Social class and the general practice consultation

Mary Boulton, David Tuckett, Coral Olson andAnthony Williams

Sodal class ai^ the general practice c<»isiiitatiiMi

Abstract After assessing the evidence on social class differences in therates and nature of general practice consultations, the BlackReport proposed that middle-class patients receive a betterservice from GPs than do their working-class contemporaries.Tiiis paper reviews the literature on which this sug^stion isbased and presents further evidence from a study of communi-cation in general practice, as to the extent and nature of sociald ^ s differences in consultations.

The results of the study suggest that middle-class patients aremore active than working-class patients in presenting their ideasto the doctor and in seeking further explanation of his viewsfrom him or her. However, this greater activity does notnecessarily mean that middle-class patients get more benefitfrom the consultation, at least in terms of its cognitiveoutcomes: similar proportions of working-class and middle-classpatients received explanations from the doctor and similarproportions misunderstood and rejected his views »id advice.These findings point to the important distinction between'process' and 'outcome' and un^rline the need for furtherresearch which assesses sodal dass differences in the outcomesas well as the processes of consultations.

IntrodvctMHi

It is now commonplace to observe that medical consultations can beviewed as exchanges between sodal actors who both use and aregoverned by sodal roles (Stimson and Webb 1975, Byrne and Long1976). As with any other sodal encounter, therefore, it may beargued that the sodal backgrounds of the individuals involvedpotentially influenoe the nature and outcome of the aimultation.llie hnport^ice of this for the provision of health care in Britain washighlighted recently by the Black Report on Inequalities in Health( 1 9 ) itKHigh primarily concer^Kl to examine the rede of sodal

326 Mary Boulton, David Tuckett, Coral Olson and Anthony Williams

and material factors outside the health care system, the workingparty also reviewed evidence on sodal class differences in theutilization of services in the NHS and, with regard to generalpractice in particular, the evidence on sodal class differences in thenature of the service received. The research they cited in relation tothe consultation presented patients from middle-class backgroundsas better able to communicate effectively with their doctors. Fromthis, they tentatively proposed that, in addition to having higherconsultation rates in relation to need, 'middle-class patients appearto receive a better service when they do present themselves thantheir working-class contemporaries (Townsend and Davidson 1982).The evidence on which they based this conclusion is limited,however, and they acknowledge that 'further analyses remain to becarried out.' The purpose of this paper, therefore, is to provideftirther evidence, from a study of communication in generalpractice, as to the extent and nature of sodal class differences inconsultations.

Most of the evidence on which discussions such as the BlackReport are based are concerned with the processes within theconsultation, whidi are most acce^ible for analysis. Only occaaonallyhave studies (Bain 1977) gone on to consider their outcomes.However, studies such as Stimson and Webb's (1975), show that theoutcomes of consultations may differ substantially from what mightbe expected from observations of the face-to-face interaction alone.A further aim of this paper, therefore, is to provide data on one typeof consultation outcome and to consider sodal class differences inthis light.

Sodal dass and the coit^tation: theoretical approaches

Although sodal class is one of the most commonly used variables insocial research, it remains a problematic concept with a variety ofmeanings relating to different views of the nature of sodal structureand a variety of measurement complications. Different omceptionsof sodal d a ^ variously emph^ize status and prestige, level ofincome, or power and authority as the bask for sodal distinctions.They also d^er both as to the number of da^es they Tecogmzc andas to wheie they draw t te major dividmg lines in sodety. Empiricalreseardi in Britain, however, has been doraiiiated by a view ofsodal dass as a grcniping of occupations whk^ share similar j iKaalstatus and which can be ordered with otiier groupngs of occupatiom

Sodal class and the general practice consultation 327

into a hierarchy of prestige. The most widely used method ofgrouping occupations has been the Registrar General's Classificationof Occupations. This method of assigning social class is very usefulin so far as occupations may be a reasonable indicator of manyaspects of lifestyle. It is also fairly straightforward to apply and hasbeen widely used in research and official statistics.

In interpreting the findings of studies which use this measure,however, it is important to bear in mind the conception of sodalclass which it embodies and the limitation inherent in it. Forexample, most systems for classifying occupations treat the householdas the unit of analysis and classify all members of the household onthe basis of the occupation of its (usually male) head. While thisapproach may work well for men, it may have less explanatorypower in the case of married women, whose class experience may bequite different from their husband's (c.f. the current debateGoldthorpe 1983, Heath and Britten 1984, Goldthorpe 1984,Erikson 1984). Other limitations that have been identified stemfrom theoretical disputes regarding the boundaries between classesand the delineation of class groupings (Abercrombie and Urry1983). In particular, the expansion of white-collar occupations(many of which involve women) and of white-collar membership oftrade unions has made problematic the major distinction betweenmanual and non-manual occupations (Allen 1982). The attitudesand orientations of such new groupings are less easily predictedfrom their position in the division of labour. One response to thistheoretical debate has been the development of alternative con-ceptions of social class such as the work of Goldthorpe and Hope(1974), which are more appropriate to a modem division of labourand can also be operationaUsed for sodal research.

A second response to this theoretical debate seems necessary inthe field of health. The perpetuation of inequalities in mortality andmorbidity is clearly revealed by current forms of measurementwhatever their limitations. At least as important as the refinementof the concept and measure of social class, therefore, is researchwhich attempts to identify and assign weights to the different sodalprocesses that are responsible for such inequalities. The point ofcontact between the individual and the doctor is another area inwhich conventional measin-ement has identified important differences.Yet several potentially competing interpretations have been givenfor such differences. Often the social processes at work in medicalconsultations have been left implidt in research and it is the task ofmedical sodology at this stage to examine more systematically

328 Mary Boulton, David Tuc^tt, Coral Olson and Anthony Williams

alternative perspectives on these processes which may have verydifferent implications. At least three different approadies to theways in which sodal class differences in consultations and theiroutcomes may arise can be identified in research in this area. Theseapproaches variously focus on the sodal skills of the patients, theirhealth knowledge and beliefs and their position as a client seekinghelp from a professional. These three, outlined below, provides auseful framework for reviewing the literature on sodal class andgeneral practice consultations and for considering the nature andextent of stxial class differences in the present study.

