J.H.K.C. Psych. (1991) 1, 37-46
Peter W.H. Lee, F. Lieh-Mak, K.K. Yu & J.A. Spinks

pdf Full Paper in PDF


The recent literature on schizophrenia generally points to a better outcome and a more benign metamorphosis in the schizophrenic illness process. We report a study on the cross-sectional outcome of 153 actively follow-up schizophrenic patients in Hong Kong. Apart from tgeir more favourable symptomatic outcome, other outcome indicators relating to social relationship, quality of life, and work adjustment can hardly be labelled as benign. Result obtained from the one year follow-up on the same group of subjects indicated that symptomatic control was less consistent and liable to fluctuate over time. Other psychosocial measures of outcome were, however, highly consistent and stable over time. This indicated that psychosocial handicaps associated with schizophrenia may be more resistant to our currently emphasized forms of treatments. The rehabilitation cost of sustaining schizophrenic patients in society is discussed in relation to the present outcome pattern, and implications for future directions in the management of schizophrenia in a more cost effective manner is proposed.


Schizophrenia in the 21st century promises an increasing optimism than ever before known. Indeed, a benign metamorphorsis (Zubin et al., 1983) in the pathological process was postulated. Lehmann (1981) estimated that the chances of a favourable schizophrenic outcome are four to five times better than they were in the early years of the century. Kraepelin originally regarded inevitable deterioration into a typical "end-state" as the defining criterion for dementia praecox. He subsequently revised his estimate to a lower 70% for ultimate deterioration. Based on their earlier survey of 800 outcome studies and also reviewing the results of other follow-up studies conducted in Europe, Zubin et al, (1983) asserted that "..... the outcome of schizophrenia appears to be changing from chronicity to an episodic course with a more favourable outlook." Wing (1987) also cautioned that the pessimistic schizophrenic outcome may be complicated by the poverty of the social environment.

Manfred Bleuler (1978) categorised 6%-15% of his long term follow-up probands as persistently chronic. He argued that schizophrenia was not a generally deteriorating condition, particularly after the first five years. His data indicated that about 25% of schizophrenic patients would recover completely with no further need for treatment. An intermediate group comprising about half of the patients would run a fluctuating course for years depending on environmental and treatment factors. Of the remaining 25%, only 10% would result in permanent hospitalisation or become invalidated for life. Bleuler also noted a general improvement in the schizophrenic course due to improvement in social and environmental conditions predating the phenothiazine era. In line with this general expectation, other studies provided a similarly positive outlook for schizophrenic outcome (Ciompi, 1980; W.H.O., 1979; Lo and Lo, 1977). Strauss et al. (1978) compared the two year outcome of a group of schizophrenic with nonschizophrenic psychiatric patients, and noted that "although there was a trend toward the schizophrenics being worse on all outcome measures, there were no significant differences between the two groups on measures of social relationships, work functioning, and symptom severity at follow-up". Studies in Berlin by Pietzcker and Gaebel (1987) supported some of Strauss et al's (1978) findings that by the end of their first year of follow-up, schizophrenic patients did not differ significantly in a number of clinical outcome measurements from their control patients with neurotic and affective disorders.

Despite the gradually unfolding consensus pointing towards a more benign outcome in schizophrenia, the concept of outcome employed in previous studies had mostly been unclear. This problem was confounded by the use of nonspecitic and non-operationally defined criteria. Instead of specifying outcome dimensions in different aspects of the patient's life functions, global descriptive terms like "improved", "unchanged" or "unimproved" were used. Falloon (1983) noted that "the increasingly rigorous attention paid to the diagnosis of schizophrenia has not yet been extended to the development of reliable and valid criteria for describing the course of the illness." Indeed recent findings indicated that outcome was unlikely to be classifiable under global and vague descriptive categories. Ciompi (1980) noted that "there is no such thing as a specific course of schizophrenia". Strauss and Carpenter's group (1972, 1974a, 1974b) argued that outcome in schizophrenia was characterised by a highly variable and heterogeneous process. Bleuler (1974) noting the wide variety of outcomes in schizophrenic patients was convinced that "a specific treatment of schizophrenia does not exist". Falloon (1983) cautioned that "successful management of an illness should effectively minimize all symptoms of the condition and any associated disability and handicap". Studies in schizophrenic outcome seemed to be reminiscent of Rachman's (1974) concept of "desynchrony" between different outcome parameters in anxiety neurotics, indicating that multidimensional treatment approaches may be needed to totally combat the incapacitating effects of the illness.

