11
THE UNITED STATES-UNITED KINGDOM PROJECT ON DIAGNOSIS OF THE MENTAL DISORDERS Joseph Zubin and Barry J. Gurland Biometrics Research Unit New York State Psychiatric Institute New York, New York 10032 and Columbia University New York, New York INTRODUCTION No one can really understand his own culture, the anthropologists tell us, until he develops an understanding of a contrasting culture. This holds true especially of the culture of psychopathology. As long as American diagnostic procedures and treatment remained isolated and self-enclosed it was difficult for us to conceive of any other system. The United Kingdom was chosen as the contrasting culture because, although the total admission rates for mental disorder in the two countries were nearly identical, the relative admission rates for the two functional psychoses - schizo- phrenia and affective disorders - were strikingly different. There were relatively only half as many admissions for schizophrenia as for the affective psychoses in the UK but twice as many in the USA, so that relatively speaking there was a fourfold difference in the ratio of first admissions for the two functional psychoses in the two countries. Similarly, the ratio of first admissions for functional psychoses to chronic organic brain syndromes was approximately ten in the US and approximately one in the UK, or a tenfold difference. Whether these discrepancies were due to cultural, biological or classification differences, or to differences in hospital utilization in the two countries was the focus of the inquiry of the USA-UK Project. The choice of the UK was also based on the fact that despite differences in culture, the English language was common to both countries and a deep and abiding interest in psychopathology and its classification permeated the teaching and practice of psychopathology in both. The first need that had to be met was the provision of a common system of interviewing that could be used in both countries. Fortunately, in the USA as well as in the UK considerable work was already in progress in transforming the free-wheeling psychiatric interviews into systematic structured instruments that could be used in a standard fashion. A combination of the items from the Present State Examination' prepared in London by the Social Psychiatry Research Unit under Dr. John Wing and his associates and from the Psychiatric Status Schedule12 prepared by our own Biornetrics Research Unit under the guidance of Dr. Spitzer and his associates was hammered out into a joint instrument, and additional forms for obtaining history information and social data were developed. The project staffs in the two countries were trained in the use of the prepared interview schedules until a high degree of reliability was obtained in the precoded ratings of psychopathology based on the responses of the patients to 676

THE UNITED STATES-UNITED KINGDOM PROJECT ON DIAGNOSIS OF THE MENTAL DISORDERS

Embed Size (px)

Citation preview

THE UNITED STATES-UNITED KINGDOM PROJECT ON DIAGNOSIS OF THE MENTAL DISORDERS

Joseph Zubin and Barry J. Gurland

Biometrics Research Unit New York State Psychiatric Institute

New York, New York 10032 and Columbia University

New York, New York

INTRODUCTION

No one can really understand his own culture, the anthropologists tell us, until he develops a n understanding of a contrasting culture. This holds true especially of the culture of psychopathology. As long as American diagnostic procedures and treatment remained isolated and self-enclosed it was difficult for us t o conceive of any other system.

The United Kingdom was chosen as the contrasting culture because, although the total admission rates for mental disorder in the two countries were nearly identical, the relative admission rates for the two functional psychoses - schizo- phrenia and affective disorders - were strikingly different. There were relatively only half as many admissions for schizophrenia as for the affective psychoses in the UK but twice as many in the USA, so that relatively speaking there was a fourfold difference in the ratio of first admissions for the two functional psychoses in the two countries.

Similarly, the ratio of first admissions for functional psychoses to chronic organic brain syndromes was approximately ten in the US and approximately one in the UK, or a tenfold difference. Whether these discrepancies were due to cultural, biological or classification differences, or t o differences in hospital utilization in the two countries was the focus of the inquiry of the USA-UK Project.

The choice of the UK was also based o n the fact that despite differences in culture, the English language was common to both countries and a deep and abiding interest in psychopathology and its classification permeated the teaching and practice of psychopathology in both.

The first need that had to be met was the provision of a common system of interviewing that could be used in both countries. Fortunately, in the USA as well as in the UK considerable work was already in progress in transforming the free-wheeling psychiatric interviews into systematic structured instruments that could be used in a standard fashion. A combination of the items from the Present State Examination' prepared in London by the Social Psychiatry Research Unit under Dr. John Wing and his associates and from the Psychiatric Status Schedule12 prepared by our own Biornetrics Research Unit under the guidance of Dr. Spitzer and his associates was hammered out into a joint instrument, and additional forms for obtaining history information and social data were developed.