(a) The social distance approach. Perhaps the most common view ofthe influence of sodal class is couched in terms of the social distancebetween doctor and patient. Tliose who are similar to the doctors insodal class terms, it is suggested, are more likely to share theircommunication style and are therefore more likely to be able tocommunicate effectively with them. By contrast, working-classpatients are likely to find communication more difficult becausetheir communication style differs from that of the doctor and theylack the necessary sodal skills to negotiate the medical encountereffectively.^ In support of this view, a number of studies have foundthat worldng-dass patients say less than middle-class patients abouttheir complaints, and ask fewer questions of the doctor (Bain 1976,Cartwright and O'Brien 1976). TTiey have also been shown to getless information and fewer explanations from the doctor (Earthrowland Stacey 1977) despite the fact that they do not differ frommiddle-dass patients in wanting as much information as possible(Cartwright and Anderson 1981). This situation is seen to persist inpart because working-class patients are more diffident aboutexpressing critidsm and less articulate in conveying their needs andinterests to the doctor (Cartwright 1964, Blaxter and Paterson1982a). For their part, doctors have twen found to feel lesscomfortable with working-class patients (Bochner 1983), to sf^ndless time with them (Buchanan and Richardson 1973, Bain 1976,Cartwri^t and O'Brien 1977), ami to perceive more communicationdifficulties in their consultations (Jaspars, King and Pendleton1983). As a result of these features, it is su^ested, working-classpatients may not derive as much benefit from a medical consultationas do middle-class patients (Bain 1977).

While the empirical evidence dted here is fairly consistent inshowing working-class patients as less skillful in achieving theirgoals in the consultation,, such evidence needs to be interpreted withsome caution. Firstly, the measures which were used in these studies

Sodal claiss and the general practice consultation 329

were somewhat limited. Cotmting the number of units of expr^»on,for example, provides a useful index of the quantity of a givenactivity, but says little about its value or effectiveness; one predsequestion may be more useful to a patient than two vague ones.Secondly, many of the sodal class differences reported are small andnot statistically significant and in some ways point to a morefiindamental similarity in the experience of all patients. Thirdly,there is some suggestion that class differences in attitudes andbehaviour with doctors may be decreasing, as working-class women,particularly in the younger generation, may be adopting a morecritical and questioning approach to medical care (Blaxter andPaterson 1982b). Finally, the significance of some of the reporteddifferences for the outcomes of medical care is unclear: for example,while some studies suggest that middle-class patients may haveslightly longer consultations than working-class patients, otherevidence questions whether longer consultations necessarily lead togreater satisfaction with the encounter (Korsh and Negrete 1972) orsuggests that, on the contrary, they may be more predictive ofcommunication difficulties (Tuckett, Boulton and Olson 1985).

(b) The health knowledge and beliefs approach. A second, morecultural approach focuses on class-related differences in knowledge,attitudes and beliefs and their infiuence on the consultation. Earlyresearch documented defidendes in working-class patients' knowledgeof medical terms and diseases and suggested that this made itdifficult for them to understand the doctor (Samora, Saunders andLarson 1961, Boyle 1970). Taking a more relativistic approach,Friedson (1961, 1970) and, more recentiy, Kleinman (1978, 1980)have argued that patients do not lack knowledge but rather draw onlay knowledge and beliefs, developed in the context of their socialnetworks, which may differ from professional medical knowledgeand beliefe. Patients' presentation of their problem in the consultation,it is suggested, is shaped by their lay theories or lay 'explanatorymodels' while the doctors' clinical examination is shaped by theirmedical explanatory models. To the extent that lay and medicalexplanatory models differ, the doctor will have problems in elicitingwhat s/he feels is adequate information from the patient, and thepatient will have problems in getting across his or her own ideas andconcerns to the doctor. When the differences between lay andmedical models are not resolved in the consultation the patient mayalso reject the doctor's advice when s/he leaves the surgery. Suchproblems, it is suggested, are more likely to occur with working-class patients than with middle-class patients, since their culture is

330 Mary Boulton, David Tuckett, Coral Olscm and Anthony Williams

further from that of the doctor and their lay theories therefore moredivergent.^

This approach has proved particularly useful in illuminating thecommunication difficulties between doctors and patients from themore deprived and ethnically diverse groups in America (Harwood1971, Snow 1974). It is not yet clear, however, how applicable it is tothe situation in Britain, where the sodal structure and ethniccomposition are quite different. Recent work on lay concepts ofhealth and disease causation has indicated that, among the native-bom white population in Britain, the differences between the socialclasses may be quite small. Several studies have shown that evenpoorly educated working-class patients use sophisticated lay modelsof illness causation, which are 'often scientifically wrong in detailbut not in principle unscientific' (Blaxter and Paterson 1982b,Blaxter 1983, Pill and Stott 1982). In addition, little variation inconcepts of health has been found among sodal classes withinBritain (Williams 1983) although there may be greater variationbetween British and other cultural groups (Herzlich 1973). Furtherwork in this area may therefore show that lay models of illness oftendiffer from the clinical model used in a consultation, and hence giverise to confiict, misunderstanding and dissatisfaction, but that suchdifferences are not predictable on the basis of social class alone.

(c) Professional control approach. In contrast to the first twoapproaches, the final approach stresses the similarities in thesituation and experience of all patients. In the view of theprofessional control approach, the over-riding infiuence on theconsultation is the professional power of the doctor. Social classdifferences among patients are acknowledged as infiuendng theinteraction but such differences are seen as minor variations within apattem that is dominated by professional-client relations. Becauseof their status as experts, it is argued, doctors are able to maintainultimate control over the interaction. They do virtually all of theinitiating in the consultation, structuring it to elidt information as andhow they want it, while largely ignoring the patient's initiatives inpresenting information they do not require or in requestinginformation they are not willing to give (Coulthard and Ashby 1976,Hughes 1982, Drass 1982). Medical routines also help to structurethe consultation in a way that preserves the medical dominance(Bloor 1976) and potential challenges to medical authority generallylead the doctor to adopt interactional strategies to inaintain control(Stimson and Webb 1975, West 1976).

Sodal daffi and the general practice consultation 331

This tight control over the interaction means that what happens inthe consultation is determined almost entirely by the doctor and thatthere is little room for the patient's actions to infiuence it. Thus,Byrne and Long (1976) describe doctors as relatively stable andconsistent in the consulting style they used and as varying it little inresponse to differences among patients. Similarly, Boreham andGibson (1978) found that the extent to which a doctor providesinformation in the consultation depends less on what the patientsays or does and more on the doctor's perception of what s/he wantsto or should know. Pendleton and Bochner's (1980) findings thatdoctors volunteer more explanations to middle-class than toworking-class patients can also be seen to support the view that it isthe doctor who determines the content of the consultation, inaccordance with his or her views of what patients want to know andwhat they are capable of understanding. That more of the variancein the lengths of the consultations they studied was accounted for bydoctor- rather than patient-variables again underlines the doctor'sdominant role.

