Intended for healthcare professionals

Primary Care

Do patients with unexplained physical symptoms pressurise general practitioners for somatic treatment? A qualitative study

BMJ 2004; 328 doi: (Published 29 April 2004) Cite this as: BMJ 2004;328:1057
  1. Adele Ring, research assistant1,
  2. Christopher Dowrick, professor of primary medical care2,
  3. Gerry Humphris, professor of health psychology3,
  4. Peter Salmon (psalmon{at}, professor of clinical psychology1
  1. 1Department of Clinical Psychology, University of Liverpool, Whelan Building, Brownlow Hill, Liverpool L69 3GB
  2. 2Department of Primary Care, University of Liverpool, Liverpool L69 3GB
  3. 3Bute Medical School, University of St Andrews, St Andrews KY16 9ST
  1. Correspondence to: P Salmon
  • Accepted 9 March 2004


Objectives To identify the ways in which patients with medically unexplained symptoms present their problems and needs to general practitioners and to identify the forms of presentation that might lead general practitioners to feel pressurised to deliver somatic interventions.

Design Qualitative analysis of audiorecorded consultations between patients and general practitioners.

Setting 7 general practices in Merseyside, England.

Participants 36 patients selected consecutively from 21 general practices, in whom doctors considered that patients' symptoms were medically unexplained.

Main outcome measures Inductive qualitative analysis of ways in which patients presented their symptoms to general practitioners.

Results Although 34 patients received somatic interventions (27 received drug prescriptions, 12 underwent investigations, and four were referred), only 10 requested them. However, patients presented in other ways that had the potential to pressurise general practitioners, including: graphic and emotional language; complex patterns of symptoms that resisted explanation; description of emotional and social effects of symptoms; reference to other individuals as authority for the severity of symptoms; and biomedical explanations.

Conclusions Most patients with unexplained symptoms received somatic interventions from their general practitioners but had not requested them. Though such patients apparently seek to engage the general practitioner by conveying the reality of their suffering, general practitioners respond symptomatically.


About a fifth of patients who consult their general practitioner present with physical symptoms that the general practitioner thinks are not explained by physical disease.1 Nevertheless, these patients often receive extensive somatic investigation and treatment, which is largely ineffective and sometimes iatrogenic2 3 and which is usually attributed to pressure from patients for somatic treatment and cure.4 5 That general practitioners feel pressurised to offer somatic interventions helps to explain reports of their dissatisfaction with these consultations69 and the widespread use of terms such as “difficult” and “heartsink”10 to describe such patients.

General practitioners' subjective feelings of being pressurised for somatic intervention are, however, unreliable as evidence of how patients present. We identified types of presentation that had the potential to lead general practitioners to feel pressurised to give somatic interventions. We tested the specific assumption that these patients demand somatic intervention, and we identified other kinds of presentation that might trigger pressure by conveying the need for a response or the general practitioner's responsibility for a response or by constraining or directing the general practitioner's response.



We wrote to general practitioners from seven practices to ask them to take part. Of 30 contacted, 28 (13 women and 15 men, with experience ranging from 5-42 years, mean 18.4 years) agreed. Three practices were urban, three were suburban, and one was rural. List sizes ranged from 2180-13 116 patients, and each practice had from one to 10 doctors. The Jarman deprivation indices for the study areas were −11 to 38 (mean 17.6). Twenty one general practitioners and all participating practices were represented in the analysed transcripts.

There are no agreed research diagnostic criteria for primary care patients with unexplained symptoms. Criteria derived from psychiatric diagnoses of somatisation disorder are problematic because of poor agreement among them or poor discrimination compared with structured diagnostic interviews.11 12 A common procedure in UK studies is to select from patients with symptoms designated by the general practitioner as unexplained those who are psychologically disturbed and who respond to a single question about the causes of their symptoms by choosing “physical.”13 This procedure risks oversimplification, restricts concern to those who manifest psychological disturbance on a screening questionnaire, and assumes that patients readily distinguish physical from psychological causes. Because we focused on the difficulties that patients present for doctors, we used less restrictive criteria to identify patients who, in the doctor's opinion, have unexplained symptoms. Immediately after consultation (see below), the doctor completed a checklist1 to indicate whether or not the consultation involved: presentation of a physical symptom; symptoms that had existed for at least three months; symptoms that caused the patient clinically significant distress or impairment; and symptoms that could not be explained by a recognisable physical disease. We retained for analysis consultations that satisfied all four criteria. These criteria have face validity in that they can be readily understood and applied by general practitioners; they allow for the clinical reality that general practitioners vary in their decisions about which symptoms are unexplained by physical pathology; and they are strongly associated with the diagnostic criteria for abridged somatisation disorder.11 Although it is possible that some symptoms identified as “unexplained” might prove to have a pathological cause, our selection procedure ensures a patient group that is defined by their clinicians' belief that such a cause is absent.