A clear and operational delineation of specific outcome criteria became all the more important given that there was only a moderate relationship between outcome measures on hospitalisation, social contacts, employment status, and symptomatology (Strauss et al,, 1972). Inter-relationship between different outcome measures accounted only for 7% to 14% of the total variance, Strauss and his colleagues concluded that ".... schizophrenic outcome is a mixture of some general factor affecting the level of all functions, together with considerable individual variation of the separate areas". This failure to acknowledge the varying characteristics of outcomes and disparate courses within the schizophrenic disorder tended to add to the existing confusion and misconception regarding its pathological nature. Rather than holding on to a global feeling of optimism, a better understanding of the more benign versus the more resistant incapacities of the schizophrenic patient becomes a vital step for better resource allocation in this area.

This research report deals with the application of a set of operationally definable criteria for studying outcomes in schizophrenic patients in our local Hong Kong Chinese setting. The outcome pattern of 153 schizophrenic subjects is reported. A clearer demarcation of the more benign symptoms compared with the more resistant handicaps resulting from the schizophrenic illness is also attempted. The issue of cost and allocation of limited and valuable resources to the alleviation of the schizophrenic plight will also be addressed.



All patients with a diagnosis of schizophrenia for at least one year since initial contact were selected from the outpatients clinic of the Psychiatry Department, University of Hong Kong. Attending psychiatrists were asked to affirm the diagnosis of schizophrenia in the prospective subject population. These patients were then invited in writing and contacted by the clinic nurses to participate in the study. Subjects were informed that the basic purpose of the study was to assess their general pattern of recovery. Patients who consented to participate were screened by the project psychiatrist on the DSM-lll-R criteria and also applying the inclusion and exclusion screen adopted by the International Pilot Study of Schizophrenia (W.H.O., 1979). The flexible system of diagnosis was also used as an additional diagnostic screen because it was developed and considered as "reflecting more than any other how the term 'schizophrenia' is actually used throughout the world" (Strauss and Gift, 1977). 153 subjects consisting of 73 females and 81 males were recruited into the study. Their age range was from 17 to 47 years with a mean age of 32 years. The length of time since first onset of the schizophrenic illness ranged from 1 year to 18 years. 22% of subjects had their schizophrenic illness for less than three years, 30% for three to six years, 16% from seven to nine years, and 32% for ten years and more. 48% of the subjects had nil or one psychiatric admission for the schizophrenic illness, 35% had two to three admissions, and 17% had 4 or more psychiatric admissions.


All subjects were assessed independently by the project psychiatrist and clinical psychologist who were blind to each other's results. The project psychiatrist evaluated and recorded subjects' psychiatric and medication status on two structured interview schedules - the Diagnostic Intake Schedule and the Illness and Symptom History Schedule. The clinical psychologist recorded basic demographic and past history data on the subjects, and also recorded their social, self and occupational adjustment on two other standard structured interview schedules: namely the Demographic and Past History Schedule and the Outcome Evaluation Schedule.


To test the reliability. of the outcome pattern, 97 subjects (52 females and 45 males) were successfully followed-up with a reassessment of their outcome one year subsequent to the initial assessment.


Four intake and outcome structured interview schedules were used at the initial assessment stage. Some sections of the different schedules were adapted from the IPSS instruments (W.H.O., 1979). All interview schedules were pretested on 5 schizophrenic subjects before the study proper was conducted. The test-retest reliability of two points of assessment as conducted by the two assessors on the same subjects over a period of two weeks yielded an average correlation of 0.85 (range: 0.8-0.98). (Details of the interview schedules used are available from the authors).



Of all the 153 subjects, 10% received psychiatrist rating of the symptomatic outcome as 'satisfactory', 61% as 'fair' while 26% was rated as 'poor' and 3% as 'very poor'. Over 70% of subjects was relatively free from florid psychotic symptoms at interview.