The theme of 'professional control' is not incompatible with theview that the patient's social class affects the interaction in theconsultation, particularly tiirough the medium of the doctor'sstereotypes. With the exception of Pendleton and Bochner's work,however, none of the studies reviewed in this section has used thesocial class of the patient in the analysis of its data. This refiects animportant bias in many of the detailed studies of doctor-patientcommunication. Interest has focussed largely on the intricacies ofthe interaction and the relationship of these patterns to the broadersocial context has often been overlooked. It is only by exploringthese relationships, however, that the role of wider sodal processescan be reoignized and the pattems themselves understood. Suchanalysis may also point to the importance of sodal variablesother than class: age, sex and ethnic background may all infiuencethe doctor's response to patients in consultations. Armitage,Schneidennan and Bass (1979), for example, compared 52 couplesin a group practice in Califomia and found that, in relation to fivecommon complaints, doctors made more extensive diagnosticefforts for men than for women. In another American study.Hooper and his colleagues (1982) found that a sample of hospitaldoctors gave more information to and showed more empathy withfemale patients; were more courteous to elderly patients; and werebetter at intervtewing patients from an Anglo-American, comparedto Spanish-American background.

332 Mary Boulton, David Ti»:kett,Q)ral Olson and Anthony Williams

HM general practice study: sam^e and nwtliods

The data presented in this paper are based on a sample ofconsultations gathered to investigate the educational potential andoutcomes of general practice consultations. The sample is composedof consultations by 405 patients with 16 doctors from a variety ofpractices in southeast England. Thirteen of the doctors were menand three were women. TTie consultations selected were those inwhich the doctor made a new decision about the nature of thepatient's problem or its treatment and included consultations abouta wide variety of problems. They were audio-recorded with thepermission of the doctor and patient and later transcribed. Bothtranscripts and recordings were used in rating the consultations. Thepatients were interviewed in their own homes within a week of theconsultation and these interviews were also recorded. For a varietyof technical reasons the interviews with 77 patients could not beused, leaving a sample of 328 patients for whom recordings of bothconsultations and interviews were available.(a) Sample: A number of background sodal variables were assessedfor each patient, including sodal class, ethnic background, gender,age and education. In this paper, we will focus largely on the socialclass of the patient in presenting the findings of the study, althoughan analysis in terms of the other sodal variables will be givenwhen particularly relevant. Sodal class was assigned accordingto the Hope-Goldthorpe classification of occupations, with themain distinctions (following Goldthorpe and Llewellyn (1977) andGoldthorpe et al 1978) drawn between classes I and II (serviceclass). III, rv and V (intermediate class) and VI and VII (workingclass). The 'service' and 'intermediate' classes were also combinedto form a 'middle class' group, to fadlitate a middle class/workingclass comparison. Tlie occupation used was that of the head of thehousehold: thus, women living on their own (or with their children)were d^sified according to their own occupation and other womenaccording to their husband's or father's occupation. Social classcould not be assessed for 11 patients.

The sample was fairly evenly divided among the three sodalclasses: 134 of the patients were in the service class, 133 in theintermediate class and 127 in the working class. The great majorityof patients, 351, (87%) were of Anglo-Saxon origin whUe only 32(8%) were from non-European back^ounds. About two thirds(269) were women. The sample was faidy young, with almost half(181) between 26 and 45 and a further quarter (105) between 46 and

Sodal da% and the general practice consultation 333

65; only 50 patients were over 66. About a fifth (75) had universityor equivalent professional qualifications, while over a half (216) hadonly basic educational quaUfications.

(b) Substantive measures. During the pilot stage, a series of ratingscales was developed to assess spedfic aspects of the consultationand its outcomes. These scales enabled us to consider our own datain relation to the issues reviewed in the previous section of thispaper.(i) Measures of patients' activity:From the recordings of the consultations, ratings were made of theextent to which patients, in their presentations to the doctor,employed each of the following strategies:^

1. indicating a lay diagnosis for the doctor's comment:Consultation 1866:Pa: I've had a few problems with my arm. I don't know whether

I've tom some tendons or not.Consultation 2214:Pa: She's got, what looks to me, the beginnings of athlete's foot

on the bottom of her foot. Whether it is or not I don't knowbut she says it's very itchy and very sore.

2. requesting ftirther explanation from the doctor:Consultation 1093:Pa: How long have I had it (thrush)? . . . Is that why I have a

blood discharge?Consultation 1284:Pa: Please, can you tell me what they (Naprosin tablets) are,

because I have been taking 3 at a time at first and they giveme nightmares . . . Anybody else had anything like that?

3. requesting clarification of instructions:Consultation 2220:Pa: Is that two in the morning and one in the evening?Consultation 1811:Pa: We've got penidllin to last until Thursday. Shall we

a>ntinue with that?4. doubting or disagreeing with the doctor:

Considtation 1367:Dr: I think this is highly likely to be a pulled muscle rather than

anything awful.Pa: I see. But wouldn't that make my knee hurt as well?Consulte^n 12M:Pa: If my blood pressure is all right.

334 Mary Bouiton, David Tuckett, Coral Olson and Anthony Williams

Dr: You should take those all the time.Pa: Yes, but let me tell you. The volume was 20, then they put it

up to 40 and these are 80. They wanted me to take an extrahalf. But if my blood pressure is all right, do I need to takeall that amount?

Rather than counting individual units of expression these scalesasse^ed the 'power' of the patient's overall presentation in relationto each strategy.(ii) Measures of doctor's explanations:The main measure for the doctors looked at the extent to which theyexplained and justified their views on the diagnosis, treatment orpreventive action advised. An explanation was defined as a reasonwhich linked the diagnosis, treatment or preventive action with thepatient's symptoms, or which ruled out alternative possibilities. Forexample:

Consultation 2214:Dr: No, it isn't athlete's foot. Athlete's foot is actually between

the toes, in there, where she is quite clear.Consultation 1598:Dr: Leave them (tablets for high blood pressure) off for a month

and let me recheck after that. Because if - your bloodpressure was up when I took it initially and I put you on thetablets. But at the moment it's not at all bad and last time,without tablets, it wasn't bad.

A 'more elaborate' explanation was one in which two or moreclasses of reasons were given. For example:

Considtation 1610:Mr. A., a 41 year old lorry driver, complained of pain in theelbow, previously diagnosed as tennis elbow, which had persistedfor 12 weeks. The doctor examined him, asMng him to push hishand down and then up. He continued:Dr: Well, what I'm doing is putting pressure on this long muscle

of yours, which is fixed to the bone there. That reallypinpoints it. It is tennis elbow. You get tennis elbow eithercaused by the muscles that you use to dose your fist or themuscle that opens it. Tlie two work very closely together. . . It does take a long time because it's a ligament that'stom and it has a very poor blood supply and thereforeeverything that happens is very slow, including healing.

