POSTGRAD. MED. J. (1965), 41, 401

STUDIES OF SOCIAL ATTITUDES AND VALUESIN RELATION TO PSYCHIATRIC EPIDEMIOLOGY

K. RAWNSLEY, M.B., M.R.C.P., D.P.M.

Professor of Psychological Medicine, Welsh National School of Medicine, (formerly memberof scientific staff, M.R.C. Social Psychiatry Research Unit, Llandough, Penarth, S. Wales).

FOR MANY years the annual rate of first admis-sions to psychiatric hospitals in England andWales has shown a steady rise (RegistrarGeneral 1964). Changes in the legal and ad-ministrative arrangements for the care and treat-ment of mental disorders are probably account-able for the greater part of this rising influx. TheMental Treatment Act 1930 established volun-tary admission and enabled local authorities toset up out-patient clinics. The development ofthe National Health Service brought a sub-stantial enlargement of the specialist establish-ment in psychiatry. More recently the MentalHealth Act 1959 has abolished the special statusof psychiatric hospitals and has removed allformality from the admission procedure for allexcept a minority of patients.

These changes in social policy towards thementally ill may arise in part out of ratherwidespread changes in attitudes towards mentalillness among the general population. Imple-mentation of the policies must, in turn, serveto generate modifications in social attitudes.The passage of patients into and out of the

specialist mental health services and the factorswhich play upon and determine this movementrepresent one stage of a cycle which beginswith the earliest recognition by the individualpatient (or by his relatives) that something iswrong. The process continues when the abnor-mality is reckoned by the patient, or by hisfamily, to have medical significance, and whena decision is made to seek advice, usually fromthe general practitioner. A further stage isencountered in the appraisal of the case bythe G.P. and in his diagnosis and treatment,and in his decision whether to handle the casehimself or to refer the patient for a psychiatricopinion.

Part of the work of the M.R.C. SocialPsychiatry Research Unit (S. Wales detach-ment) has been to examine certain aspects ofthe elaborate social process whereby psychiatric

cases are defined in the community, recognisedby community members and by medical andsocial agencies, and dealt with by one meansor other. Studies of this kind are very relevantto the epidemiology of mental disorder since,by the nature of such illness, the detection andenumeration of cases is intimately linked withprevailing social "yardsticks" pertaining to theacceptable bounds of "normal' behaviour andexperience and also to the categorisation ofdeviant behaviour as falling within the doctors'province.Attitudes to the Psychiatric In-patientA feature of post-war British psychiatry has

been the mobilisation and rehabilitation ofmany long-stay patients in mental hospitals.The increasing number of cases with residualsymptoms living either with their families, or atinstitutions in closer proximity to the localpopulation than traditional mental hospitals,demands much tolerance and sympathy fromrelatives, friends, neighbours and members ofthe general public. The Mental Health Act1959 emphasised the desirability of developinga comprehensive scheme of care and treatmentin the community. The success of this policywill depend, in part, on the attitudes whichprevail towards mental disorder in thecommunity.An enquiry was carried out in S. Wales to

examine the attitudes of relatives to a familymember in a mental hospital and especiallythose attitudes which may influence familybehaviour if and when the time comes for thepatient to leave hospital (Rawnsley, Loudonand Miles, 1962). The survey was basedupon all patients, 230 in number, who were inmental hospitals at the time of the enquiryand those whose home addresses on admissionlay in one of three defined areas in S.E. Wales-a mining valley, a rural area, and a smalltown.

by copyright. on D

ecember 24, 2021 by guest. P

rotectedhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.41.477.401 on 1 July 1965. Dow

nloaded from

POSTGRADUATE MEDICAL JOURNAL

Patients were assessed by the hospital staff interms of their social capabilities and poten-tialities. Attitudes of relatives were studiedprincipally by home interviews conducted witha sample of relatives from each family and bythe examination of records of contacts-visits,letters, parcels--between relatives and patients.Twenty per cent of the patients, many of

them with very long hospital stay, were foundto lack any contact with their families for thepast year or more. It was a little surprising,therefore, to discover relatives of thesepatients-usually close relatives-living locally.Furthermore, offers of accommodation in theevent of discharge were forthcoming in aquarter of the cases where contact had beenlost for so long.6.1% of the patient sample had been con-