This outcome measure assessed the symptomatic status over the course of the patient's illness. The categorical classification of the course of illness was based on the criteria stipulated in the DSM-III-R (A.PA, 1987). The majority (75.5%) of patients were 'in remission' at the time of study assessment. 2.5% were rated as 'subchronic', 12.2% as 'chronic', 1.8% as 'Subchronic with acute exacerbation' and 8.0% as 'chronic with acute exacerbation'.


A breakdown of individual symptoms at first psychiatric contact (as recorded in patient files) compared to those recorded in the study's assessment is presented in tables 1 and 2. A marked reduction of all prominent psychotic symptoms is noted.


Looking at the various ratings on the continuity and quality of the patients' work engagements over the previous one year, 46.4% were employed continuously, 17.5% for about 3/4 of the year's working hours, 11.4% for about 1/2 of the year's working hours and 9.0% for about 1/4. 15.7% had no useful work at all.

For the quality of useful work in the past year, 11.4% had productive and good work record most of the time, 34.9% with satisfactory work record for over half of the time, 25.9% with mediocre work record for about half of the time and 6.0% had poor work output for more than half of the time. 21.8% showed inability to cope with work most of the time.

So slightly more than half of all the 153 subjects were regularly employed and about half of all subjects had a satisfactory work record.


29.5% of subjects were married without divorce or separation, or dates regularly, 3.6% were married with conflicts causing brief separation, or dates sometimes. Of those who were unmarried or separated: 5.4% dates infrequently, 1.2% rarely dates and 60.3% never dates. So less than one-third of all the subjects were married. Of the reminder, a great majority lived a celibate life and engaged in no dating behaviour.


Concerning the quality of social relations in the past year, 1.8% of all subjects reported to have one or more relationships that is 'close', 9.6% 'rather close' and 21.7% 'moderately close' relationships. 18.7% had one or more 'only rather superficial' relationships and 48.2% had only very superficial relationships. So about one-third of the subjects reported having a rather close interpersonal relationship. Others admitted to having only superficial acquaintances.

In terms of the frequency of meeting with friends, about half of the subjects were seeing friends at least once a month. 12.7% met with friends on average at least once per week, 10.3% about once every 2 weeks and 24.8% about once a month. Others made only very irregular and infrequent social contacts outside of their home and work settings. 38.3% of the subjects did not meet with friends except "over back fence", or at work or school. 13.9% did not meet with friends at all under all conditions.


Relax intentionally 97% of all subjects were able to independently meet Engage in conversation with their own basic needs.


The frequency and range of daily life chores and leisure activities engaged in by the subjects is listed in Table 3 below. Most subjects engaged predominantly in rather passive forms of activities (e.g. watching TV, listening to radio). Activities involving self discipline (e.g. go to church, exer­ cise), social visits/interaction (e.g. visit friends, play mah­jong) and self initiatives (e.g. go to movies, picnics, relaxation, and cultivating self interests) were much less frequent. Most subjects were adequate in activities of a routine and repetitive nature, e.g. helping in household work, and going to work for those who were employed.


Quality of life

The quality of life rating was based on all the information gathered by the clinical psychologist through the individual interview with subjects. This rating was made on a global clinical intuitive basis. 40% of the subjects were judged to be living at least a moderately f ull life at the time of assess­ ment (2% very full life, 10% full life, 28% moderately full life). However the majority of subjects were rated as living rather impoverished (10% of vegetative existence) and re­ latively empty lives(50%).


An intercorrelation matrix was calculated between the different major outcome measures and the results are listed in table 4. The quality of life measure had the best correlation with most other measures including total symptoms, work quality, social relationship. The most independent measure was the number of hospital admissions which had no significant correlation with any of the other outcome measure. Most of the outcome measures, despite their significant correlations, had only a low to moderate level of correlation with one another. Of the statistically significant correlations, the psychiatric symptoms rating had the lowest correlation with other psychosocial measures of outcomes. The general pattern of significant intercorrelations ranged from a low of -0.20 between quality of life and social relationship.


To obtain a more reliable pattern of outcome in a more succinct summary form, a factor analysis on all the outcome measures was performed. Using a principal components analysis, and an oblimin rotation on the SPSS-X program (SPSS Inc., 1988), five outcome factors accounting for 73.4% of the variance were generated. The five factory are listed in table 5.

The quality of life factor was loaded highly with measures of frequency and quality of social interaction, quality of life rating by the psychologist, and the psychiatrist's rating of overall symptomatic outcome.