Consultation 2391:Mrs. B., a 28 year old mother of a young baby opened the

Sodal class and the general practice comultation 335

consultation saying:Pa: Well, I'm here about this thrush. I've had treatment for it

twice and I've still got this rash.The doctor examined her and said:Dr: It looks like thrush. I don't know what sort of advice to give

you . . . I think we better think about doing some otherthings. I think quite often what happens is that you can getthmsh organisms in your bowel. I mean it actually growsinside you, in your gut, and so that's the source of yourinfection. So I think it might be worthwhile bashing you withNystatin by mouth . . . We may have been clearing it uppartially, but you have been reinfecting yourself. Because ifwe give you the Nystatin by mouth, you don't absorb it intoyour body, it just stays in your gut, and just fiows straightthrough. . . I think that the other way of perhaps trying totackle it is that - theoretically, thrush grows where it is niceand warm and so on. Some people who have been plaguedby the problem have been helped with wearing skirts andcotton pants and not nylons and trousers.

(iii) Consultation times:All consultations were timed. 'Total time' refers to the entire timethe patient was with the doctor. 'Conversation time' excludes longsilences, for example while the doctor was writing notes or thepatient was dressing.(iv) Mei^ures of patents' cognitive outcomes:From the interviews with the patients after the consultation, ratingswere made of the patients' interpretation of and commitment to thedoctor's views and advice. These measures were initially designed totest out the often implidt assumption that more explanation andjustification from the doctor leads to more understanding andacceptance of his views by the patient. They are rather complexmeasures and have been described in detail elsewhere (Tuckett et al1985). The measure of 'interpretation' looked at the way thepatient made sense of what the doctor said to him or her anddistinguished between those who understood it 'correctly', usingbiomedicd criteria, and those who misunderstood it in at least onesignificant way. Two case examples will serve to illustrate the waypatients 'misinterpret' the doctor's views:

Consultation 1468:Mrs. C , a 40 year old woman with a history of kidney problems10 years ago, made a second visit to the doctor in a week, 'still

336 Mary Boutton, David Tuckett, Coral Olson and Anthony Williams

feeling this sickness'. The doctor tested her urine and said'There's not much wrong with your kidne)^, actually, at themoment.' He nonetheless rang the hospital and arranged for herto be seen again by the ainsultant. When interviewed, Mrs. C.said that the doctor thought there was something seriously wrongwith her kidneys. Hie urine test had shown only that it was notan infection. In her own view, there was something pushing intoher kidney, causing a chronic problem which only a majoroperation could remedy. She was inclined to attribute this viewto the doctor as well.Consultation 2276:Mr. D., a 45 year old bus driver complained of stomach painwhich had been diagnosed as a stomach ulcer. The doctorrecommended Tagamet tablets, which 'stops you producingadd', and were to be taken continuously for 3 months. He alsoadvised a week off work in order to 'eat regularly'. Wheninterviewed, Mr. D. said he was to have a week off work becausehis stomach pain broke his concentration while driving. He wasalso to have a course of Tagamet tablets, which he understoodwould disperse the poison in his stomach and dry up his ulcer. Hedid not recall anything about 'eating regularly'.

The measure of 'commitment' looked at the way the patientevaluated what the doctor said and distinguished between those whoaccepted it and those who rejected it. Two further case examplesillustrate the way patients 'reject' the doctor's views and advice.

Consultation 1841:Mrs. E., a 29 year old woman took her 4 year old daughter to thedoctor because she had been coughing and sick at nursery school.Hie doctor said that the child's cough was not whooping cough,but just an 'ordinary' cough.When interviewed, Mrs. E. disagreed with the doctor, saying shewas sure it was whooping cough. The daughter of a friend hadthe same symptoms as her daughter and had been told it waswhooping cough. In addition, the nursery school teacher, whosees many more children than the doctor, and sees them for alonger time, said it was whooping cough.ConsidttUion 1367:Mrs. F., a 24 year old woman, consulted the doctor about a paindown her leg. The doctor dedded that the pain was a 'mu^ilarpain' and recommended r^ t and aspirin.When interviewed, Mrs. F. rejected the doctor's diagnosis.

Sodal class and the general practice consultation 337

saying it felt more like a nerve pain - 'a toothache in the leg' -because it had a 'shooting' sensation. She therefore also rejectedthe treatment advice as inappropriate. She thought a nerve waspinched by a displaced leg joint and wanted to have the leg'whacked back into place'.

Results

Analysis of the sample as a whole shows that the patients were farfrom passive. In addition to describing the presenting problem, overthree-quarters of the patients tried to influence the coui^e of theconsultation through at least one of the four patient strategies,though the manner in which they did so was generally subtle orweak. The doctors provided some sort of explanation on at least oneof the topics in over eighty per cent of the consultations and in halfof these consultations a 'more elaborate' explanation was given. Inonly half of all the consultations, however, did the patient bothcorrectly interpret and accept the doctor's views and advice.

The social class pattems within this picture are presented inTables 1 to 4. Table 1 shows that service-class patients were moreactive than their working-class counterparts in all of the strategiesexcept in expressing doubts and disagreement with the doctor'sviews. These class differences are largest with regard to indicating alay diagnosis for the doctor's comment and with regard torequesting further explanation from the doctor. Intermediate-classpatients were also more active than working-class patients in thesetwo areas, particularly in indicating to the doctor their laydiagnoses. TTiese results support the 'sodal distance' view thatmiddle-class patients are, on the whole, more active than working-class patients and therefore potentially able to gain more benefitfrom the consultation. In this context, the similarity between theclasses in their rates of registering doubts or disagreement with thedoctor is particularly interesting.

In contrast to the picture of patient activity. Table 2 shows noclear chss differences in the a c t i v e outcomes of the consultations:service- and intermediate-class patients misinterpreted or rejectedthe doctor's advice as often as working-class patients did. It appearsthat the greater activity of the middle-class patients during theconsultation does not result in greater understanding of or commit-ment to its outcomes when the a)nsultation is over. These findingssuggest that the distinction between process and outcome is an

338 Mary Boulton, David Tuckett, Coral Olson and Anthony Williams

Table 1: Patient Strategies and Sodal Class

PatientStrategies

1. Indicating lay *diagnosis

2. Requesting *furtherexplanation f

3. Requestingclarification ofinstructions

4. Doubting ordisagreeing

* p<0.01t p<0.05

Service

56 (42%)

123

61 (46%)

100

74 (55%)

130

73 (55%)

146

Social Qass

Intermediate

67 (50%)

(46%)

39 (29%)I I I •'

(38%)

56 (42%)t

(49%)

73 (55%)i

(55%)

Working

33 (26%)

34 (27%)

59 (46%)

73 (58%)

Total

156 (40%)

134 (34%)

189 (48%)

219 (56%)

important one, and requires more attention in research on theconsultation.

