Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A vulnerability-stress model for the course of schizophrenia ? Erickson, David Harry 1994

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1994-893524.pdf [ 4.05MB ]
JSON: 831-1.0099205.json
JSON-LD: 831-1.0099205-ld.json
RDF/XML (Pretty): 831-1.0099205-rdf.xml
RDF/JSON: 831-1.0099205-rdf.json
Turtle: 831-1.0099205-turtle.txt
N-Triples: 831-1.0099205-rdf-ntriples.txt
Original Record: 831-1.0099205-source.json
Full Text

Full Text

A VULNERABILITY-STRESS MODEL FOR THE COURSE OF SCHIZOPHRENIA 7 by DAVID HARRY ERICKSON B.A.  (Honours), University of Regina, 1978 M.A., Simon Fraser University, 1984  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY  in THE FACULTY OF GRADUATE STUDIES Department of Psychology  We accept this thesis as conforming o the required standard  November 1993 David H. 7 Erickson  95  In  presenting  this  thesis  in  partial  fulfilment  of the requirements for degree at the University of British Columbia, agree that the Library freely available for reference and study. I further agree that permis sion copying of this thesis for scholarly purposes may be granted by the department  or  by  his  or  her  representatives,  an  advanced  shall make it for extensive head of my  It is understood that copying or publication of this thesis for financial gain shall not be allowe d without my written permission.  (Signature)  Department of The University of British Columbia Vancouver, Canada  Date  DE-6 (2/88)  1].  ABSTRACT  Despite a prevailing paradigm that emphasizes an interaction of vulnerability and stress to account for the etiology of schizophrenia, diathesis—stress models of subsequent course and outcome of this disorder are rare.  Even the simpler stress—  process model, where the influence of stressors is mediated by supportive social relationships, has received little attention in studies of the course of schizophrenia. The objective of this study was to assess the following components of a diathesis—stress model as they predict the fiveyear outcome of first-episode schizophrenia: events;  (1) stressful life  (2) supportive social relationships;  ventricle size; and  (3) brain lateral  (4) smooth pursuit eye movements.  As part of the Greater Vancouver M.A.P. Project, we recruited first-episode DSM-III schizophrenia and affective psychosis patients.  At intake to the study, their social relationships,  smooth pursuit eye movement function, and brain ventricle size were assessed. at intake;  Life events in the previous year were measured  events over the following 18 months were assessed in  two later interviews.  Five years later we assessed outcome,  using a global rating of social and occupational functioning. Descriptive results showed substantial variability within the schizophrenia group at intake and outcome.  The trajectory of  adaptive functioning over time was remarkably similar for the schizophrenic and affective psychosis groups.  Of the four  11]  hypothesized predictors, associated  (p=.O3)  only social relationships were  with five—year outcome.  The number of life  events was not associated with five—year outcome, nor was either of the biological risk factors.  As a result, the predictor  variables could not be combined in either a stress—process model or a vulnerability—stress model of the course of schizophrenia. That social relationship variables are associated with fiveyear outcome supports earlier findings regarding 18-month outcome,  including the differing predictive roles for family and  nonfamily relationships.  The absence of hypothesized results  for the life events data probably indicates that too much time had passed between outcome and the events as measured.  Finally,  that brain ventricle size and eye-movement dysfunction predict 18-month but not five—year outcome may indicate that impairment due to biological factors is expressed only in the early stages of schizophrenia.  iv Table of Contents (1) Abstract (ii) (iii) (iv)  ii  Table of Contents  iv  List of Tables  vi  List of Figures  viii  Cv) List of Appendices (vi)  I.  II.  III.  ix  Acknowledgements  x  INTRODUCTION  1  OBJECTIVES  4  BACKGROUND AND LITERATURE REVIEW  6  Social Factors in the Course of Schizophrenia Stressful Life Events Social Relationships Stress—Process Models Biological Factors: I. Lateral Ventricle Size CT Imaging and Ventricle Size: General Issues Lateral Ventricles in Schizophrenia Correlates of Ventricle Size Ventricle Size and Affective Disorders Summary: VBR as a Predictor of Outcome  9 9 18 24  .  Biological Factors: II. Smooth Pursuit Eye Movements Systems Governing Eye Movements Recording and Scoring SPEM Theoretical Requirements for a Genetic Marker . Correlates of SPEM Dysfunction Summary General Issues in the Prediction of Course and Outcome Sociodemographic and Premorbid Factors Duration of Onset and the Course of Schizophrenia . Predictors of Short— vs. Longer—Term Outcome .  .  .  .  