The hospitalization factor was best repesented by the two measures of number and total duration (days) of psychiatric admissions.

The psychotropic medication adjustment and psychiatric symptoms factor consisted predominantly of the psychiatrist's evaluation of upward or downward adjustment of psychotropic medication and the manifestation of incoherence and psychotic symptoms in the subjects.

The work and independence factor consisted primarily of measures on subjects' independence in basic needs, and continuity of employment and quality of work production.

The daily life function and heterosexual adjustment factor was loaded primarily with two measures: namely daily life pattern of activities, and presence/absence and quality of heterosexual contacts.


A comparison of the outcome status of the 96 subjects who were assessed both initially and successfully followed up after one year was conducted. The one year outcome on all psychosocial adjustment measures was stable and highly consistent with the previous assessment. The range of the significant correlations between initial assessment and follow-up assessment ranged from a low of 0.42 relating to continuity of work employment to a high of 0.78 relating to heterosexual relationship adjustment. In contrast, symptomatic measures of patients' outcome were much less stable over time. The only statistically significant correlation, which was low in comparison with the psychosocial measures, was in the overall symptomatic rating. Other measures relating to adjustment of psychotropic medications, number and duration of psychiatric hospitalization had a low and non-significant correlation over time (see Table 6).


This study aimed to look into the outcome patterns in schizophrenia. The specification of the nature and heterogeneity of outcomes within the schizophrenic population may shed light on low one may best plan and allocate the existing limited and valuable therapeutic resources.


All subjects selected fulfilled the DSM-IlIR criteria for schizophrenia, and the mean number of diagnosable symptoms in the flexible system was 7.5. This means that a relatively stringent criteria for the selection of schizophrenic subjects had been adopted with only 0-1% of false positives (Carpenter et al., 1973). However, no chronic schizophrenic subjects were included. Also, our group of subjects did not consist of the "cured" group of schizophrenics who either defaulted follow-up on their own or who were discharged with an "open appointment" to come back only when problems arise. The subjects recruited are thus not representative of patients at the two extreme ends of the schizophrenic disease spectrum. The distinct advantage in selecting the bulk of the middle range group of schizophrenics is that since they form the bulk of patients who will utilise the most active therapeutic resource, whose wellbeing may also be more vitally dependent on how resources are being utilised, findings therefrom may enable a more confident estimate in terms of cost and resource allocation planning.

Due to tir-:ie limitation, a sizable number of the original subjects were not followed up. However, the follow-up sample was similar to the original sample on many major parameters: including educational background, occupational and marital status. Students' t-test and Chi Squared tests also indicated no difference in the number of previous admissions, duration of psychiatric admissions, symptomatic presentation at onset of the illmess, and duration of their illness.


The factor analysis conducted on all the outcome mea­sures (Table 5) provides a better underestanding of the pat­ tern of outcomes in schizophrenia. The largest outcome factor is the social adjustment and quality of life factor. So­cial adjustment, quality of life, and to a lesser extent the overall symptomatic outcome are all loaded highly on this factor. The quality of life rating seems to be an important summary statement of the patient's standing in relation to him/herself and his social environment. The symptomatic rating is on the other hand, a good summary statement of the patient's psychiatric state.

The second factor consists of measures of previous psychiatric hospitalization. This factor has a significant lack of any relationship with all the other outcome measures. The number of previous hospitalization should therefore not be taken as indicative of a poor outcome prognosis. This finding seems to be partly supportive of Zubin and Spring's (1977) vulnerability hypothesis and stress diathesis model where they noted that the only permanent ':lttribute of the schizophrenic patient was his/her vulnerability. The floridly psychotic presentation of the schizophrenic patient which often leads him/her to be hospitalized seems to be a less persistent state of the illness, The deteminants of the frequency and duration of psychiatric hospitalization are probably qualitatively different from the determinants of the other enduring psychosocial handicaps.