The project staffs in the two countries were trained in the use of the prepared interview schedules until a high degree of reliability was obtained in the precoded ratings of psychopathology based on the responses of the patients t o

676

Zubin & Gurland: Diagnosis of Mental Disorders 677

the specific items of the interview schedules. The eighth edition of the ICD (International Classification of Diseases) and the accompanying Glossary of Mental Disorders compiled in the UK by the General Register Office, which is quite similar t o the American Psychiatric Association's Diagnostic and Statistical Manual, were utilized t o yield operational definitions and categories.

There were three types of studies undertaken: ( 1 ) examination of admissions t o mental hospitals, (2) showing of videotapes of the same patients to audiences of clinicians in many parts of the UK and USA and Canada and (3) analysis of case records of earlier cases for historical analysis of trends.

HOSPITAL STUDIES

In order t o compare the data on admissions in the two countries, the individual items in the interview schedule were clustered according t o clinical judgment on the standard dimensions of psychopathology - viz., anxiety, depression, depersonalization, etc. These item clusters were then subjected t o a factor analysis in order to attain greater internal consistency for each dimension and better differentiation between the dimensions, and the resulting factors standardized with a mean of 50 and a standard deviation of 10. Profiles of the standard scores for each patient were then drawn u p and the average profiles for the patients in the two contrasting categories of schizophrenia and affective dis- orders based o n project diagnosis are shown in F I G U R E 1 and F I G U R E 2. It will

64 - SCHIZOPHRENIC (N=43) 62 ~ - - AFFECTIVE (N=53)

60 I

44= 42

U

FIGURE 1. Mean section score profiles for Brooklyn patients diagnosed by the project team as schizophrenic or affective.

678 Annals New York Academy of Sciences

60 w 58- a s 56- v) 54.

52. a $ 50- 5 48-

46.

42L

- SCHIZOPHRENIC (N=44) --- AFFECTIVE (N=66)

FIGURE 2. Mean section score profiles for Netherne patients diagnosed by the hospital staff as schizophrenic or affective.

be noted that the dimensions reflecting mood disorder, lying to the left, show the schizophrenics with low scores and the affectives with high scores, whereas the dimensions reflecting behavioral and conceptual disorganization show the schizophrenics with a high profile and the affectives with a low profile.

The initial studies covered the age range 20-59 in two large public mental hospitals in New York and London. The local diagnoses made by the hospital staff lived up to expectations based on the national statistics revealing the great discrepancies between schizophrenic and affective disorders in the two countries. The project diagnoses told quite a different story. The differences practically disappeared though a small residue of differences remained. In order to get a more representative comparison, all the admissions to public mental hospitals in metropolitan New York and London were sampled and the results are shown in

It is clear that the differences between the two cities were inherent not in the patients but in the psychiatric culture.

Since the differences between the diagnoses in the two cities seemed to be based upon the attitudes of the psychiatrists towards the utilization of the two diagnostic categories rather than upon differences in the patients, an appeal was made to signal detection theory to determine whether the differences inhered in the criterion for judgment of the presence or absence of the disorders or in the sensitivity to the presence of the two disorders.

F I G U R E 3.

Zubin & Gurland: Diagnosis of Mental Disorders 679

NEW YORK LONDON

@ @ . Project Dlagnosis

FIGURE 3. The ratio of schizophrenia and affective disorders among admissions to public mental hospitals, as judged by hospital and project diagnosis.

There was little or no difference in the sensitivity of the two groups of psychiatrists to the presence of psychopathology of the affective disorders or of the schizophrenic disorders, but -there w&e significant differences in the criterion. The American psychiatrists were reluctant t o use the label of affective disorders but were free in their use of the schizophrenic label whereas the reverse held true of the British psychiatrists.*

VIDEO TAPE STUDIES

In order to replicate the finding that the differences inhered in the psychiatric culture rather than in the patients, the same video tapes of patient interviews were shown to groups of psychiatrists in the various parts of the USA, UK and Canada. The results of this study are shown h F I G U R E 4.