tinuously in hospital for a year or more, werefit to live at home (according to the hospitalstaff) and had a home waiting for them (accord-ing to the family). Although on the face of itthese patients would appear to be promisingcandidates for rehabilitation, certain reserva-tions must be entered. There is evidence thatcontact with close relatives at home may havea deleterious effect upon schizophrenics(Brown, Carstairs and Topping, 1958). Disquiet-ing reports have also been published of the dis-rupting effects on the family of the returningpatient with residual disability (Wing, Monck,Brown and Carstairs, 1964).These studies, carried out in an area where

the community aftercare services were inade-quate, indicate very plainly the need for inten-sive supervision of the partially remittedschizophrenic after leaving hospital.

In the S. Wales investigation, analysis offrequency of visits by relatives and of theirwillingness to accommodate patients in theevent of discharge by age of patient and byduration of stay in hospital indicates that,although interest expressed through visiting issustained in the elderly group of patients, this isnot matched by willingness to house them.Judging by the response of relatives, it is clearthat a policy which seeks to discharge long-stayelderly patients of either sex must look foraccommodation outside the family. The pro-vision of special hostels for the elderly long-staypatient with mild symptoms is an obvioussolution. Social work with families would bebest reserved for the younger group whatevertheir length of stay in hospital.

Married schizophrenics of less than two-years' stay command a higher level of activeinterest (as judged by visiting frequency) and

better prospects of accommodation on dischargethan do single schizophrenics. This findingilluminates earlier statistical enquiries into thedifferential probability of discharge for singleand married patients. Thus Norris (1956)showed that amongst schizophrenic admissionsin 1947-49 to some London hospitals, singlepatients had the greatest chance of retentionfor two years and that the married had least,with the widowed and divorced occupying anintermediate position. Brooke (1959) showedthat 19%/ of single schizophrenic first admis-sions in England and Wales in 1954-56 werestill alive and in hospital two years later, while12% of the patients who had been married atsome time were likewise retained. Wing,Denham and Monro (1959) found that singleschizophrenic patients in two cohorts of admis-sion to Long Grove Hospital had a worseprognosis for discharge within two years thanmarried patients.Comparison of the attitudes of relatives to

patients from the three geographical areas inS. Wales indicates that despite equivalence of"active interest" revealed through visiting,patients from the town have a substantiallylower proportion of relatives willing to housethem on discharge than have patients from therural area. Patients from the mining valley areintermediate in this respect.One important factor not systematically ex-

amined in this study is the attitude of thepatients themselves to their future and, inparticular, to the prospect of leaving hospital.Although many patients would undoubtedlyprefer an early discharge, others are too appre-hensive or too settled to want it. Folkard(1960) found that the expectations of a groupof selected chronic patients regarding prospectsof discharge and level of performance afterdischarge, were more optimistic than thoseexpressed by their relatives. Wing, Bennett andDenham (1964) have shown that the desire ofsome chronic male schizophrenics to leavehospital can be substantially sharpened byexposing them to an Industrial RehabilitationCourse at a Ministry of Labour Unit.The General Practitioner

Patients with psychiatric illness form a subs-tantial fraction of the G.P.'s case load. In asurvey carried out by the General RegisterOffice in collaboration with the College ofGeneral Practitioners (Logan and Cushion,1958) 171 volunteer G.P.s kept an account oftheir contacts with patients, including a recordof the nature of the illness, for one year.

402 July, 1965by copyright.

on Decem

ber 24, 2021 by guest. Protected

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.41.477.401 on 1 July 1965. D

ownloaded from

RAWNSLEY: Social Attitudes and Psychiatric Epidemiology

During the year, 5% of all patients on thepractice lists consulted their doctors for anailment regarded as psychiatric. This is anaverage value for a large number of practices.Other studies by individual G.P.s reveal a largevariation in the estimates of psychiatric morbid-ity in different practices which can be accountedfor, in part, by the way the frequency of caseshas been calculated and expressed (Kessel andShepherd, 1962). Several other factors probablycontribute to this variation and the influenceof any single factor is difficult to assess. Thepractice populations may differ in their dis-tribution by sex, age, occupation, social andgeographical mobility and other characteristicsknown to be related to the occurrence ofpsychiatric disorder. The range of psychiatricdisorders included in the count is subject towide variation, depending on the aim of theenquiry and the views of the observer, and isdifficult to control because of differences in theusage of diagnostic terms.