Before consultation a researcher approached consecutive patients (n = 659) attending participating doctors on study days and asked for written consent for audiorecording their consultation; 110 were excluded (age under 16 years; inability to consent because of visual impairment, learning disability, or extreme distress), and 420 (77%) consented. Each doctor used a minidisk and microphone to record consultations with consenting patients. Twenty three consultations of consenting patients were not recorded because of equipment failure or operator error. The general practitioners completed the checklist immediately after each consultation, yielding 42 consultations for analysis. Of these, one recording failed and five transcripts contained insufficient discussion of physical symptoms. The 36 remaining consultations were anonymously transcribed, including all speech, and noting silences exceeding 10 seconds and simultaneous speech.


Analysis was inductive. Initial reading and discussion of 10 transcripts by all authors showed that overt demands for somatic intervention were rare. We therefore sought to identify more subtle ways in which patients' presentation might pressurise general practitioners. The common link between these forms of presentation was functional, in that they conveyed the patients' problems and needs in ways that were liable to pressurise general practitioners by obliging doctors to make a response, or by directing or constraining their responses.

One author (PS) carried out the preliminary analysis. This was then developed and tested by inclusion of subsequent transcripts, which were read and discussed by all authors. Analysis focused exclusively on verbal content in identifying recurring ways in which patients presented; it excluded non-verbal or contextual factors and avoided imputations of participants' motives. The analysis was unchanged by the final 16 transcripts. We have given examples to illustrate the range and commonality within each type of presentation. To demonstrate the completeness of the analysis we have shown the numbers of transcripts in which we identified each major type of presentation.

By cycling between data and the developing analysis, by involving authors from different disciplinary backgrounds in the analysis, and by presenting raw data to substantiate our findings we have enhanced the trustworthiness of the analysis. In addition, we aimed to maximise the coherence of the analysis and its “catalytic validity,”14 which refers, essentially, to its usefulness—that is, it should have the potential to change clinical practice or research.


Sample characteristics

We analysed transcripts from 36 patients (26 women) aged 19 to 81 years. All but two were white European. The most common symptoms were abdominal complaints (n = 10), pain in limbs (n = 9), or headaches (n = 7); others included chest pain, back pain, dizziness, fatigue, skin problems, and gynaecological or genitourinary symptoms (see table on Patients presented one to seven symptoms and received a range of somatic interventions including prescriptions (27 patients), investigations (12 patients), specialist referrals (4 patients), and sick notes (5 patients). Five patients were prescribed psychotropic drugs, and one was referred for physiotherapy.

Patient presentations

No patient asked for investigation or medical referral. Ten asked for physical interventions of other kinds: one sought physiotherapy, four sought repeat prescriptions, and six sought a new drug prescription—for example “I tell you what you could do while I'm here… something for heartburn.” Five requested sick notes. Although direct requests for somatic intervention were therefore few, patients presented in several ways that had the potential to pressurise general practitioners (box). All but one presented in at least one of these ways and most in several.

Twenty four patients routinely reported how their symptoms impaired activities of daily living or social behaviour:

At the moment all I'm worried about is my work, problem is my sleep, I can't really sleep… if I could sort my sleep out at least I'd be able to work. (P23)

Emotive words such as “nightmare” or “horrendous” conveyed the intensity of many patients' symptoms (n = 22):

I get terrible, terrible pains in my stomach… Just terrible, really sharp, sharp pains… it's really swollen and it's absolutely solid and it's excruciatingly painful… Sometimes it'll be like somebody has literally stabbed me. (P8)

This language extended to emotive metaphors or analogies:

I just feel as though somebody's got me by the back and the shoulders and it's just really horrendous round there. This side is worse, a lot worse than that side. (P15)

Nineteen patients proposed physical explanations for their symptoms. Some, such as “wind,” did not signify responsibility for the doctor. However, many, such as “pleurisy,” “arthritis,” or “ulcers,” signified disease that the doctor would be responsible for treating. Explanations were presented as proposals for discussion, rather than firm beliefs:

P2: I don't know where to start. I thought I took a heart attack a week last Saturday, severe chest pains. I rush myself into [hospital] and they said it wasn't the heart that was bothering me at that particular time they said “it's pleurisy in the chest.”

GP: Good.

P2: But the pains have just started again.