The predilection for the diagnosis of schizophrenia on the part of the American psychiatrists and the corresponding preference for the diagnosis of affective disorders on the part of the British is quite clearly borne out.

Those patients who inspired transatlantic diagnostic disagreement were generally diagnosed as schizophrenic by the American psychiatrists; they received a wide range of British diagnoses, however. Diagnoses of the manic-depressive disorders were common, but also neuroses and personality disorders were diagnosed. There was no group of patients whom the British tended to call schizophrenic more often than did the Americans.

* This analysis is reported by Dr. W. Crawford Clark in an unpublished study. Although the two circles for London representing the hospital and project diagnoses seem quite similar, there was a considerable change between these two diagnoses for many patients, but the proportions remained quite similar.

680 Annals New York Academy of Sciences

NEW YORK LONDON NEW YORK LONDON

FIGURE 4. Eight videotaped interviews as diagnosed by psychiatrists at The New York State Psychiatric Institute and at The Maudsley Hospital, London.

Some biases? toward the diagnosis of schizophrenia were about equally powerful in the two countries, and therefore did not alter the level of transatlantic diagnostic disagreement. For example, in both countries younger patients were more likely to be called schizophrenic than were older patients with similar psychopathology. Surprisingly there was no evidence of bias associated with the social class of the patient in either country; the significance of this result is questionable, however, since there was only a very narrow spread of social classes within the patient population under examination.

In both countries, there was a strong association between race and the hospital diagnosis of schizophrenia. ' In London, this association was con-

? Here and elsewhere, a reference to diagnostic bias means that there was an association between diagnosis and a demographic variable (e.g., age, race, sex) even when psycho- pathology was held constant.6

Zubin & Gurland: Diagnosis of Mental Disorders 68 1

founded by the fact that all of the black patients were also recent immigrants, but the symptomatology of this group was consistent with the hospital diagnosis. In New York, it became clear that there was a bias on the part of the hospital psychiatrists toward diagnosing schizophrenia in black patients. Project psychiatrists in New York, on the other hand, showed no such bias and found no increased frequency of schizophrenia in the black patients; nor were the symptoms shown by the patients diagnosed schizophrenic by the project team any different in the black than in the white patients.

FOLLOW-UP STUDY

A two-year documentary unpublished follow-up of duration of hospital stay showed about the same predictive power for project and hospital diagnoses in London, but the hospital diagnoses in New York more sharply discriminated the prognosis of affective and schizophrenic groups than did the project diagnoses (although both the hospital and project diagnoses showed a worse prognosis for the schizophrenic than for the affective groups of patients). However, whereas the project diagnosis was made without foreknowledge of the patient’s progress, the hospital diagnosis may have reflected a n awareness of the patient’s response to treatment. It is possible, for example, that in those few cases that the New York hospital diagnosed as affective disorder, the diagnosis was predicated o n the patient’s rapid recovery. There are additional reasons for a cautious interpretation of follow-up data based on routine hospital records. Apart from the usual errors in documentary information and the omission of data on the progress of symptoms, the duration of hospital stay may be altered by administrative events such as a strike of staff members, availability of beds, o r fiduciary considerations, and rehospitalization rates may depend more on the family structure than the patient’s symptoms, as shown in another study undertaken by project members. These follow-up data, consequently, may be of limited value as a test of the predictive validity of the diagnosis.

Insofar as the operational definition of schizophrenia in the ICD of DSM I1 is based o n descriptive psychopathology, there appears t o be greater internal consistency in the British than in the American concepts of the disorder. The British hospital diagnoses show a stronger relationship to psychopathology (independently measured by the project) than the American hospital diag- n o s e ~ . ~ 11 0 $

CASE RECORD STUDY

Whatever the relative utility of the narrow and the broad concepts of schizophrenia, held respectively by psychiatrists in the UK and USA, it is

$ For more incisive testing of the usefulness of the different concepts of schizophrenia, one must turn to studies in the field of genetics by other workers,’ to studies on the relationship between diagnosis and treatment response, or between diagnosis and physio- logical variables. Recent studies by the Cross-National Project, in collaboration with the Biometrics Laboratories under the direction of Dr. Samuel Sutton, have used an “iterative technique”’ to reduce inconsistencies between neurophysiological and clinical classifica- tions. In this way, data have been collected on the concurrent validity of the two modes of classification.