In addition, doctors probably differ in theextent to which they perceive a particularcondition as psychiatric, especially if thepatient presents initially with what appear tobe "physical" symptoms, such as pains in theabdomen, back or head. The influence ofcertain characteristics of the G.P. upon recogni-tion of psychiatric cases has been elegantlydemonstrated in the report by Mowbray, Blair,Jubb, and Clarke (1961) of a pilot study bythe West of Scotland Faculty of the College ofGeneral Practitioners and the Department ofPsychological Medicine, University of Glasgow.Fifteen G.P.s in seven groups or practicessampled their practices at the point oftermination of each transaction with a patient(including obvious terminations and also,where less obvious, points at which therewas no undertaking to see the patientwithin a period of one month). Diagnoses werecategorised as 1. physical illness (62%); 2.psychological factors in physical illness (21%);3. psychosomatic illness (5%); 4. psychiatriccases (3%); 5. personal problems (2%); 6.other category (7%). A significant negativecorrelation (r = -0.7) was found betweennumber of years since medical qualification andthe proportion of patients placed in combinedcategories 2 and 3. The corresponding correla-tion between years since qualification andproportion in category 4. was also negative,but not significant ;(r = -0.27). The level ofinterest in psychiatry of the 15 doctors wasrated on a 5-point scale by an official of theCollege who knew all the participants. A

positive association was found between levelof interest and the proportion of patients placedin each of the categories 2, 3 and 4 separately.Having recognised or suspected the existence

of a psychiatric disorder, what factors weighwith the G.P. in his decision to refer a patientfor psychiatric opinion and treatment? In astudy by Kessel (1960), neurotic patientsreferred to hospital by a group of G.P.s didnot differ in the form of their illnesses frommany of those not so referred. It was difficultto decide what factors had led to their referral.In the enquiry reported by Mowbray and others(1961) a series of G.P.s' letters referring patientsto a psychiatric clinic were analysed. The find-ings suggested that few practitioners referredpatients on the basis of a positive diagnosticappraisal, tending rather to stress abnormalitiesof conduct, the existence of social problemsor inappropriate responses to medical attentionas reasons for referral. It was concluded thatvariation in type and number of referrals couldbe due to a wide variety of attitudes to psy-chiatry on the part of practitioners.An interesting opportunity occurred in

S. Wales to make a close comparison of thefactors influencing referral of patients to psy-chiatric services from six general practices,including eight practitioners, situated in thesame mining valley (Rawnsley and Loudon,1962, a and b). Information about casesreferred during the period 1951-59 was gatheredfrom hospital and clinic records. This includedclinical data, name of G.P. and a statementas to whether the patient was referred directlyby the G.P. to a psychiatrist or came by wayof another specialty, e.g., the general medicalclinic. Data about the practice populationswas gathered principally from a private censustaken throughout a defined area in the miningvalley by Professor A. L. Cochrane of theMedical Research Council EpidemiologicalResearch Unit. One of the items on the censusschedule requested the name of the G.P. Thepopulations of the six practices were found tobe closely similar in their distributions by sex,age, occupation, number in household, placeof birth and education.

Despite this homogeneity, the rate of referralof patients directly to psychiatric services showsa subtantial variation among the practices,so that, for females, the highest rate (36.8)*is almost twice the total average (19.4) andmore than three times the lowest (10.8). Onereason for this diversity of rates could be the*Average annual rate of direct referrals per 10,000population at risk.

