1.1Introduction
1.2historical view of course of psychosis
1.3Heterogeneity of outcome in schizophrenia
1.3.1Questioning the Dichotomy of Outcome
1.3.2Courses of Illness
1.3.3World Health Organisation Prospective Longitudinal studies
1.4Single episodes
1.5Predictors of outcome
1.5.1World Health Organisation Studies
1.5.2Other Prospective studies
1.5.3Retrospective and Cross-sectional Studies
1.5.4Other Factors including Psychological Factors
1.5.5Summary of Predictors of Outcome
1.5.6Reviews of Long-term Studies
1.6methodological issues in outcome studies
1.7Conclusions about Course of Illness Studies
1.8why consider outcome in psychosis?
1.8.1Economic Cost of Psychosis
1.8.2Critical Period Hypothesis
1.8.3Early Intervention
1.8.4Summary
1.9conclusions
CHAPTER 1. PREDICTORS OF OUTCOME IN SCHIZOPHRENIA
1.1Introduction
The idea of this thesis began with a question aboutgood outcome. Research and clinical practice usually has a focus on poor outcome and ways of ameliorating that undesirable situation. Those who recover leave services; the goal is achieved. However, some people, sometimes contrary to all expectations, do very well; recovery is unpredictable.
Recovery in schizophrenia is particularly unpredictable (Fenton and McGlashan, 1987). This thesis begins by addressing questions about those with best outcome in schizophrenia or schizophrenia-related disorders; those with a single episode of illness. Who are they and can we learn anything about those with a single episode which will help those with worse outcome? In this chapter the literature relating to course of illness and outcome in schizophrenia will be reviewed,followed by reasons for understanding more about outcome in schizophrenia.
1.2historical view of course of psychosis
The course and outcome of schizophrenia have received considerable attention since the condition was first described as dementia praecox by Kraepelin in 1896. While a century’s research has given rise to an extensive literature regarding its natural history, some of the original questions regarding the prognosis and outcome of schizophrenia remain just as pertinent today.
Kraepelin originally considered dementia praecox to have a deteriorating course, altering this view when he found 13% of his severe hospital-based patients to show spontaneous remission. He later revised this estimate, stating that only 2-4% showed lasting recovery. Bleuler (1911, translated 1950) described similar findings regarding the course of schizophrenia, stating both that it was possible for people with mild schizophrenia to exist outside hospital and not get “very much worse” and that it was impossible to consider his patients cured of schizophrenia as they never attained “restitutio ad integrum”. This mixed picture, emphasising the expectation of a poor prognosis without precluding good outcome, guided subsequent thinking as is illustrated by the Diagnostic and Statistical Manual (DSM-III; American Psychiatric Association, 1980) which states that
“a complete return to premorbid functioning is unusual – so rare, in fact, that some clinicians would question the diagnosis. However, there is always the possibility of full remission or recovery, although its frequency is unknown” (American Psychiatric Association [APA], 1980 p. 185).
While this explicit statement concerning poor prognosis has been dropped from subsequent revisions, deterioration in function remains a central criterion for the diagnosis of schizophrenia in DSM-IV (APA, 1994). For example, DSM-IV states that social or occupational dysfunction will be present, asserting that individuals will display a level of functioning in a major areasuch as work, interpersonal relationships or self-care below that achieved prior to the onset of a schizophrenic disorder.
Despite the continued prominence of deterioration of functioning associated with a label of a schizophrenic disorder, the early acknowledgment of the chance of remission suggestsother outcomes are possible.
1.3Heterogeneity of outcome in schizophrenia
1.3.1Questioning the Dichotomy of Outcome
Traditionally,with the finding of the possibility of remission, there hadbeen a tendency to dichotomise outcome intoeither a chronic, deteriorating course or a full remission. The extent of this dichotomy of outcome had led some to suggest that those with better outcome were suffering from a different disorder, distinct from that of true schizophrenia (e.g. Langfeldt, 1969). This idea was criticised, but itspossibility emphasised the need to have a better understanding of the course of schizophrenia.