682 Annals New York Academy of Sciences

Year 25 30 35 40 45 50 55 60 65 L

F n t sample Second sample

FIGURE 5 . Total admissions, including first and readmissions. Profiles of proportion schizophrenic for first admissions and readmissions separately are only available for 1934-1957, and are nearly identical.

interesting to explore the development of these concepts in the two countries. For this purpose, a case record was carried out at two institutes of psychiatry (New York State Psychiatric Institute and the corresponding Institute of Psychiatry at Maudsley Hospital). These two institutes were chosen because they have wielded considerable influence in shaping the concepts of schizophrenia. The proportion of admissions annually diagnosed by the hospital psychiatrists as schizophrenic between 1930 and 1970 was found to be fairly constant in London, but to have increased dramatically in New York across the years 1942-1952. A sample of the New York case records was drawn from the years 1932-1 941 and 1947- 1956, censored to remove references to diagnosis and year of admission, and then presented to 16 American-trained psychiatrists and one British-trained project psychiatrist for diagnosis. The rediagnoses showed no change in the proportion of schizophrenics admitted annually across the decades (see F I G U R E 5). The conclusion was that the concepts of schizophrenia had been much the same in London as in New York in the decade 1930-1940, but that the concept of schizophrenia held by New York psychiatrists had greatly broadened after the second World War.§

The psychiatric traditions emanating from Pinel in France, Kraepelin in Germany, and Freud in Austria-Hungary, which so largely determined the trends in psychiatry today, underwent a confluence in both the United States and the United Kingdom, even though each tradition was weighted differently. Furthermore, Adolf Meyer, who was such an important influence in the United States, also had disciples in the United Kingdom, especially in Scotland, so that the two countries were exposed to similar influences that had differential impacts. By comparing the United States and the United Kingdom in the 1970s, we are examining the influence of the same traditions and the way they were incorporated in the two countries. A discussion of some of the factors that led to a broadening of the concept of schizophrenia in the United States in the 1940s, and later, has been broached el~ewhere.~ Certainly, the growth of psychoanalytic therapy in the United States brought with it a heightened interest in the detection of milder or borderline forms of schizophrenia, and possibly an increased tendency to infer the presence of this disorder from subtle or intuitive signs.

Zubin & Gurland: Diagnosis of Mental Disorders

DISCUSSION

683

In the absence of clear evidence in favor of one or the other concept of schizophrenia, one might ask which concept, if wrong, is likely to do the least harm to the patient. If wrong, the broad concept of schizophrenia would include patients suffering from depressive and manic disorders, and there would be a theoretical danger that the patient's depression or mania would not receive specific treatment and that the relatively poor prognosis expected in schizo- phrenia would lead to a delay in discharging the patient from the hospital. An analysis of the data from the hospital studies showed that patients with mainly depressive symptoms were less likely to receive antidepressive treatment in the New York than in the London hospitals and that this was at least partly due to the fact that many of the New York patients received a hospital diagnosis of s~hizophrenia.~ The New York-London contrast in therapy, however, might have been more serious had the New York psychiatrists not paid more attention to symptoms rather than diagnosis in selecting treatment (i.e., a large proportion of New York patients with depressive symptoms received antidepressive treatment despite a hospital diagnosis of schizophrenia). Furthermore, there was no evidence of a self-fulfilling prophecy, since the patients with mainly depressive symptoms had a relatively short duration of hospitalization, whether or not they were called schizophrenic by the hospital psychiatrists. Nevertheless, it seems that the narrow concept of schizophrenia has an edge over the broad concept until further evidence accrues.