27 29 31 37 46 50 53 53 55 57 60 64 66 66 67 71  Current Vulnerability—Stress Models  73  Summary and Discussion of Literature Review  77  V  THE MAP PROJECT  IV. METHODS:  V.  83  Samples  83  Procedures  85  Measures  91  HYPOTHESES AND ANALYSES  103  Hypotheses  103  Analyses  104  Formal Statistical Considerations  113  VI. RESULTS  VII.  114  Phase 1.  Methodological and Descriptive Results  Phase 2.  The Independent Role of Predictor Variables  .  .  115 132  Building Models  156  Summary of Results  156  DISCUSSION  159  Methodology  159  Descriptive and Preliminary Results  166  Predictive Results  172  Models of Course and Outcome  190  Significance of the Project  191  Limitations  192  Future Research  195  VIII. REFERENCES  199  vi List of Tables  1.  Stressful Life Events, Onset and Relapse  13  2.  Social Relationships and Schizophrenia  23  3. Ventricle Size and Premorbid Adjustment  39  4.  Lateral Ventricle Size and Prognosis  43  5.  Ventricle Size and Affective Disorders  47  6.  Onset Characteristics and Outcome in Schizophrenia  7.  Summary of Measures  8.  Sample Size at Three Stages of Intake,  9.  Interview Schedule for Social Interaction:  10.  .  .  87 by Diagnosis  .  .  Scale Items  Summary of Proposed Analyses  Sources of Five-Year Follow-up Data Used for Analyses  93  .  95  117 119  .  13. Intercorrelations Among Social Relationship Measures, for Two Diagnostic Groups 14. Factor Analyses of Social Relationship Measures: Communality Estimates and Loadings For ‘Quality’ and ‘Quantity’ Factors for All Cases Combined and by Diagnosis 15. Sample Characteristics: Functioning by Diagnosis  .  105  11. Use of Rules in Selecting for Analyses Data From Multiple Sources Reflecting Five-Year Outcome 12.  69  .  123  .  125  Demographics and Baseline 129  16. Proportions of Patients in Three Categories of Adaptive Functioning Over Time, by Diagnosis  133  17. Correlations Between Five-Year Outcome and Demographic Characteristics and Baseline Functioning  134  18. Social Networks, Social Resources, Intake for Two Groups  and Perceived Support at 137  19. Correlations Between Five-year Outcome and Six ‘Ratings of Social Relationships, for Two Groups and All Cases Combined. 139 20. Hierarchical Regression: The Collective Role of Social Relationships in Predicting Five-year Outcome  141  vii 21. Hierarchical Regression: The Unique and Independent Role of Social Relationships in Predicting Five—year Outcome . . 144 22. Number of Life Events at Three Points in Time, by Diagnosis 23. Correlations Between Five-year Outcome and the Log of the Number of Life Events, by Time and Diagnosis .  148  .  .  150  .  24. Eye-Tracking Error and Ventricle-to-Brain Ratio, by Diagnosis 25. Correlations Between 18-Month Outcome and the Number of Life Events or Baseline Characteristics, By Diagnosis 26. Two Regression Equations: Stressful Life Events and Baseline Characteristics Predicting 18—Month Outcome Separately for Two Diagnostic Groups  153  .  .  176  178  viii  List of Figures  1. Nuechterlein’s (1987) Vulnerability—Stress Framework for Possible Factors in the Development of Schizophrenic Episodes  2.  Axis V Ratings at Intake and Five Years,  76  by Diagnosis  .  .  3. Smoothed Distribution of Eye—Tracking RMS Error Scores for 124 First—Episode Psychosis Patients  4. Axis V Ratings at Three Points 5.  in Time,  by Diagnosis  Life Events in Two Periods Prior to Onset,  .  by Diagnosis  131  154  168  .  .  184  ix  List of Appendices  1.  Summary of DSM—III Axis V  .  .  .  213  x ACKNOWLEDGEMENTS  Whereas the title page indicates a single author,  in fact  many individuals and groups have made important contributions to this dissertation and this degree.  Dianne E. Chappell has made  personal and professional contributions that will not cease with convocation.  Dr. Morton Beiser has inspired  learning and a love of life”, empirical research.  “  a love of  in providing an apprenticeship in  The expert and patient guidance of Dr.  Dimitri Papageorgis has been invaluable in the presentation of the ideas described herein,  as well as in the completion of a  challenging degree program. The MAP Project has been a team effort.  It has been a joy  to work with Jon Fleming, Janice Husted, Bill lacono, Kathy Keetley, Neil Kyle, Tsung—yi Lin, and more recently Grahm Bean and Jia-hui Zhang.  Four others helped with five-year data:  Val  Geddes, Richard Hsu, Elizabeth Kao and Janice McLauglin. The biggest collective contribution is from more than 300 patients and 200 members of their families, from the community.  and 200 volunteers  They gave freely of their time,  their  thoughts, and their feelings in order that staff of the MAP Project might make a small contribution in understanding schizophrenia. Finally, thanks to the taxpayers of Canada and the National Health and Research Development Program of Health and Welfare Canada for National Health Ph.D. Fellowship #6610-1816-47.  1  I.  