Nineteen patients communicated emotional distress associated with their symptoms, using words including “worry,” “distressing,” and “bothered.” They did not always attribute distress directly to the symptoms, but often attributed it to their uncertainty or fears about the cause:

I don't really know what's happening. Just worry about it all. (P27) The other thing is I'm worried… scared… is two of my aunties, two of my uncles and my gran had all died of lung cancer which all started with a lump in my [sic] neck. (P18)

Other individuals, usually family members, were cited as authority for the reality and severity of patients' symptoms and for the need for medical attention (n = 16):

With my stomach I keep thinking “oh it's probably just wind or something like that,” but with it being so sharp and so regular, really because my boyfriend said “you're going to have to go to the doctors you know it's getting bad.” (P8)

Potential sources of “patient pressure”

  • Effects of symptoms on patient's life

  • Graphic and emotive language

  • Biomedical explanations

  • Emotional distress caused by symptoms

  • External authority

  • Criticism and negation

  • Complexity of presentation

This included linking their symptoms to diseases of other family members:

P3: It's exactly the same symptoms as my mum has.

GP: Right, and what does she have.

P3: Rheumatoid arthritis.

No patient explicitly contradicted the doctor. Nevertheless, 17 negated in indirect ways the general practitioner's attempts to explain or manage their symptoms and, in particular, to exclude disease. Some offered additional perspectives that effectively invalidated the doctor's position:

GP: The first thing I would have checked was your gall bladder.

P7: That's what they checked.

GP: And I've seen that scan there and that was normal.

P7: But I mean at the time they took it, it was a good day so whether it's something that's flaring up.

Additional information about symptoms could also invalidate doctors' explanations:

GP: You know you talked about stress before. Are your headaches ever related to when you're feeling stressed?

P8: Yes it's quite possible that it could be… You see sometimes it's at the weekend though when I'm not, I could be lying on the couch…

GP:… So you can get the headaches when you're not actually at work?

P8: Yes but sometimes I'll have them all weekend.

Patients also negated doctors' explanations by offering alternative diagnoses or by emphasising the ineffectiveness of previous treatments:

GP: You don't get any buzzing in your ears or noises that aren't there?

P10: Yes.

GP: Cos that's called tinnitus and that sometimes goes along with occasional noises feeling a lot louder because it's to do with the nerve that supplies the ear.

P10: It's strange, oh God I can't explain.

GP: I don't know that there's anything I can really give you to make any difference.

P10: You know I'm putting it down, now I wonder if that's got something to do with me cholesterol or you know.

P15: I came to see you two weeks ago and you gave me some, not quite sure, some pills to sort of hopefully take the pain away from my shoulder. It's still there, as bad as ever.

GP: Right… So those tablets have done absolutely nothing.

P15: No.

GP: Not a sausage.

P15: No.

GP: Oh dear, that's a shame.

P15: I don't know why I'm taking them.

In addition, we noted complex temporal patterning of symptoms and striking complexity in the diversity of some patients' symptoms. Twenty one presented at least three somatic problems, including 11 who presented four or more. In many cases, the unfolding complexity of symptoms negated doctors' explanations. One patient introduced three distinct symptoms in the course of rejecting the doctor's attempt to attribute the first to stress:

GP: Anything else happening around the time when it [abdominal pain] first began?

P7: No.

GP: In your life or…

P7: No, nothing that I can, yes it could be stress maybe you know. Went through a bad time with my husband and that, everything's fine now but that was like two and a half years ago but maybe it could be that you know I suppose there's a lot of things… But I don't know whether it could be that because stress comes in different forms doesn't it—headaches, and I had a migraine the other day and I haven't had a migraine for a long time and then this kicked off and I thought “God.”

GP: Is it ever associated with migraines normally?

P7: No.

GP: No, that was just a coincidence?

P7: And you just think “oh I wonder what,” you know I wonder, I just don't know what's causing it. But I can't sleep, I can't sleep on that side, I always tend to sleep on this side, or if I sleep on my side I have to like—how can I explain it—not on my side fully, I have to sleep just like that you know what I mean, like that side's.

GP: Is there anything you've done that makes the pain easier?

P7: No. It seems to be all this side ‘cos I get throbbing in my leg as well, like, boom, boom boom all the time. Feel like a hypochondriac you know but you think “oh God what's to do with me.” You know there's something just not right. There's something definitely there.


Main findings in relation to existing literature

It is widely assumed that patients with unexplained symptoms pressurise doctors for symptomatic intervention and that this explains the difficulty or dismay that general practitioners feel in dealing with these patients and the high level of investigation and treatment that they provide.4 5 9 15 Our direct examination of patients' presentations does not support this assumption. General practitioners' provision of symptomatic intervention cannot be attributed to patients' overt demands.

As would be expected from previous evidence,3 all but two patients in this sample received somatic interventions. Most were prescribed drugs, a third were referred for investigations, and four were referred to hospital doctors. Nevertheless, none had requested investigation or specialist referral, and few asked for prescriptions. Explanation for why general practitioners treat such patients somatically, and why they apparently feel pressurised to do so, must be sought in other aspects of patients' presentation.