The medication dosage and adjustment factor is a measure indicating increase, decrease or no change of the subjects' usual intake of psychotropic medications is also highly loaded onto this factor. As this factor was separated in the factor analysis from the hospitalization factor, two points may be noted. Firstly, the frequency and duration of past psychiatric hospitalization is not a key determinant of the extent and severity of present psychotic symptoms. Secondly, hospitalization and positive symptoms and medication adjustment may be determined by different factors. This was noted earlier in Brown et al.'s (1968) study that hospitalization of schizophrenic patients may be dependent on the patient's symptomatic manifestations, as well as other factors such as the patient's relative position and hierarchy in the surrounding social environment. Increase in psychotic symptoms, on the other hand, may be due to a host of other factors ranging from increase in stresses, default in taking medications, and other idiosyncratic vulnerability factors.

The fourth and the fifth factor together comprise about 17% of the outcome variance. Work adjustment and level of independence are the two major measures making up the fourth factor. The measures on pattern of daily life activities and presence of any meaningful heterosexual relationship are loaded most highly onto the fifth factor. The measure on level of independence in factor four is less important in characterizing differences between subjects as the variance within this measure is very small; and 97% of all subjects were rated as being fully independent in daily chores and required no help at all.

The clear delineation of the outcome measures into five relatively distinct factors is indicative that different outcome measures may not necessarily be concordant with one another. Varying combinations of positive and negative outcomes in different areas of functioning are possible, where each may be dependent on a complex nexus of intervening and historic factors. The same picture is consistently seen in the intercorrelational analysis and will not be further discussed.


The outcome status of the schizophrenic subjects varies depending on the area of functioning being considered. The best outcome relates to symptomatic ratings. Over 70% of all the subjects were assessed to be asymptomatic and in remission.

Outcome measures on other indexes are notably poorer than the symptomatic ratings, and can hardly be rated as optimistic. Only 46% of the subjects were employed continuously and had a satisfactory or good work record. About 30% of all subjects were working marginally with a poor work output. A significant 20% of the subjects were unable to cope with any work at all.

Heterosexual relationship is the poorest of all the ourcome indexes. Less than one-third of all subjects were married or dating regularly. The majority of the remaining twothirds were unmarried and reported no dating experiences over the previous 12 months prior to assessment.

65% of the subjects had only a ]imited to superficial social relationship. Only one-fifth of the subjects had one rather close to close friend. More than 59% of the subjects had no social interactions or only met with others "over the back fence". Another 22% of the subjects met with friends at least once every two weeks. The remaining subjects (25%) reported meeting with others only irregularly.

The daily life activities measure is a reflection of the range of daily life activities subjects usually participated in. It is also a reflection of the extent and amount of pleasure subjects derived from living. It was noted that most activities that the subjects participated in are of a basic, routine, and rrepetitive nature. The majority are apparently unable or unwilling to be active or self-initiated in cultivating any interests or a constructive life pattern. For example, while going on picnics or morning walks form a common past-time of most people in Hing Kong, over 70% of the subjects had never engaged in such activities. Likewise, 66% of the subjects had never intentionally relaxed themselves, 69% of the subjects never engaged in past-times of an interpersonal nature like playing mah-jong, or other socialisation. 72% of the subjects had no hobbies. The overall impression gained of the schizophrenic subjects is that the majority may get by in life, but the variety of their life activities is low, their self initiative is poor, and the enjoyment derived from living is limited. This impression is testified in the global quality of life rating where over 60% of the subjects were rated as living a relatively empty life to a vegetative existence. Only 12% of the subjects were rated as having a full to a very full life.

Commenting on the overall pattern of results, "inevitable deterioration" (Kraepelin, 1913) cannot be said to be an adequate representation of our schizophrenic outcome findings, To a limited extent, the results are supportive of Zubin et al.'s (1983) notion of a benign metamorphosis in schizophrenia. This statement must however be interpreted with great caution. An increasingly positive outcome may only be confidently applied to areas of daily self care, and positive symptomatic manifestations of the schizophrenic subjects. However, non-encourageing outcomes in work, social, and self adjustment still characterise over half of the subjects studied. The majority of the schizophrenic subjects may be expected to live a celibate life. At least half of all the schizophrenic subjects studied may be considered as marginal and incapable of enjoying life as normal people. While the majority of the schizophrenic subjects are capable of basic self care, have a reduced level of disrupting psychotic symptoms, and may generally be expected to be able to stay clear from long stay psychiatric institutions, the majority are still disabled when the ability to live a fruitful, constructive life with adequate development of one's potentials are concerned.