INTRODUCTION  The prevailing paradigm for the etiology of schizophrenia invokes a diathesis-stress model, wherein the presence of both a predisposing biological vulnerability as well as some kind of psychosocial risk factor(s) must come together to account for the ‘causation’  of the disorder.  Gottesman & Shields,  1982;  Many prominent authors  Zubin & Spring,  (Crow,  1988;  1977) have described  conceptual models consistent with the diathesis-stress paradigm. Typically, however,  investigators gather evidence to support  factors from only the biological or the psychosocial realm.  The  results may be conceptually nested within a discussion of a diathesis-stress paradigm, but empirical investigations of etiology seldom bridge the chasm separating the biological and social traditions. Compared to investigations of etiology,  integrated  “biopsychosocial” models for course and outcome are even more rare. This paucity is curious:  such a model of outcome seems even  more compelling than one regarding etiology,  in light of the  weight of evidence regarding the efficacy of both psychosocial and somatic forms of intervention.  However compelling,  integrated, empirically—based models may not exist:  such in the  literature, there are no thorough assessments of the concurrent and interactive influence of biological and social factors affecting the course of schizophrenia.  2 Introduction  This dissertation attempts to provide just such a model of course and outcome.  From the biological domain two vulnerability  factors are examined, smooth pursuit eye movement dysfunction and brain ventricle size.  There is good evidence that this eye  movement dysfunction is under partial genetic control, and is probably specific to schizophrenia.  Large brain ventricles, on  the other hand, can be seen as a vulnerability factor that is neither familial nor specific.  To date, neither of these  biological factors has been adequately assessed as a predictor of the course of schizophrenia,  in a manner that examines the unique  and independent role of each. From the psychosocial domain,  I will examine a stress-process  model of the course of schizophrenia.  In particular, I will  assess the separate and combined influences of stressful life events and social support as they affect long—term prognosis.  No  assessment of such a model has been reported in the literature. Many authors have called for interactive models, both in schizophrenia in particular  (Nuechterlein,  1987;  Zubin & Spring,  1977) and in the prediction of behaviour in general 1978; Endler & Edwards,  1988).  In this project,  (Bern & Funder,  I will attempt to  examine not only the concurrent influence of social and biological factors, but also the interaction between some of these predictor variables. This dissertation is based on a large prospective study of the course of first-episode psychosis known as the MAP Project. Begun in 1981 by Drs. Morton Beiser and William Iacono, the study  3 Introduction is based on a representative sample of 175 patients, recruited shortly after they experienced their first lifetime episode of psychosis.  The participants were followed over the next five  years, with the last of the five—year follow-up interviews being conducted in 1989. groups: disorders  Most of the patients are in three diagnostic  schizophrenia, and either unipolar or bipolar affective Beyond  (both of which had psychotic features).  representative sampling, a number of methodological characteristics provide a solid foundation on which to construct an integrated, biopsychosocial model of the course of schizophrenia. The structure of this thesis is as follows.  First, the  specific aims are presented, with the rationale provided by a review of the literature.  Subsequently, the methodology of the  MAP Project is presented.  Next,  five hypotheses and related  statistical analyses will be outlined, presentation of the results.  followed by the  Finally, the discussion section will  consider the larger significance of the present findings.  4 II.  OBJECTIVES  The principal aim of this project is to propose an empiricallybased biopsychosocial model of the course of schizophrenia, using data from a prospective study of a large group of first—episode patients.  In particular,  I wish to examine the following  predictors of five—year outcome: 1. 2. 3. 4.  Stressful life events Supportive social relationships Brain lateral ventricle size (ventricle—to brain ratio; VBR) Smooth pursuit eye movement (SPEM) dysfunction  The objective is to evaluate the unique and independent role of each set of predictor variables.  To accomplish this aim,  prediction equations will control for the baseline measure of functioning.  Other control variables, such as age, sex,  socioeconomic status insidious vs.  (SES),  and the duration of onset (e.g.,  acute) will also be incorporated into the predictive  equations in order to control for the more obvious “third variable” artifacts.  These controls are necessary in order to  support possible causal interpretations of the prospective correlations. To the degree that each of the predictors is related to outcome, they can be considered concurrently.  