A further point of interest in these data relates to the discussion inthe 'health ^owledge and beliefs' approach of possible social classvariations in the lay theories which underpin patients' presentations.When the patients' accounts of their consultations were looked atclosely, it became apparent that their interpretation and evaluationof the doctor's views were shaped largely by their own lay theoriesof their problem and its management (Boulton 1981). Theyunderstood and accepted the doctor's views when they were inaccord with their own lay theories; when the doctor's views differedfrom their own, and these differences were not resolved in the(x>nsultation they generally misunderstood or rejected the doctor'scomments. The fact that there are no class differences in theproportion who misunderstood or rejected the doctor's views pointsto the hypothesis that, at least among the largely native-bompopulation of south-^ast England that we studied, working-class andmiddle-class patients drew on a similar body of lay knowledge andexplanatory modeb in making sense of their illness experience. The

Sodal dass and the general practice consultation 339

interview material also suggested that the lay models of causation,prognosis and treatment they used had been developed in relationto scientific medical thinking, though in any given consultation theywere as likely as not to confiict, in detail or particular application,with the doctor's models. This is not to say that the doctor's modelswere always more 'sdentific' or 'correct' in biomedical terms thanthe patient's. The 'operational' models which the doctors themselvesused drew heavily on commonsense knowledge and lay concepts ofcausation (c.f. Helman 1978).

Tables 3 and 4 show no class differences either in the duration ofthe consultations or in the extent to which the doctors explained andjustified their views. The greater activity which middle-class patientsengaged in was therefore not refiected cUrectly in longer consultationsor in more extensive explanations from the doctor. This would seemto support the 'professional control' view that the informationconveyed in a consultation is determined largely by the doctor and isaffected little by initiatives from the patient. There is some evidencethat patients who made an effort to elidt explanations from thedoctor - those who requested further explanation and those who

Table 2: Cognitive Outcomes and Social Class

Social Class

OutcomeService Intermediate Working Total

1. InterpretationAt least oneimportant element 32 (30%) 37 (34%)misunderstood

2. Commitment(rated on 'correct'interpretations only)

At least oneimportant element 16(15%) 20(18%)rejected

3. OveraUAt least oneimportant elementmisunderstood or 48(45%) 57(52%)rejected

43(42%) 112(35%)

13 (13%) 49 (15%)

56(54%) 161(50%)

340 Mary Boulton, David Tuckett, Ck>ral Olson and Anthony Williams

Table 3: Consultation Length and Sodat Oass

Using 'three class' Using 'two class'groupings groupings

Total Consultation r = -0.0218 r = 0.0238time p > 0.05 p > 0.05Total Conversationtime (excluding longsilences - eg. patient r = -0.0651 r = -0.0318undressing, doctor p > 0.05 p > 0.05reading notes)

Table 4: Doctor's Explanation and Patient's Social Class

Doctor'sExplanation

Moreelaborateexplanationprovided

Service

43 (32%)

80

Sodal Class

Intermediate

40 (30%)

(31%)

Working

43 (34%)

Total

126 (32%)

Less elaborate 66 (49%) 56 (42%) 51 (40%) 173 (44%)explanation ' y ^provided 122 (46%)No explanation 25 (19%) 37 (28%) 33 (26%) 95 (24%)provided ' ' ' . i '

62 (23%)Total 134(100%) 133(100%) 127(100%) 394(100%)

expressed doubts or disagreement - were more likely than otherpatients to be given more elatorate explanations (Table 5).However, about 40 i^r cent of such explanations were given topatients who did not make efforts to obtain them, and about 60 percent of those who tried to obtain more elaborate explanations werenot given them. As the following excerpts illustrate, doctors oftenignored their patients' doubts or requests for further e^lanation:

Sodal dass and the general practice c(Misultation 341

Consultation 2212:Dr: 'There certainly isn't any sign of a pleurisy developing.'Pa: 'No. As I say, it's been going on for the last fortnight at

least, you know.Dr: (intenupting): 'And I would take something like panadol or

disprin I think, just to ease the pain and otherwise leave itbecause I don't think any antibiotic is going to help.

Pa: (interrupting): 'Do you reckon the headache's something todo with that then?'

Dr: 'Yeah. Take panadol or something. . .'Pa: 'Yeah. As I say, it seems to be worse early on in the day,

when I get up in the morning and then after a couple ofhours,

Dr: (interrupting) 'Yes, yes.'Pa: 'Yes, it seems to go off.'Dr: 'Yes, yes.'Pa: 'O.K.'Dr: 'O.K.'

Consultation 1996:Mrs. G, a 24 year old woman, was feeling 'fat and unhappy' aftera course of steroids. The doctor offered to weigh her saying:Dr: 'It is a side effect of the tablets, though we had to give it to

you to stem the other illness. Cortizone does that to you.'Pa: 'I was sick quite a lot.'

Table 5: Patients requesting further explanation or expressingdoubts or disagreements and Doctor's explanation.

Doctor'sExplanation

More elaborateexplanationproviitedLess elaborateor no explanationprovidedTotal

Patients requesting explanation orexpressing doubts/disagreements

Yes No

75 (40%) 53 (24%)

112 (60%) 165 (76%)

187 (100%) 218 (100%)

Total

128 (32%)

277 (68%)

405 (100%)

1 = 0.35 1X0.01

342 Mary Boulton, David Tuckett, Coral Olson and Anthony Williams

Dr: 'Darling, we've got very little time; don't take your shoesoff, I'll allow a pound. Slip this off.'

Pa: 'Is the sickness caused by (ximing off the tablets?'Dr: 'O.K. 67V2 kilos. There's much more of you. Let me check

your blood pressure, because they can push your bloodpressure up too. . .'

Thus, while patients can infiuence to some extent the explanationsthey are given, doctors maintain, and exerdse, an over-ridingcontrol over what is discussed in the consultation.

Since the social class of the patient could not account for theexplanations given in a consultation, in the main study we alsolooked at other patient characteristics which might be of importance.The ethnic background of the patient appeared to have littleinfiuence: patients from Anglo-Saxon backgrounds were no morelikely than other patients to receive a more elaborate explanation.Similarly, the age category of the patient made little difference:there were no differences in the proportion of mothers presentingfor babies under two, mothers presenting for children under 10, andadults who were given more elaborate explanations. Women (aloneor with children) were no more likely than men to receive a moreelaborate explanation. The degree to which the patient was 'active'in the consultation made no difference in any of these analyses. Atpresent, then, it is still unclear why some patients were given moreelaborate explanations while similar patients, with similar problems,were not.