July, 1965 403by copyright.

on Decem

ber 24, 2021 by guest. Protected

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.41.477.401 on 1 July 1965. D

ownloaded from

POSTGRADUATE MEDICAL JOURNAL

selective recruitment of psychiatric casesto those G.P.s regarded by the populationas being especially competent or sympatheticin handling such problems. This hypothesisis not supported by evidence availablefrom the material. It may be supposedthat doctors differ in the criterion of clinicalseverity which they apply in deciding to refera patient, or in the relative proportions ofdiagnostic categories referred. The findings,however, show no significant difference bet-ween the six practices in either of these factors,nor in the distribution of referrals by age,civil state or occupation.A clue to the factors which may influence

referral came from interviews conducted withthe G.P.s themselves. The commonest reportedfactor was the failure to respond to treatmentprovided by the G.P. However, a medley of"non-clinical" factors was also mentioned, eachof which appeared to weigh in varying degreewith individual doctors. Examples are, 1. pres-sure from relatives for something else to bedone; 2. request by patient to see a specialist;3. serious impairment of patient's workingcapacity; 4. lack of emotional support forpatient from members of the family; 5. G.P.'sopinion that the patient may find it moreacceptable to be told he has nervous troubleby a specialist, rather than by his own doctor.

It was not possible to make a quantitativeestimate of the influence of each of these factorsseparately upon referral practice. Their divers-ity, however, even among the eight G.P.sstudied was noteworthy. The varying weightaccorded to these non-clinical factors by differ-ent G.P.s could perhaps account, in part, forthe variation in direct referral rates.The variation in rates of referral to psy-

chiatric services shown in the above studyhas implications for epidemiological researchin psychiatry based on specialist-treated cases.Since the G.P. is the principal agent by whompatients are passed to the mental health services,he must exercise a powerful influence on mentalhospital and clinic morbidity statistics. Thehabits of G.P.s in referring cases may well bedetermined, in part, by the nature of thetraining in psychiatry received at the medicalschool. The psychiatric morbidity statisticsfor a large population might be influenced bythe teaching policy in psychiatry in the medicalschool which produces a substantial propor-tion of the doctors for the area.

Attitudes Values and Symptom PatternsThe findings suggest that some doctors take

their cue for referral from the relatives' attitudeor from that of the patient. To this extent,therefore, referral will depend in part uponthe attitudes prevailing in the populations toillness, to doctors in general, and to psychiat-rists in particular. There may well be varia-tions in such attitudes which are related tosex, age, social class, area of residence andother factors.The study of the relationship between the

prevalence of mental disorder measured bydirect survey of a population and the complexweb of social attitudes, values and standardswhich also prevail in the same populationraises difficult issues both theoretical, method-ological and technical.

Detection of the common psychiatric ail-ments-neuroses and personality disorders, forexample-depends upon reports of behaviouralanomalies or of changes in inner experiencewhich will, in turn, be governed by the stand-ards of "normal" behaviour and experience ofpatients themselves, or subscribed to by theirrelatives or by other members of their socialworld. Quite apart from the awareness of theexistence of abnormality, attitudes of diffidencearising from the possibility of stigmatisationmay lead to concealment of such disordereven during special enquiry. Beliefs concerningdepression or morbid anxiety may cause adenial of such phenomena. Potent in this re-gard may be the notion that these manifesta-tions are 'not of medical importance but ratherindicate moral defect or a weak character.The neurotic may be held personally responsiblefor his symptoms which are seen, in the lastanalysis, to be susceptible of voluntary controlin a way which does not apply to manifestly"organic" symptoms.An opportunity to study the influence of the

social climate upon the pattern of psychiatricsymptomatology arose when the entire popula-tion of the South Atlantic island of Tristan daCunha was evacuated to England in 1961following a volcanic eruption (Rawnsley andLoudon, 1964).Although there has usually been fairly

regular contact between Tristan and the restof the world, the community may be regardedas closed in that, for half-a-century before thevolcano erupted, there had been virtually nopermanent migration in or out of the island.The nature and circumstances of life on Tristan-close proximity of residence in a corner ofan inhospitable island mountain; universalinter-relatedness through blood or marriage;

July, 1965404by copyright.

on Decem

ber 24, 2021 by guest. Protected

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.41.477.401 on 1 July 1965. D

ownloaded from

RAWNSLEY: Social Attitudes and Psychiatric Epidemiologyan economic life requiring much co-operationin certain processes-had led to a remarkablehomogeneity in social values and attitudes anda low tolerance for departures from generallyaccepted standards.

