors’ perceptions of patients’ expectations—a

questionnaire study. BMJ 1997;315:520–3.

[26] Britten N, Ukoumunne O. The influence of patients’ hopes of

receiving a prescription on doctors’ perceptions and the decision to

prescribe: a questionnaire survey. BMJ 1997;315:1506–10.

[27] Leopold N, Cooper J, Clancy C. Sustained partnership in primary

care. Fam Pract 1996;42:129–37.

[28] Robinson WD, Priest LA, Susman JL, Rouse J, Crabtree BF.

Technician, friend, detective, and healer: family physicians’ responses

to emotional distress. J Fam Pract 2001;50:864–70.

[29] Barsky AJ, Borus JF. Functional somatic syndromes. Ann Intern Med

1999;130:910–21.

[30] Page LA, Wessely S. Medically unexplained symptoms: exacerbating

factors in the doctor–patient encounter. J R Soc Med 2003;96:223–7.

[31] Wessely S, Nimnuan C, Sharpe M. Functional syndromes: one or

many? Lancet 1999;354:936–9.