With regard to the cognitive outcomes of the consultation - theway patients understood and evaluated the doctors views - thepicture is much clearer. Misunderstandings of the doctor's viewswere associated with several characteristics of the consultationprocess (Tuckett et al 1985) but only one characteristic of thepatients: education. Patients with only the statutory minimumeducation were most likely to misinterpret at least one major topicof the consultation (7 = 0.34; p<0.05). This relationship isparticularly interesting given the lack of relationship betweenpatients' interpretation of the doctor's views and a measure of sodalclass based on oixupation. Education may be a better indicator ofthe more 'cultural' - as distinct from 'material' - aspects of sodalclass. A number of other characteristics were assodated with theway patients evaluated the oynsuttation. Those most likely to rejectthe doctors views were patients from minority ethnic b^kgrounds(-y = 0.49; p<0.05), patients over 60 (-y = 0.31; p<.05), and women

Sodal class and the general practice consultation 343

alone (y = 0.52; p<.05) or presenting for children under 10 (•> =0.44; p<0.05). These groups of patients are among those who wouldbe expected to have lay explanatory models differing from those ofthe doctor. Patients with little education and patients from minorityethnic back^ounds, for example, may be culturally more distantfrom the doctor and perhaps more likely to tum to folk models ortraditional explanations of and treatments for their problems.Women, on the other hand, have a responsibility for their family'shealth and, over the years, may become 'experts' in the area,particularly in relation to their children. In both instances, it isbecause they draw on altemative sources of medical thought whichhave an authority and accepted validity in their own social groupthat these patients 'misinterpret' or reject the views of the doctor.

IMscusdon

The overall picture which has emerged from this analysis can besummarised as follows: in making sense of their illness experienceand in discussing their complaints with the doctor, middle-class andworking-class patients alike drew on lay theories about their illnesswhich derived from sdentific medical theories. Middle-dass patients,however, were somewhat more active than working-class patients intrying to put these lay theories explicitly on the agenda of theconsultation and in seeking further explanation of his views from thedoctor. TTiis greater activity of middle-class patients did notnecessarily mean they got more benefit from the consultation, atleast in terms of its cognitive outcomes: similar proportions ofworking-class and middle-class patients received explanations fromthe doctor and similar proportions misunderstood or rejected thedoctor's views and advice. (This is not to say that middle-d^patients may not have been more successful in other terms, such asin obtaining the investigations, referrals or prescriptions theywanted.)

In looking at the influence of sodal class on the communicationproce^es and outcomes of the consultations studied, all threetheoretical s^^oaches outlined above were valuable. The approachwhich emphasized the 'sodal dktance' between doctor and patientwas useful m hi^bli^ting ^gnificant differences between the sodalelates in their styles of talking to the doctor. Taken on its own,however, outside the conteiA of the doctor's responses and theconsultati(Hi outojiiies, tiiis approach runs the risk of exaggerating

344 Mary Boulton, David Tudcett,CoralCMson and Anthony WUliams

the importance of these differences to the patient's experience.Indeed, they may be more important for the doctor's experience inthe consultation than for the patient's: the number of questions thedoctor faces varies a>nsiderably according to the socisd cl^^s of hisor her patient, while the extent of the explanations the patientreceives varies little with his or her sodal class. Patients mayattempt to control the consultation - and these attempts may varyaccording to sodal class - but the decision to respond to or to ignorethe patient's efforts remains with the doctor. The 'professionalcontrol' approach was useful in pointing to this &ct and in focusingon the common experience shared by all patients in relation to thedoctor.

The approach which focused on the explanatory models of thedoctor and patient, proved useful in accounting for the differentviews of the problem which they held following the consultation.The sodal class of the patient did not appear to be an importantvariable in explaining these differences, but a number of other sodalvariables, including age, education, gender, and ethnic backgrounddid. The more 'cultural' approach then, was valuable in pointing toa range of sodal variables which are important in shaping patients'experience with doctoi^. It was also valuable in drawing attention tothe content of comultations - what was said in the consultation andwhat was understood afterwards - which other approaches havegenerally ignored. Most research on the cognitive outcomes ofconsultations has focussed almost entirely on the quantity ofinformation conveyed or remembered (cf. Ley et al 1976, Bertakis1977, Andereon 1979) and has largely avoided Itxaking at the clinicalcontent of that information (cf. Fitq)atrick and Hopkins 1983,Fitzpatrick 1984). Not all information is of equal importance to thepatient, however, and assessments of consultations based oncounting items of information stated may be misleading. Similarly,limiting attention to the amount of information recalled does not dojustice to the way in which patients actively interpret and evEduatethe meaning of what the doctor has said, and may also give amisleading view of a>nsultation outcomes. By contrast, an approachwhich highlights the explanatory models of doctor and patientprovides a more meaningful basis for assessing consultation out-comes and one which more closely approximates the patient's ownperspective. Further research using this approach is clearly calledfor.

While the main analysis has been in terms of the patients' sodaldass, tl^se latter findings suggest that future work may benefit from

Sodal dass and the general practice consultation 345

taking account of the complex interactions between social class andother variables such as gender, age and ethnic background. Anotherdirection in which future work may go is towards closer examinationof the spedfic social and material conditions within class experience,which affect actions in response to illness. There is a great deal ofdiversity within the broad occupational groupings embodied inconventional measurement of sodal class and a more sensitivedifferentiation may prove useful in explaining variation. Looking atwholly working-class samples, for example. Pill and Stott (1982)found significant differences in attitudes to responsibility for healthin a sample of Cardiff women, according to level of education andhousing tenure. Similarly, Blaxter and Paterson (1982b) founddifferences in use of services in a sample of Aberdeen women, bymeans of an index of material and sodal deprivation. The currentmeasure of social class for women - dividing them along the Unes oftheir husband's occupation - is particularly problematic. Sodal classdistinctions developed with reference to the situation of men maynot always reflect meaningful distinctions among women. Oassifyingthem according to their husband's occupation, then, may serve toconfuse rather than to clarify our understanding of their experience(Oakley 1981). Since women consult general practitioners morefrequentiy than do men, the need for more precise and sensitivesodal indicators is clearly important. Research into general practi-tioners' perceptions of the sodal class of their patients, particularly'housewives', would also be of value.

Using conventional measures of sodal class, however, thefindings of the present study suggest that middle-da^ patients aremore active in the strategies they use in dealing with tfie doctor butthat this greater activity does not necessarily lead to greatereducational l^nefits from the consultation. This, it has beensuggested (Health Education Studies Unit 1982), is because doctorsdo not elidt or respond to the explanatory models of any patients,with the result that differences in lay and medical views persistthroughout the consultation and distort the patient's underetandingand evaluation of the doctor's advice. If this is the case, we wouldspeculate that future developments in general practice mightincrease the significance of the sodal class differences in thepatients' presentation. If styles of consulting move towards a more'negotiated' or 'mutual participation' model, the scope for patients'ideas, attitudes and behaviour to influence the consultation mayincrease. In this situation, the more active mode of presentationamcMBg middle-clt^s patients may put them at a definite advantage in

346 Mary Boutton, David Tuckett, Coral Olson and Anthony Williams

getting the most benefit from the consultation and sodal inequalitiesin medical care may be increased. In the meantime, however, thedata in this study on the whole do not support the thesis, as arguedin the Black Report, that middle-class patients are currently gettinga better service from ^ne ra l practice consultations than areworking-class patients."*