The major thrust of this paper has been in the direction of demonstrating the reliability of psychopathological ratings and of diagnoses. The only validity aspects dealt with was the demonstration that schizophrenics tended to require longer hospitalization than affectives. But there is more to validity than outcome. If we appeal to the six scientific models of etiology: genetic, internal environment and neurophysiology on the one hand and ecological, develop- mental and learning, on the other hand,' we can demonstrate that depending upon the model one adopts, certain diagnostic approaches seem to be more valid than others. Thus, accepting the genetic model, Essen-Moller's more middle-of- the-road concept of schizophrenia, neither as wide as the American nor as narrow as that of the British, shows the highest rate of genetic transmission of the disorder in relative^.^ Accepting the neurophysiological model, it can be demonstrated that with processing of information as our touchstone, certain groups of schizophrenics, e.g., those with thought disorder, differ in their central nervous system functioning from other schizophrenics, depressives and normals insofar as they have a shorter critical duration for sensory integration. Similarly, manic-depressives tend to differ from schizophrenics and normals in their auditory threshold as well as in their capacity to benefit from an additional energy input that follows 15 msec after the initial input, than do the other groups. There are several other findings of this type indicating that the information processing is different in certain types of psychopathology, and in all of these findings, the patient group in question excels the normals, thus eliminating the explanation that patients do worse because they are unmoti- vated.' These laboratory findings can provide indicators that would fortify the diagnostic category of the patients who exhibit the indicators.

Similar findings could be based on the application of the learning theory model, which would indicate that those who are more responsive to immediate rather than nonimmediate stimulation by the environment are more likely to be schizophrenic.8 Similar indications can be found for the developmental model,

684 Annals New York Academy of Sciences

the internal environment model, and the ecological model, but these would take us too far afield from our topic. Perhaps the best summary one can make of the approach suggested here is t o indicate the need for culture-free and culture-fair indicators of psychopathology. We must distinguish between two cultures: the culture of the patient and the psychiatric culture of the psychiatrist. Armed with such transcultural measures, it would become possible t o make comparisons cross-culturally despite the occluding influences of disparate patient cultures. The neurophysiological indicators mentioned above might well serve this purpose.

The work reported here identifies a t least one of the factors that confound cross-cultural comparisons of psychopathology, namely, the effect on perception and labeling of psychopathology of variation in psychiatric culture. The methods used by the USA-UK project staff were designed t o control the effects of variation in psychiatric culture by such means as structured interview tech- niques, a glossary of psychiatric labels, and classification based on statistical analysis of discrete ratings. We have presented evidence of the success of these methods in reducing spurious findings of cross-national differences in the distribution of types of mental disorder.

Unfortunately, methods of assessing and defining mental disorders that are based on phenomenology (as are the methods reported here) are vulnerable t o cultural biases other than those operating through the psychiatrist. Most importantly, the patient’s culture may influence the way in which a given mental disorder is manifested. It is taken for granted that the content of abnormal behavior is culture-bound and little is yet known of the extent, if any, to which the form of psychopathology is also culture-bound. Thus, the degree t o which a given culture elicits, contains, suppresses or occludes manifestations of psycho- pathology is still unknown. I t would clearly be of value to cross-cultural research to have available indicators (of mental disorders) that are not affected by the culture of the milieu nor of the psychiatrists. Tests of neurophysiological function offer promise in this respect.

The Venn diagram ( F I G U R E 6) can be used to illustrate some conceptual prob- lems in developing objective indicators of mental disorders for cross-cultural re- search.Three subpopulations of patients with mental disorders are shown: (1)those who actually have mental disorder X; ( 2 ) those so diagnosed by psychiatrist A and (3) by psychiatrist B. It will be noted that both psychiatrists include in their diagnoses individuals who d o not belong to category X of mental disorder. Both the diagnoses of psychiatrist A or of psychiatrist B can be objectified by use of reliable methods for assessing phenomenology or by finding, for example, physiological correlates of one or other set of diagnoses. Yet these objective methods might indicate subpopulations that overlapped but did not coincide with the actual cases of mental disorder X. Thus it is necessary t o be clear about what we wish t o objectify and what we wish t o indicate.

Ideally, an objective indicator of mental disorder should be positive only with actual cases of a given mental disorder, with all such cases and with no other cases. It is more likely, however, that objective indicators available now or in the near future will be positive with a substantial but unknown proportion of actual cases of a given mental disorder and with some cases of other mental disorder. It might therefore seem reasonable in the interim t o attempt a study of the pathoplastic effects of culture by the following expedient: (1) select two o r more defined populations, each representing contrasting cultures, and screen them for cases of putative mental disorder X by use of a method that is free of variation due t o psychiatric culture; and ( 2 ) test the two resulting subsamples

Zubin & Gurland: Diagnosis of Mental Disorders 68 5

Patients with Mental Disorders

0 I - Patients with Mental Disorder X

2- Patients with Dmgnosis X by Psychiatrist A

0 3- Patients with Diagnosis X by Psychiatrist B

FIGURE 6.

with the most relevant of the available neurophysiological indicators. If the proportions of each subsample found to be positive on the neurophysiological indicator are essentially equal then it is likely that the patients’ culture has only a weak effect, if any, on the form of mental disorder X. In that event we will be in a position to proceed to examine the effects of the patients’ culture on the incidence, and course of that mental disorder. For this purpose, and under the circumstances outlined above, indicators of mental disorder X could be chosen either from objectified phenomenological assessments (keeping psychiatric culture constant) or from the neurophysiological or other biological fields.