Sources of patient pressure

Patients did, indeed, have striking ways of conveying the extent and intensity of their suffering and need for help and of constraining doctors' attempts to help. They described their suffering with graphic words and metaphors, emphasised disabling effects of their symptoms, and cited friends and family as authority for their suffering or concern. These strategies reflect the difficulty of conveying the reality and nature of a symptom for which no objective evidence exists.16 However, unlike some strategies used by patients presenting unexplained symptoms in outpatient clinics1720—such as explicitly requesting treatment, claiming catastrophic consequences of being untreated, and blaming doctors for making symptoms worse—these patients' strategies do not overtly challenge the authority of the doctor over management decisions. Similarly, although these patients did enter one area that is traditionally the doctor's preserve—diagnosis—in offering biomedical explanations, they did not assert explanations but offered them as hypotheses. Patients' avoidance of overt challenge to the general practitioner's authority probably reflects the need to maintain long term relationships with general practitioners, which contrasts with the single opportunity that patients have to engage doctors in outpatient consultations.

As well as presenting the need for general practitioners to engage with their problems, patients constrained or impeded that engagement. They offered both confirmatory and conflicting evidence about physical and psychosocial explanations and presented symptoms that defied simple explanation or reassurance. The combination of instigation for, and constraint on, engagement helps to explain the feelings of pressure, difficulty, and helplessness that such patients provoke in doctors who seek to help them.5

Doctors' responses to patient pressure

Why did so many patients receive symptomatic intervention from doctors who believed that they had no disease? The subjective feeling of pressure and helplessness might be sufficient. Emotional distress promotes habitual behaviours rather than ones that are well thought out. For general practitioners, the habitual response is to intervene symptomatically. Alternatively, even though they did not overtly request it, perhaps patients sought such intervention and their doctors simply detected and met these wishes. There is, however, no evidence to substantiate this assumption about patients' goals, and it contradicts two kinds of evidence: patients consult doctors for explanation and support rather than medical intervention,16 21 and general practitioners are generally inaccurate in detecting patients' intentions.22 Other explanations are possible. The simplest is that patients sought to engage doctors with their problems.16 Therefore they struggled to convey the reality of their symptoms and, when simple explanations threatened to end the doctor's engagement, they presented with such complexity or intensity that engagement continued. General practitioners might then have responded symptomatically either because they mistook patients' insistence on engagement as desire for intervention or because they lacked another response to evident suffering. Finally, general practitioners' provision of prescriptions, investigations, and referral may be an attempt to establish authority by emphasising a role that remains exclusively theirs in a situation in which they feel powerless.5

Strengths and limitations of this study

Because our study was a qualitative analysis in a small sample, we must be cautious in making generalisations. Moreover, our decision to use doctors' assessments of whether or not patients' symptoms were unexplained means that our results cannot be directly compared with those from studies that used external assessments of the causes of symptoms. The proportion of patients with unexplained symptoms recruited to the study was lower than anticipated. Participating doctors suggested that patients with unexplained symptoms were less likely to consent to audiorecording. Our results may not therefore adequately characterise the full range of interactions with these patients.

Implications for research and practice

Our qualitative study, based on what patients and doctors said in consultation, cannot test our hypotheses about participants' goals. Nevertheless, we have shown that present assumptions about patients' goals are unlikely to be correct. If unnecessary symptomatic intervention is to be avoided in patients with unexplained symptoms, general practitioners will need educational interventions that reflect the ways that patients influence their decisions and that are based on a clear understanding of the goals that shape patients' presentations and doctors' responses.

What is already known on this topic

Many patients with unexplained physical symptoms receive unnecessary somatic investigations and treatment in primary care

This discrepancy has been attributed to patients pressurising doctors for somatic management

What this study adds

Most patients in this study did not request symptomatic interventions

The ways in which patients presented their symptoms had the potential to challenge general practitioners by conveying the need for a response while constraining the general practitioner's opportunity for response

The key to understanding why this group of patients receives somatic intervention lies in understanding why general practitioners respond to such challenge by offering somatic interventions

Embedded Image A table of symptoms presented by patients can be found on


We are grateful to T Hak for assistance in the design of the study, and for the enthusiastic participation of the anonymous general practitioners.


  • Contributors PS conceived the study, jointly developed the methods and led the analysis, and contributed to the preparation of this manuscript; CD jointly developed the methods and led the analysis; AR collected the data, contributed to the analysis, and led the preparation of the manuscript; GH jointly developed the methods and contributed to analysis and the preparation of the manuscript. PS and CD are guarantors.

  • Funding UK Medical Research Council.

  • Competing interests None declared.

  • Ethical approval Liverpool ethics committee.