Most of the discussions about the benefits of health care todifferent sodal classes have focussed almost entirely on processeswithin the consultation: presenting problems, asking questions,giving information and explanations. With the increasing use ofaudio and video recorders in general practice surgeries, these arethe aspects of the consultation which are most readily accessible foranalysis. By contrast, consultation outcomes - whether they arecognitive, affective, behavioural or medical - require more effort toevaluate. At present, there are few studies which assess classdifferences in terms of the consultation's outcomes, and the sum ofthe evidence is equivocal. For example, whilst Bain (1977) founddifferent levels of knowledge of diagnosis, dmgs and medical advicein the upper and lower sodo-economic grou{», our study found nosignificant differences in the proportions of middle and working-class patients who understood and accepted the doctor's diagnosis,treatment and preventive advice. The overall conclusion, therefore,has to be that the contribution of the social sdences must followtrends within medical practice where emphasis is increasingly uponthe evaluation of the outcomes of procedures (cf. Cochrane 1971,Donabedian 1980, Butler and Vaile 1^4) as well as on the processesof care themselves.

Department of Community MedicineSt. Mary's Hospital Medical School, Praed Sti'eet,

London W2 lPGUK

paper is a revised and extended version of a paper prepared fora Joint Conference of the British Sociological Association and theRoyal College of General Practitioners, Eynsham Hall, Oxford,Nov. 29-30,1983.

Hie researdi was carried out t^ule aU the authors were at theHealth Education Studies Unit, Hughes Hall, Cambridge. Hie Unitwas entirely funded by t l ^ Health Education Cbundl. We would

Sodal class and the general practice considtation 347

like to thank David Armstrong, David Blane, Michael Bury andRay Fitzpatrick for their helpfril comments on earlier drafts of thepaper.

Notes

1 A variation on the 'sodal distance' approach, deriving from the work of BasilBernstein, sees sodal class differences in presentation as related to morefundamental differences in the *deep structures' of communication itself.Bernstein has argued that particular features of the sodal context in which themiddle- or working-class child grows up orientate him towards quite differentmodes of language use. On the one hand, working-class children learn to use a'restricted code' in vtiiich meanings are implidt and context-tied, and so fullyunderstandable only to those who have access to their context. On the otherhand, middle-class children learn to use both a restricted and 'elaborated code' inwhich meanings are explidt and context-firee and so understandable to all.According to Bernstein's theory, these differences in language use, and the sodalrelationship realized in them, mean that communication between working-classand middle-class individuals is difficult. In particular, working-dsiss patients havedifficulty in performing e^ctively in situations such as the consultation which arepredicated on an elaborated code and its system of sodal relationships. (See, forexample, Bernstein, 1971; Bernstein, 1972, and Robinson and Rackstraw, 1978.)Bernstein's theories have been invoked to predict and to account for majorcommunicaticm difficulties when working-dass patients encounter the middle-class style which dominates medical institutions. (See for example, Tuckett,1976, and Hauser, 1981.) They have never been tested directly in theconsultation, however, so it is difficult to assess their explanatory power in thatcontext. Bernstein's work has also been widely oitidzed in recent years and the'superiority' of the elaborated code for expressii^ ideas clearly has beenquestioned (see for example, Labov, 1972).

2 Kleinman, for example says:The study of the interaction between practitioner EMs [explanatory models] andpatient EMs ( ^ r s a more p^edse analysis of prcMems in dinical communication. . .[EMs] differ in analytic power, level of abstraction, logical articulation,metaphor and idiom. . . Obvioudy enthnidty, sodal class and educationinfluence choice of metaphor and idiom (Kleinman, 1980:105,107)

3 Details of the concepts and measures veloped in the study can be found inHealth Education Studies Unit, (1982). In this paper, the original 4 and 7 pointscales have been amplified to (Uchotomous 'present-absent' scales. The onlyexception is 'requesting further expixa&tioa from the doctor' where thedistinction is drawn between Uiose deto'ty requesting an e3q)lanation and thosenot requesting one or only hinting they would like one.

4 The analysis in this ptper is in tenns of individual patients. It is possible that theservice provicted by HK NHS varies according to the sodal dass of the area rathertbaa of the indS^mteal patioit, widi all patients in a predominantly working-dassarea tting a pooier service than aH jratients in a predominantly middle-dassarea. (See J. Tudor Hart, 1971; and Skrimshire, 1978 dted in Townsend andD

348 Mary Boulton, David Tuckett, Coral Olson and Anthony Williams

Abercrombie, N. and Urry, J. (1983) Capittd, Labour and the Middle ClassesLondon: Geoi^e AUen & Unwin.

Allen, S. (1982) Gender inequality and class formation. In Giddens, A. andMackenzie, G. (eds) Social Class and the Division of Labour Cambridge:Cambridge University Press.

Anderson, J. (1979) Patients' recall of information and its relation to the nature ofthe consultation. In Oboume, D. , Gruneberg, M.M. and Eiser, J.R. (eds)Research on Psychology and Medicine 2 Linidon: Academic Pre^.

Armitage, K., Sdmeiderman, L.J. and Ba^, R. (1979) Response of physidans tomedical complaints in men and women. Journal of the American MedicalAssociation, 241,2186-7.

Bain, D.J.G. (1976) Doctor-Patient communication in general practiceconsultations. Medical EductUion, 10,124—31.

Bain, D.J.G. (1982) Patient knowledge and die content of the consultation in generalpractice. Medical Education 11, pp. 347-50.

Bernstein, B. (1971) Class, Codes and Corttrol: Volume I, London: Routledge &Kegan Paul.

Bernstein, B. (1972) Sodal class, language and socialisation in Gigliomini PP (ed)Language and Social Context, Harmondsworth: Penguin.

Bertakis, K. (1977) The communication of information from physidan to patient:A method for increasing patient retention and satisfaction. Jounud of FamilyPractice 5,211.

Black Report (19^) Report of the Working Party on inequaMes in Health,London: DHSS.

Blaxter, M. and Paterson, E. (1982a) Consulting behaviour in a group of youngfamilies. Journal of the Royal College of General Practitioners 32,657-62.

Blaxter, M. and Paterson, E. (1982b) Mothers and Daughters: A three-generationalstudy of health attitudes and behaviour, pp. 159-163 London: Heineman.

Blaxter, M. (1%3) Hie causes of disease: women talking. Social Science andMedicine, 17,59-69.

Bochner, S. (1983) Doctors, patients and their cultures, in Pendleton and Hasler(eds) Doctor-Patient Communkation. London: Academic Press.

Boreham, P. and Gibson, D. (1978) The informative process in private medicalconsultations: a preliniinary investigation. Social Science and Medicine, 12,409-16.