In 1937 the population was subject to anepidemic of major hysteria which was ex-tremely well documented by the doctors of aNorwegian expedition which arrived shortlyafter the epidemic began (Christophersen,1946). Twenty-one islanders were affected,11% of the population. It was possible toidentify these cases by name from the detailsprovided in the Norwegian report and 19 ofthem came to England as evacuees, almost25 years later.The hysteria took a variety of forms-faints,

convulsions, "sleeping spells", "fighting spells"and "choking spells". The epidemic graduallysubsided over a period of several months andalthough sporadic cases have occurred since,the condition has never recurred on the samescale. It is difficult, in retrospect, to say whatfactors may have precipitated this outbreak.Strained relationships between families; sexualrivalries and jealousies; the isolated monoton-ous life, are among the causes mentioned bothby the Norwegian investigators and by membersof the community. In the early stages, at least,this series of dramatic exhibitions attracted agreat deal of attention and evoked muchinterest in the population.

In a socio-medical survey conducted in 1962by a social anthropologist (J. B. Loudon) anda psychiatrist (K. Rawnsley), the investigatorswere impressed early in the course of theenquiry by the high frequency of headaches(59% of adult population) and by the remark-ably stereotyped manner in which these weredescribed. They were bifrontal in distribution,the position often being indicated by a cha-racteristic gesture. They were common bothon Tristan and in England and sufferers wereaccustomed to have them every week or two.Sometimes they were disabling, causing thepatient to cease work for a while, but usuallythey were said not to interfere with life act-ivities. They were not associated with eyesymptoms or vomiting and were relieved byaspirin. The commonest provoking factorswere exposure to strong winds or bright sun-shine; menses; and worry.An association was found between the occur-

rence of headache, especially of worry-provokedheadache, and a history (from the Norwegianreport) of previous hysterical attacks. Thus,

of the 19 individuals known to have hadhysteria 25 years previously, 16 now statedthey were subject to headaches associated withworry. Only three members of a control groupof 19, matched for sex and age, had headachesof this kind.

Sixty per cent of the islanders who reportedheadaches denied the influence of anxiety orworry in provoking attacks. In considering thephysiogenic as well as the psychogenic basisof the Tristan headaches one is reminded ofthe experimental work by Holmes, Goodell,Wolf and Wolff (1950) on responses of thenasal mucosa to a variety of stimuli in normalsubjects. Swelling of the turbinates, hyperaemia,increased secretion, obstruction, lowered painthreshold, and sometimes the development ofa rather characteristic headache occurred inresponse to the following: pain elsewhere inthe body; cold; bright lights; menses; allergensand certain emotional states, notably anxiety,resentment, anger, guilt, humiliation, frustra-tion. To recapitulate: the precipitating causesof headache reported by the islanders wereworry; bright sunshine; strong winds; andmenses.Two hypotheses may be advanced but can-

not be resolved on the evidence provided bythis survey:-1. A high proportion of the population is

liable to nasal congestion with consequentheadache i'n response to a number of provok-ing agents. Those of neurotic disposition mayfind that their undue load of anxiety or othermorbid affect is especially potent in producingthe response.

2. There is a nucleus of people, perhaps quitea small one, with headache due to nasalcongestion. In addition, however, there areislanders who, without a physiogenic mechanismof this kind, 'have adopted the headache res-ponse to anxiety as a convenient, sociallyacceptable, commonplace symbolic reaction.The homogeneous nature of Tristan society,together with the high degree of social inter-action, may have powerfully influenced theestablishment and spread of this symptom. Ifone accepts the Norwegian figures for preval-ence of headache in 1937, it must be concludedthat this malady has become much more ex-tensive in recent years. Headache may nowhave become, in part, an endemic neuroticsymptom modelled on a physiogenic disorder,but spreading through the community in a lessdramatic though more enduring fashion thanthe convulsive hysteria of 1937.