REFERENCES

1 .

2.

BARRETT, J . E., Jr., J . B. KURIANSKY & B. J. GURLAND. 1972. Community tenure following emergency discharge. Am. J. Psychiat. 128: 958-964.

BRUDER, G., S. SUTTON, H. BABKOFF, B. J . GURLAND, A. YOZAWITZ & J. L. FLEISS. 1975. Auditory signal detectability and facilitation of simple reaction time in psychiatric patients and non-patients. Psychiat. Med. 5: 260-272.

FLEISS, J . L., B. J . GURLAND, R. J . SIMON & L. SHARPE. 1973. Cross-national study of diagnosis of the mental disorders: Some demographic correlates of hospital diagnosis in New York and London. Int. J . SOC. Psychiat. 19: 180-186.

GURLAND, B. J., J. L. FLEISS, J. E. COOPER, L. SHARPE, R. E. KENDELL & P. ROBERTS. 1970. Cross-national study of the diagnosis of mental disorders: Hospital diagnoses and hospital patients in New York and London. Compr. Psychiat. 1 1 : 18-25.

3.

4.

686 Annals New York Academy of Sciences

5.

6.

7 .

8.

9.

10.

11.

12.

13.

14.

15.

GURLAND, B. J. , J. L. FLEISS, J . E. BARRETT, Jr., L. SHARPE & R. J. SIMON (U.S.)with COPELAND, J . R. M., J . E. COOPER & R. E. KENDELL (U.K.). 1972. The mislabeling of depressed patients in New York State hospitals. In Disorders of Mood. J. Zubin & F. Freyhan, Eds.: 17-31. Johns Hopkins Press. Baltimore, Md.

PROFESSIONAL STAFF O F THE US-U.K. CROSSNATIONAL PROJECT. 1974. The diagnosis and psychopathology of schizophrenia in New York and London. Schiz. Bull. No. 11 (Winter): 80-102.

ROSENTHAL, D. & S. KETY. 1968. The Transmission of Schizophrenia. Pergamon Press. New York, N.Y.

SALZINGER, K. 1973. Schizophrenia: Behavioral Aspects. John Wiley. New York, N.Y.

SHIELDS, J. & I. GOTTESMAN. 1972. Cross-national diagnosis of schizophrenia in twins. The heritability and specificity of schizophrenia. Arch. Gen. Psychiat. 27: 725-730.

SIMON, R. J., B. FISHER, J. L. FLEISS, B. J . GURLAND & L. SHARPE. 1971. Relationship between psychopathology and British or American-oriented diagnosis. J. Abn. Psychol. 78: 26-29.

Depression and schizophrenia in hospitalized black and white mental patients. Arch. Gen. Psychiat. 28: 509-512.

SPITZER, R. L., J. ENDICOTT, J. L. FLEISS & J . COHEN. 1970. The Psychiatric Status Schedule: A technique for evaluating psychopathology and impairment in role functioning. Arch. Gen. Psychiat. 23: 41-55.

SUTTON, S. 1973. Fact and artifact in the psychology of schizophrenia. In Psychopathology: Contributions from the Social, Behavioral and Biological Sciences. M. Hammer, K. Salzinger & S. Sutton, Eds.: 197-213. Wiley. New York, N.Y.

WING, J., J. E. COOPER & N. SARTORIUS. 1974. The Measurement and Classification of Psychiatric Symptoms. Cambridge Univ. Press. New York, N.Y.

ZUBIN, J . 1972. Scientific models for psychopathology in the 1970’s. Seminars in Psychiatry. 4: 283-296.

SIMON, R. J., I. L. FLEISS, B. J . GURLAND, P. STILLER & L. SHARPE. 1973.