Boulton, M. (1981) Lay Theories and Ae Consultation Process. Paper presented atthe BSA Medical Sodology Conference, York University, York.

Boyle, C. (1970) Differences between patients' and (toctors' interpretation of somecommon medical terms. British Medical Jourrud 2,186-9.

Buchan, I. and Richardson, I. (1976) Time Study of Consultations in GeneralPractice, Scottish Health Studies No 27, Scottish Home & Health Dept.

Butler, J. and Vaile, M. (1984) He(dth and HecJth Services: An Introduction toHealth Care in Britain, London: Routledge & Kegan Paul.

Byrne, P. and Long B. (1976) Doctors taking to patients. London: HMSO.Cartwright, A. (1964) Human Relations and Hospital Care, London: Routledge

& Kegan Paul.Cartwright, A. and O'Brien, M. (1976) Sodai dass variations ia health care and in

Sodd class and the generd practice consultation 349

the nature of nersd practice constdtations, in Stacey, M. (ed), 7%e Sociologyof the National Hetdth Service, Sodology Review Monograph No 22, KeeleUniversity, Keele.

Cartwri^t, A. and Anderson, R. (1981) General Practice Revisited. London:Tavistock.

Cochrane, A.L. (1972) Effectiveness and Efficiency: Random Reflections on theHeaM Service, London: Rock Carling Lecture.

Coulthard, M. and Ashby, M. A. (1976) Linguistic description of doctor-patientinterviews in Wadsworth M. and Robinson D. (eds) Studies in Everyday MedicalLife. London: Martin Robertson.

Donabedian, A. (19^) The definitions of Quality arul Approaches to itsAssessment. Health Administration Press: Ann Arbor.

Drass, K. (1982) Negotiation and the structure of discourse in medicalconsultations. Sociology of Health and Illness, 4,320-41.

Earthrowl, B. and Stacey, M. (1977) Sodal class and children in hospital SocialScience and Medicine, 11,83-8.

Erikson, R. (1984) Sodal class of men, women and families. Sodology 18 ^X).Fitzpatrick, R. and Hopkins, A. (1983) Problems in the conceptual framework of

patient satisfaction research: an empirical exploration. Sociology ofHeidth and/ffness, 5,297.

Fitzpatrick, R. (1^4) Satisfaction with health care. In Fitzpatirick R, et al 7%eExperience of Illness, London: Tavistock.

Friedson, E. (1961) Patients' Views of Medical Practice. New York: Russell Sage,Foundation.

FriedscHi, E. (1970) Profession of Medicine. New York: Dodd Mead.Goldtiiorpe, J.H. and Hope, K. (1974) Hie Social Grading of Occupations: A New

Approach and Scale. C>xford: Oxford University Press.GoldAorpe, J. and Uewellyn, C. (1977) Qass mobility in modem Britain: Three

theses examined. Sociology 11,257-87.Goldthorpe, J. Payne, C. and Llewellyn, C. (1978) Trends in class mobility .

Sodo/ogy 12,441-68.Goldthorpe, J.H. (1983) Value and limitations of using their husband's occupation

in assigning sodal class to married women. Sociology 17.Goldthorpe, J.H. (1%4) Women and dass analysis: A reply to the replies.

Sociology 18,491.Harwood, A. (1971) TTie hot-ojld Aeory of disease: implications for treatment of

Puerto Rican patients. Journal of the American Medical Association 216,1153-8.Hauser,S. (1981) Hiysidan-Patient relationships. In Mishler, E. et al. Social

Context ofHeiM Illness and Patient Care. Cambridge: CUP.Heath, A. and Britten, N. (1984) Women's jobs do make a difference. Sociology

18,475.Health Education Studies Unit (1982) Hrud Report of the Patient Project London:

Health Education Coundl.Helman, C. (1978) 'Feed a Cold, Starve a Fever' - Folk models of infection in an

English suburban community and their relation to medical treatment. Culture,MetHdne and Psychimry2,107-37.

BetzMi, C. (1973) HeMi and IUness. London: Academic Press.Hooper, E., Comstock, L., Goodwin, J. and Goodwin, J. (1982) Patient

(Wacteri^cs that influence {diy^dan beha\^our. Medical Care 20,6X

350 Mary Boulton, David Tudcett, Coral Olson and Anthony Williams

s, D. (1982) Control in the medkal consultation: organiang tdk in asituation where co-partidpants have differential competence. Sociology 16,359-76.

Kleinman, A. (1980) Patients and Hetders in the context of Cidture, London:University of California Press.

Kleinman, A., Eisenberg, L. and Good, B. (1978) Culture, IUness and Care:clinical lessons from anthropologic and cross-cultural research. Annals ofInternal Medicine 88,251-8.

Korsh, B. and Negrete, V. (1972) Doctor-patient communication. ScientificAmerican 117,66-74.

Labov, W. (1972) TTie logic of non standard English In Gigliomi, P.P. (ed)Language and Social Context Harmondsworth: Penguin.

Ley, P., Whitworth, M.A., Skilbeck, C.E., Woodward, R., Pinsent, R., Pike, L.,Qarkson, M. and Qark, D. (1976) Improving doctor-patient communication ingeneral practice Journal of the Royal College of General Practitioners 26,720.

Oakley, A. (1981) Subject: Women, Oxford: Martin Robertson.Pendleton, D. and Bochner, S. (1980) The communication of medical information

in general practice consultations as a function of patients' sodal dass. SocialScience and Medidne 14A, 669-73.

Pill, R. and Stott, N. (1982) Concepte of illness causation and responsibility. Somepreliminary data from a sample of working dsss mothers, Socml Science andMedicine 16,43-52.

Robinson, W.P. and Rackstraw, S. (1978) Social da% differences in posingquestions for answers. Sodology 12, pp. 265-80.

Samora, J., Saunders, L. and Larson, R. (1961) Medical vocabulary knowledgeamong hosfMtal patients. Journal of Health and Social Behaviour 2,83-92.

Skrimshire, A. (1978) Area Disadvantage, Sodtd Class and the HeaMi Service,Department of Sodal and Administrative Studies, Oxford University.

Snow, L. (1974) Folk medical beliefs and their implications for care of patients.Annals <rf Intern^ Medidne 81, ^ - 9 6 .

Stimson, G. and Webb, B. (1975) Going to see the Doctor. London: Routledge &Kegan Paul.

Townsend, P. and Davidson, N. (1982) Inequalities in Health. Harmondsworth:Penguin.

Tuckett, D., Boulton, M. and Olson, C. (1985) A new approach to themeasurement of patients' imderstanding of what they are told in medicalconsultations. Journal ofHeakh and Social Behaviour 26,27-38.

Tuckett, D. (1976) An Introduction to Mediad Sodology, London: Tavistock.Williams, R. (1983) Concepte of health: an analysis of lay lope, Sodology 17,