Vaillant (1963, 1978) investigated the issue of diagnostic differences between those who recovered and those with a chronic course of illness. In a 50 year retrospective study of people “recovered” from schizophrenia, he reports that three outcomes were shown by the 12 people with “remitted” schizophrenia after their first episode: some were found to remain well with no further admission, some were readmitted but with a non-schizophrenic diagnosis and some were hospitalised again, sometimes some years later. In a later 10 year follow-up study of schizophrenia, Vaillant (1978) reviewed the casenotes of people diagnosed at least ten years previously. Thirty of this sample were identified retrospectively and 26 were identified prospectively as having achieved complete remission. Twenty of this group had subsequently had chronic courses of illness. Twenty-one of the thirty-one patients were interviewed ten years after their admission (not necessarily a first episode) and were found to be doing well in terms of social independence and employment, even though some had had brief relapses. While over 30 percent of the remitted sample had had their diagnosis changed to an affective psychosis, so had 30 percent of those with chronic illness. He argues that remission in schizophrenia would not be improved by refining the diagnosis of schizophrenia as those who improved were similar to those who had a chronic course of illness. He concludes that “remitting schizophrenics are not necessarily a discrete type of schizophrenic requiring new labels” (p. 84).
Following a study by Strauss (1969) in which he argued that symptoms of schizophrenia are best seen along a continuum, Strauss and Carpenter (1972) investigated outcome in schizophrenia. Citingthe work of Zubin et al. (1961) whoreport an excess of 800 studies examining outcome in schizophrenia conducted by that date, Carpenter and Strauss state that the studies had led to a number of valuable suggestions about the importance of certain variables in predicting outcome in schizophrenia. However, they criticise the studies for their failure to investigate varying degrees of outcome dysfunction and their use of terms such as “improved” or “unchanged”. In a follow-up study of admissions to a psychiatric hospital, Strauss and Carpenter conducted the Present State Examination (Wing et al., 1967), the Psychiatric History Schedule and Social and Demographic Information Interview developed by the authors and the Inpatient Multidimensional Psychiatric Scale (Lorr et al., 1962). Information at two years’ follow-up was available for 111 people, 85 of whom had a schizophrenia-related disorder. They found that use of global terms such as “improved” or “unimproved” masked the reality of a wide range of possible outcomes in schizophrenia. In their sample, outcome ranged from severe dysfunction to no dysfunction in the four areas of social contacts, employment, duration of hospitalisation and symptom severity. They argue that their findings give no support for “conceptualizing schizophrenia as having homogenously poor outcome”(p. 745); in fact the population with non-schizophrenia diagnoses showed considerable overlap with those with a schizophrenia diagnosis.
Further studies have emphasised the heterogeneity and complexity of outcome (e.g. Strauss and Carpenter, 1977; Carpenter and Strauss, 1991; Harding et al., 1992). For example, Harding, Zubin and Strauss (1992) reviewedfive long-term studies of schizophrenia. Quoting the work of Bleuler (1972), Harding et al. (1987), Ciompi and Müller (1976), Huber et al. (1979) and Tsuang et al. (1979), in which the courses of illness of a total of1303 patients were tracked over a range of 22-37 years using both retrospective and prospective methods, Harding et al. conclude that one-half to two-thirds of the sample achieved significant recovery. They claim that diagnostic criteria cannot successfully predict long term functioning and statethat the roles of personality and environment in “aiding or interfering with the expected outcome are too powerful to permit uniformity” (p. 31).
1.3.2Courses of Illness
In an attempt to understand the various outcomes in schizophrenia, people havetaken the mode of illness onset, number of episodes and level of remission into account, and described a number of possible illness courses. Using data from 1642 people diagnosed with schizophrenia between 1900 and 1962, of whom 289 were alive and followed up an average of almost 37 years later, Ciompi (1980a,b) described a minimum of eight main patterns of outcome. Finding 49 percent to have had a favourable course of illness and 42 percent to have had unfavourable courses of illness, he describes four illness courses with insidious onsets, and four with an acute onset of illness. The other aspects of course consider “type of evolution” meaning whether or not the psychotic episodes remitted and if they did, to what extent they did so, and “end-state”, referring to the presence and severity of ongoing psychotic symptoms.