July, 1965 405

by copyright. on D

ecember 24, 2021 by guest. P

rotectedhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.41.477.401 on 1 July 1965. Dow

nloaded from

POSTGRADUATE MEDICAL JOURNAL

The association between certain social at-titudes and the prevalence of symptoms hasbeen studied in the much more complex, muchless homogeneous society of a rural area inS. Wales. The work of a social anthropologistin this population (Loudon 1961, 1964), to-gether with the private sociological census madein collaboration with Professor Cochraine's Unitlaid -the basis for a division into populationsections which differed socially in many respects.A comparative survey has recently been

made of the prevalence of a large number ofsymptoms in random samples drawn fromeach of the social sections of the rural popula-tion. Special techniques were devised for appli-cation in home interviews for the assessmentof symptoms and of associated attitudes. Theattitudes measured were (a) level of sympathymanifested towards certain symptoms; (b) ex-tent to which the same symptoms are regardedas proper objects of medical care.Symptoms were assessed using techniques

designed to minimise the observer's activeparticipation with the consequent distortiondue to prejudice and bias in the observer. Forcomparative purposes, main reliance was placedon two "objective" procedures-a modifica-tion of the Cornell Medical Index HealthQuestionnaire which, in its original form com-prises some 200 questions (to be answered'Yes' or 'No') about physical and psychologicalsymptoms; and specially designed scales for alimited number of symptoms (Ingham, 1965).The scales had an advantage over conventionalquestionnaire forms in providing a method ofgrading the severity of the symptom. They also,by their design and method of presentation,served to reduce the influence of certainspurious response "sets", e.g., a tendency toanswer "Yes" to questions whatever the con-tent, which might otherwise yield spuriousvariations between social categories.Other measures of morbidity were also em-

ployed, including re-interview of a subsampleby a psychiatrist who was ignorant of theperformance on the first interview; specialobservation by G.P.s for a period of threemonths; records of attendance at psychiatrichospitals or clinics in recent years.The results of this investigation have yet

to be published, and no attempt will be madeto present them in this paper. The survey ismentioned principally in order to draw attentionto the outline strategy and design of a particu-lar research project aimed at uncovering asso-

ciations between the complex web of socialvalues and attitudes (viewed as elements in thesocial structure and organisation of a popula-tion) and the pattern of psychiatricsymptomatology.

It was predicted that there would be positiveassociations between the social attitudes (ofsympathy, and of readiness to seek a doctor'sadvice) in a section of the population towardsparticular symptoms and the prevalence ofthese symptoms in the population section. Thesocial anthropologist, deploying his skillsinitially at the "micro" level of observation ofsocial relations, behaviour and values, was ableto proceed at the "macro" level to the construc-tion of an ad hoc social classification of theentire rural population yielding divisionspeculiarly relevant to the purpose of the enquiry.The psychologists devised techniques for thequantitative estimate of both attitudes andsymptoms with built-in safeguards to offsetthe prejudices of the interviewers and to reducethe influence of response "sets". The psy-chiatrist, using more conventional clinical skills,was able to provide data for comparison withthe psychological measures.

Sir Aubrey Lewis (1961) has bemoaned thedependence on symptom counts in psychiatricsurveys rather than on diagnoses. He regardssuch a state of affairs as "humiliating becauseit throws us back to the infancy of medicine;it aligns us with the school of Cnidus, so sharplyrebuked by the rival Coans twenty oddcenturies ago for differentiating a host of typesof disease solely according to the subjectivesymptoms complained of, without regard towhat examination might reveal or a commoncause underlie." Nevertheless he concludes, "Itmust be sadly admitted that we cannot escapeit in our present state."

Progress in the epidemiology of mental dis-orders hinges in the first place upon thesharpening of methods for the reliable andvalid estimation of psychiatric morbidity.Attempts to design such methods quicklyevoke fundamental questions concerning thenature and definition of psychiatric disorder.The answers involve, inter alia, considerationof social values and attitudes pertaining tohuman behaviour. In addition to influencingthe recognition and disposal of the psychiatriccase by the G.P. and affecting the fate of thepatient in hospital, social attitudes may makea powerful contribution to determining theoccurrence and content of psychopathology.

406 July, 1965by copyright.

on Decem

ber 24, 2021 by guest. Protected

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.41.477.401 on 1 July 1965. D

ownloaded from

July, 1965 RAWNSLEY: Social Attitudes and Psychiatric Epidemiology 407

Grateful acknowledgment is made for the facilitiesand help provided by Professor A. L. Cochrane, Hon.Director of the M.R.C. Epidemiological ResearchUnit, and by Dr. J. C. Gilson, Director of theM.R.C. Pneumoconiosis Research Unit. The workin South Wales owes a great deal to the inspirationand support provided constantly by Sir Aubrey Lewis.