The pattern of the schizophrenic outcome obtained at the one year follow-up testifies to the consistency of the above picture (Table 6). The most consistent of the outcome measures relate to heterosexual relationship (rho = 0.78), independence in badic needs (rho=0.71), quality of work output (rho = 0.66)m and daily life activity pattern (rho = 0.60). The least consistent of the outcome measures are the duration (rho = 0.03) and number of psychiatric hospitalization (rho = 0.13), and adjustment of psychotropic medications (rho = 0.11). The different pattern and consistency of the above measures indicates· that the traditionally emphasized measures of symptomatic improvement, psychiatric hospital discharge, and medications are much less enduring and probably the easier of the schizophrenic related problems to tackle.


The schizophrenic associated psychosocial handicaps may be due to a disruption of the development of the vital skills required for successful adaptation caused by the schizophrenic process, or may be a direct accompaniment and "symptom" of the schizophrenic process. Whatever the underlying cause/s, the major and most sticky of the schizophrenic handicaps seem not to be their florid psychotic symptomatology. Rather, very fundamental aspects of life functions and skills for optimisation of socialisation, enjoyment, work, and heterosexual relationships are clearly deficient. The cost of such handicaps for the individual and society are still pessimistically high. Nelson and Smart (1980) drew attention to the direct as well as the indirect costs of any illness. Direct costs are treatment costs covering costs of diagnosis, hospitalisation, drugs, rehabilitation and special nursihg care. Indirect costs are those incurred because of the illness, e.g. earnings loss, premature death. The authors would also like to add to the list of indirect costs one of reduced quality and enjoyment of life. The direct cost allocated to the treatment and alleviation of any illness should ideally be spent to reduce as much as possible the indirect costs incurred by the individual and society. Resource allocation in the alleviation of the schizophrenic plight, however, are primarily devoted to symptomatic control and inpatient facilities. This focus is clearly misplaced, and it is difficult to foresee how the predominant hospitalisation and the medication foci in treating active schizophrenic symptomatology may contribute to improving the individual's life functions. Lapsely and Cass (1981) was in favour of expanding on outpatient crisis care as a more cost-effective form of treatment. Hoult et al. (1983) also noted that such care were more acceptable to patients and their families when compared to inpatient care. The schizophrenic patient may have a much better chance of improving and developing his resources in living a more productive life when he is out of the artificially protective hospital environment, but aided systematically within his community in real life functioning skills. Important guidelines of community support systems developed by the NIMH (Turner et al., 1978) indicated that long-term (and even indefinite) supportive services in both living and work settings may be much more cost effective than inpatient care.

Another promising approach recently proposed (English and McGarrick, 1989) was the active treatment concept. Resource allocation, must as a fundamental condition be based on an individually oriented treatment package. Psychiatric services should be provided under an indi­ vidually worked out diagnostic plan covering the undividual patient's restorative needs and potentials. Such treatment plans should ideally also be multi-model in nature given the present finding that different outcome measures may be determined differently and their progression somewhat desynchronous with one another.

Resource allocation must also be more cost effectively evaluated. The cost involved in the treatment of schizophrenia is no longer one of control of the florid psychotic symtomatology. The real challenge of the day is in the alleviation of the persistent and significant psychosocial handicaps which are still a long way from witnessing a "benign metamorphorsis".


The authors would like to express their grateful thanks to the Committee on Research and Conference Grants, the Lee Wing Tat Fund, and the Medical Faculty Fund, University of Hong Kong, for generously supporting our research.


American Psychiatric Association (1987): Diagnostic and Statistical Manual of Mental disorders, (DSM-llIR), APA, U.S.A.

Bleuler M. (1974): The long-term course of the schizophrenic psychoses. Psychological Medicine, 4, 244-254.

Bleuler M. (1978): The schizophrenic disorders: long-term patient and family studies. Clemens S.M. (trans.). New Havens, Yale University Press.

Brown G.W. and Birley J.L.T. (1968): Crises and life changes and the onset of schizophrenia. Journal of Health and Social Be­ havior, 9, 203-209.

Carpenter W.T.Jr. and Strauss J.S. (19730: Flexible system for the diagnosis of schizoprenia: report from the WHO international pilot study of schizophrenia. Science, 182(Dec): 1275-1278.