Huber, Gross, Schüttler and Linz (1980) investigated course of illness in 502 people with schizophrenia an average of 22 years later. They reduced 73 possible courses of illness to 12,ranging from best outcome of “monophasic” type (single episode with complete remission) and “polyphasic” (average of five episodes ending in complete and permanent recovery), to Type XII which constituted ongoing “typically schizophrenic defect psychoses”(p. 597). Social remission was used in the classification of outcome in addition to presence of psychotic symptoms.Remission was possible over the entire duration of illness (9 to 59 years) and occurred “at approximately the same rate in patients whose durations of illness range from short to long” (p. 595). While finding that those with the shortest duration of illness (9 to 14 years) had the most favourable outcome, theystate that “there are some data to indicate that psychopharmacological treatments are responsible for the better prognosis in this group” (p.595).
A set of studies investigating long term outcome of individuals with a first episode of schizophrenia has been conducted by the Scottish Schizophrenia Research Group (1987, 1988, 1989, 1992). The number of participants in these studies is much smaller than that of a number of other studies, and are not conducted with an epidemiological sample. However, the group were able to follow up many of their sample of 49 people for five years, investigating the influence of a number of factors on course of illness, including the influence of medication and any cognitive intellectual decline. In addition, family members were included in the study.
In a report of one year outcome (Scottish Schizophrenia Research Group, 1988), 41 of the original sample of 49 were followed up. At the twelve-month follow-up, assessments of social and demographic situation were repeated, together with psychometric assessments (Block Design and Similarities subtest of the Weschler Adult Intelligence Scale; Weschler, 1955; Progressive Matrices; Raven, 1938, Mill Hill Vocabulary Scale; Raven, 1948 and the Digit Copying test; Gibson and Kendrick, 1979). No significant changes were found in marital or domiciliary arrangements at twelve months, but more were found to be unemployed. Thirty-five percent of those followed up showed no positive symptoms at twelve months, but of those whose outcome could be determined (n=36), 44 percent had “good” outcome (meaning no positive or negative symptoms at the follow-up assessment and no relapse over the whole period; i.e. a single episode) and 56 percent had “poor” outcome. Performance on the psychometric assessments was found to be the same as at the initial assessment or better: improvements were found on the matrices, digits and Block Design. Interestingly, patients’ symptoms of schizophrenia (but not non-schizophrenic symptoms) at one-year follow-up correlated with distress expressed by relatives.
An examination of two year outcome for 47 of the original sample (Scottish Schizophrenia Research Group, 1989) showed that unemployment had risen slightly. Of the 38 people whose outcome could be determined, 37 percent had a good outcome according to the same criterion reported above. More men and people with negative symptoms of schizophrenia were found to be in the poor outcome group, and more people in the good outcome group had been found to have good response to medical treatment during the first five weeks of their first admission and were continuing with maintenance medication. Performance on the assessments of cognitive function showed further improvements in block design, matrices and digits, and no change in other measures. Again, significant associations were found between the relatives’ distress and the patients’ positive and negative symptoms of schizophrenia.
In a report of five-year outcome (Scottish Schizophrenia Research Group, 1992), 44 of the original sample of 49 were followed up. No differences were found in marital status from that recorded at first assessment for 37 (88%) of the sample. A further reduction in employment status was found. Analysis of the survival rate to first relapse found that approximately 20 percent of the sample relapsed each year over the first 3.5 years. Only two people were found to relapse for the first time in the subsequent 18 months. Thirty percent of the sample had not relapsed by five years’ follow-up, but some of these had symptoms of psychosis at follow-up. Only 18 percent (3 men and 4 women) had not relapsed and reported no symptoms at assessment. Those who had relapsed were more likely to be unemployed and had taken antipsychotic medication for a longer duration. No differences were found between the groups on gender or age at onset. Unlike previous findings, there was no association between relatives’ distress and patients’ symptoms at five years’ follow-up.
Shepherd, Watt, Falloon and Smeeton (1989) investigated outcome in 107 people with schizophrenia using a prospective follow-up over five years. Using the data from those in their first episode at intake into the study (N=49), they report four possible courses of illness. These were: one episode only, no impairment; several episodes with no or minimal impairment; impairment after the first episode with subsequent exacerbation and no return to normality and impairment increasing with each of several episodes and no return to normality. Using the terminology of Ciompi (1980), these courses of illness describe the “type of evolution” of illness and the “end state” but not the type of onset.