REFERENCESBROOKE, E. M. (1959): A Longitudinal Study of

Patients First Admitted to Mental Hospitals, Proc.roy. Soc. Med., 52, 280.

BROWN, G. W., CARSTAIRS, G. M., and TOPPING, G.(1958): Post-hospital Adjustment of Chronic MentalPatients, Lancet, ii, 685.

CHRISTOPHERSEN, E. (ED.) (1946): Results of theNorwegian Scientific Expedition to Tristan da Cun-ha 1937-38, Oslo: Det Norske Videnskaps-Akademi.

FOLKARD, S. (1960): Comparative Study of Attitudesto the Rehabilitation of Psychiatric Patients, Brit.J. prev. soc. Med., 14, 23.

HOLMES, T. H., GOODELL, H., WOLF, S., and WOLFF,H. G. (1950): The Nose, Springfield, Illinois:Charles C. Thomas.

INGHAM, J. G. (1965): A Method for ObservingSymptoms and Attitudes, Brit. J. Soc. Clin. Psychol.4, 131.

KESSEL, W. I. N. (1960): Psychiatric Morbidity ina London General Practice, Brit. J. prev. soc.Med., 14, 16.

KESSEL, W. I. N., and SHEPHERD, M. (1962): Neurosisin Hospital and General Practice, J. ment. Sci., 108,159.

LEWIS, A. J. (1961): "Mental Disorders in Britain"-in "Comparative Epidemiology of the MentalDisorders" ed. Hoch, P. H., & Zubin, J. New Yorkand London: Grune and Stratton.

LOGAN, W. P. D., and CUSHION, A. A. (1958):"Morbidity Statistics from General Practice", Vol. I(General): General Register Office: Studies onMedical and Population Subjects No. 14., London:H.M.S.O.

LOUDON, J. B. (1961): Kinship and Crisis in S. Wales,Brit. J. Sociol. XII. No. 4, 333.

LOUDON, J. B. (1965): Religious Order and MentalDisorder: a Study in a Rival Community inS. Wales in "The Social Anthropology of ComplexSocieties", A.S.A. Monograph 4. London:Tavistock.

MOWBRAY, R. M., BLAIR, W., JUBB, L. G., and CLARKE,A. (1961): The General Practitioner's Attitude toPsychiatry, Scot. med. J., 6, 314.

NORRIS, V. (1956): A Statistical Study of the Influ-ence of Marriage on the Hospital Care of theMentally Sick, J. ment., Sci., 102, 467.

RAWNSLEY, K. and LOUDON, J. B. (1962a): FactorsInfluencing the Referral of Patients to Psychiatristsby General Practitioners, Brit. J. prev. soc. Med.,16, 174.

RAWNSLEY, K., and LOUDON, J.B. (1962b): TheAttitudes of General Practitioners to Psychiatry.Sociological Review Monograph No. 5, Universityof Keele.

RAWNSLEY, K., and LOUDON, J. B. (1964): Epidemio-logy of Mental Disorder in a Closed Community,Brit. J. Psychiat., 110, 830.

RAWNSLEY, K., LOUDON, J. B., and MILES, H.L.(1962): At'itudes of Relatives to Patients in MentalHospitals, Brit. J. prev. soc. Med., 16, 1.

REGISTRAR GENERAL (1964): Statistical Review ofEngland and Wales for the year 1960. Supplementon Mental Health. London: H.M.S.O.

WING, J. K., BENNETT, D. H., and DENHAM, J. (1964):The Industrial Rehabilitation of Long-stay Schizo-phrenic Patients. Medical Research Council Memo-randum No. 42. London: H.M.S.O.

WING, J. K., DENHAM, J., and MONRO, A. B. (1959):Duration of Stay in Hospital of Patients Sufferingfrom Schizophrenia, Brit. J. prev. soc. Med., 13, 145.

WING, J. K., MONCK, E., BROWN, G. W., andCARSTAIRS, G. M. (1964): Morbidity in the Com-munity of Schizophrenic Patients Discharged fromLondon Mental Hospitals in 1959, Brit. J. Psychiat.,110, 10.

by copyright. on D

ecember 24, 2021 by guest. P

rotectedhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.41.477.401 on 1 July 1965. Dow

nloaded from