Ciampi L. (1980): The natural history of schizophrenia in the long term. British Journal of Psychiatry, 136, 413-420.

English J.T. and McCarrick R.G. (1989): The econimics of psychiatry; in Kaplan J.I. and Sadock B.J. (eds.): Comprehensive Textbook of Psychiatry, 5th ed., vol. 2., Williams & Wilkins, Baltimore, London.

Falloon !.R.H. (1983): Relapse in schizophrenia: a review of the concept and its definitions. Psychological Medicine, 13, 469- 477.

Hoult J., Reynolds I., Carbonneau-Powis et al. (1983): Psychiatric hospital versus community treatment: the results of a randomosed trial. Australian and New Zealand Jounal of Psychiatry, 17, 160-167.

Kraepelin E. (1913): Psychiatrie; ein Lehrbuch fur Studierende und Artzte, 8th ed. Leipzig, Barth.

Lapsely H. and Cass Y. (1981): The economics of institutional and noninstitutional care for psychiatric patients. In Tatchell P.M. (ed.): Economics and Health, 1981, Australian National University Press, Canberra.

Lehmann H.E. (1981): Psychopharmacological treatment of schizophrenia. Schizophrenia Bulletin, 7, 27-45.

Lo W.H. and Lo T. (1977): A ten-year follow-up study of Chinese schizophrenics in Hong Kong. British Journal of Psychiatry, 131, 63-66.

Nelson T. and smart M. (1980) The social context of mental disorders. Journal of Mental Science, 121, 324-329.

Pietzcker A. and Gaebel W. (1987): Prospective Study of course of illness in schizophrenia: Part I. Outcome at 1 year: Schizophre­ nia Bulletin, 13, 287-297.

Rachman S. and Hodgson R. (1974): synchrony and desynchrony in fear and avoidance. Behaviour Research and Therapy, 12, 311-318.

SPSS Inc. (1988) SPSS-X: User's Guide, 3rd ed., SPSS Inc., Chicago, U.S.A.

Strauss J.S. and Caroenter W.T.Jr. (1972): The prediction of outcome in schizophrenia: I. Characteristics of outcome. Archives of General Psychiatry, 27, 739-746.

Strauss J.S. and Carpenter W.T.Jr. (1974a): the prediction of outcome in schizophrenia: II. Relationships between predictor and outcome variables. Archives of General Psychiatry, 31, 37-42.

Strauss J.S., Carpenter W.T.Jr., and Bartko J.J. (1974b): speculations on the processes that underlie schizophrenic symptoms and signs. Schizophrenia Bulletin, 1, 61-75.

Strauss J.S. and Gift T.E. (1977): Choosing an approach for diagnosing schizophrenia. Archives of General Psychiatry, 34, 1248-1253.

Strauss J.S. and Carpenter W.T.Jr. (1978) the prognosis of schizophrenia: rationale for a multidimensional concept. Schizophrenia Bulletin, 4, 56-67.

Turner J.C., Ten Hoar W.J. (1978): The NIMH community support program: Pilot approach to a meeded reform. Schizophrenia Bulletin, 4, 347.

Wing J.K. (1987): Has the outcome of schizophrenia changed? British Medical Journal, 43, 741-753.

World Health Organisation (1979) Schizophrenia: and international follow-up study. John Wiley & Sons, Toronto.

Zubin J. and Spring B. (1977): Vulnerability-a new view of schizophrenia. Journal of Abnormal Psychology, 86, 103-126.

Zubin J., Magaziner J., and Steinhauer S.R. (1983): The metamorphosis of schizophrenia: from chronicity to vulnerability. Psychological Medicine, 13, 551-571.

* Peter W.H. Lee MSoc Sc, PhD. Senior Lecturer, Department of Psychiatry, University of Hong Kong.
F. Lieh-Mak, MD, FRCPsych, FRANZCP, FAPA, JP. Professor and Head, Department of Psychiatry, University of Hong Kong.
K.K. Yu, MBBS, MRCPsych. Formerly Lecturer, Department of Psychiatry, University of Hong Kong.
J.A. Spinks BSc, PhD. Senior Lecturer, Department of Psychology , University of Hong Kong.* Correspondence: Departm<::nt of Psychiatry, Queen Mary Hospital, Pokfulam, Hong Kong.

View My Stats