Thara, Henrietta, Joseph, Rajkumar and Eaton (1994) report results of a ten-year follow-up study of a cohort of 90 first-onset people with schizophrenia. Assessments with the Present State Examination (PSE; Wing, Cooper and Sartorius, 1974), Psychiatric and Personal History Schedule (PPHS; XX) and Interim Follow-up Schedule (IPS; Sartorius et al., 1986) were made every year for ten years. At ten years, information was available for 76 (84%) of the sample. The most common course of illness was that of one or more relapses (two or more episodes), and 37 (49%) of the sample were foundto have this course. Eleven (14%) were found to have a single episode with complete recovery, 2 (3%) had a single episode with ongoing residual symptoms, 21 (28%) had multiple episodes with residual symptoms and 5 (7%) were continuously psychotic over the ten year period. Factors found to predict relapse will be described in section 1.5.
Thus while it is now generally accepted that course of illness is heterogeneous, there is little consensus regarding the complexity of this heterogeneity.[I wondered if I should reproduce some of the visual representations of course. Eg Shepherd et al., Thara,Ciompi, and others. Not all have visual representations, but some do. There are slight differences between them.]
1.3.3World Health Organisation Prospective Longitudinal studies
Longitudinal studies are valuable for providing a comprehensive evaluation of the outcome of psychosis. One important collection of studies is the World Health Organisation (WHO) Collaborative Study on the Determinants of Outcome of Severe Mental Disorders known as the International Study of Schizophrenia (DOSMeD; Jablensky et al., 1992) and its precursor, the International Pilot Study of Schizophrenia (IPSS; Sartorius et al., 1977, 1986; Leff et al., 1992).
Leff, Sartorius, Jablensky, Korten and Ernberg (1992), describe the five year follow-up results of the IPSS. Recruited from successive admissions to psychiatric services in centres from nine countries, 1202 people with a diagnosis of schizophrenia or an affective psychosis constituted the original cohort of participants. Two-year follow-up information was available for 77 percent of the original sample (see Sartorius, Jablensky and Shapiro, 1977) and 74 percent of the sample at five years. One centre (Taipei) withdrew from the study before the five-year point. The sample were assessed using the Present State Examination (PSE; Wing, Cooper and Sartorius, 1974), the Follow-up Psychiatric History schedule, the Follow-up Social Descriptions schedule and the Follow-up Diagnostic Assessment schedule, all designed specifically for the set of studies. A total of 52 patients (4.9% of the original cohort) had died by the five year follow-up, with suicide being the most common cause, accounting for 38 percent of the deaths. One third of the sample originally diagnosed with schizophrenia spent less than five percent of the five years in a psychotic episode, while one fifth spent over 75 percent of the time in an acute psychotic state. Agra, Cali, Ibadan and Washington were the centres with the highest proportion of people spending less than five percent of their time in a psychotic episode, and Agra and Ibadan also had the fewest number with worst outcome according to percentage of time spent in a psychotic episode.
Looking at patterns of illness course rather then percentage of time spent in a psychotic episode, seven patterns of illness course were identified: full remission of episode of inclusion with no further episodes, partial remission with no further episodes, at least one further non-psychotic episode with full remission between, at least one non-psychotic episode after inclusion admission with partial remission between, at least one subsequent psychotic episode full remission between, at least one subsequent psychotic episode, incomplete remission between and episode of inclusion continued with no subsequent remission. The results using these criteria are similar to those using time spent in a psychotic episode, with Agra and Ibadan having high percentages of people with one episode (42 and 33 percent respectively) and Aarhushaving 40 percent of their sample to have worst outcome in terms of being psychotic for the whole follow-up period. Information on social outcome mirrored that of psychotic outcome, although the majority of patients with schizophrenia did not suffer severe impairment of their social functioning.
Comparing results of the course of illness between the two-year follow-up and the five year point found that some people who had remained well at the two year point had subsequently had episodes of psychosis from which they had recovered. Also, however, it was found that people who had shown an unremitting course at two years went on to have a more favourable course of illness at the five-year point. Overall, as with the results of the two-year follow-up, people from the developing countries had a more favourable course as measured by both clinical and social outcome than those in developed countries.