Thus,  a second  objective is to consider a stress—process model of the course of schizophrenia, whereby supportive social relationships may mediate the stressor—distress relationship. A third objective is to construct a diathesis—stress model of the course of illness by considering concurrently life events,  5  Objectives  social relationships, and the two putative indices of biological liability, VBR and SPEM. year outcome,  To the degree that each predicts five-  one can assess the manner in which their  interactions affect the course of illness. The results from any empirical investigation are always more robust if replicated.  As will be elaborated below, publications  from the MAP Project have already demonstrated that supportive social relationships  (Erickson, Beiser,  lacono, Fleming, & Lin,  1989) and brain lateral ventricle size (Katsanis, Beiser,  1991) each predict short—term outcome.  lacono,  &  If the hypotheses  regarding the effect of these and other variables on longer-term outcome are supported, then the final objective of the replication of previous results may be fulfilled.  6  III.  BACKGROUND AND LITERATURE REVIEW  Conceptions of the course and outcome of schizophrenia have changed dramatically in the past 15 years.  The traditional view  had portrayed schizophrenia as a disorder with an inevitably debilitating outcome (Ciompi, 1970s, however,  Beginning in the late  investigations in North America (Harding, Brooks,  Ashikiga, Strauss, Carpenter,  1988).  & Brier,  1987; Prudo & Blum,  1977) and Europe  (Bleuler,  1987;  1978; Ciompi,  Strauss &  1980) have  demonstrated that the outcome of the disorder is not universally dismal.  Rather, under most diagnostic systems,  only about one—  third of afflicted individuals attain the state of chronic, perpetual impairment that resembles the traditional schizophrenic prototype.  At the opposite, good—outcome pole, up to one-third  of cases will experience only an occasional  subsequent psychotic  episode, with little inter—episodic impairment.  The remaining  third has been shown to have an intermediate outcome.  Thus,  for  many patients and their families in Western societies, the more recent view of course and outcome has provided hope. Results from cross-cultural studies have stimulated hypotheses about psychosocial factors that might be responsible for heterogeneity in outcome.  Affected individuals in developing, as  opposed to industrialized, countries have a more benign course (Murphy & Raman, [WHO]  1972,  1971; Waxler,  1979).  1977; World Health Organization  For example, people with schizophrenia in  India and Colombia had fewer and more brief relapses than those in Washington, Moscow, and Prague  (WHO,  1979).  These two sets of  7  Literature Review  findings--heterogeneity within and between cultures--point to sociocultural influences on course and outcome. At the same time, developments in technology have provided the tools for a revolution in the neurosciences, and in particular for the study of the etiology of schizophrenia. advent of computerized tomographic technology allowed vivo study of  ‘deep structures’ within the brain.  The  the j.  This kind of  viewing had not been previously available, and provided direct evidence of changes in brain morphology in schizophrenia. Parallel developments in psychophysiology, also due in part to computer technology, stimulated the search for genetic markers of schizophrenia.  Because the study of family aggregation  patterns had not been as fruitful as behavioural geneticists had hoped  (Gottesman, McGuffin,  & Farmer,  1987),  have pursued the study of genetic markers.  other investigators These genetically-  determined traits-—while not necessarily directly involved in disease etiology——could be used to  ‘triangulate’  the presence of  a vulnerability or predisposition in family members. By and large, the biological and psychosocial traditions have proceeded in parallel.  Generally,  investigators interested in  psychosocial aspects of schizophrenia have conducted research on the course of schizophrenia, while those interested in biological variables pursued etiology.  One reason for the “two solitudes”  is likely that investigators are simply not trained in both traditions.  The strength of these disciplinary boundaries can  serve only to impede progress in understanding the disorder and  Literature Review  8  providing aid to people with an often disabling and usually unpredictable condition. For at least two decades, the diathesis—stress paradigm has provided the conceptual integration of the two traditions. Usually associated with the work of Joseph Zubin (e.g., Spring,  1977),  Zubin &  the model assumes that varying degrees of  biologically—based vulnerability exist in a large number of people.  Among some portion of those individuals, an accumulation  of environmental stressors occurs, resulting in profound distress.  The form that distress takes is determined by the  biological vulnerability,  in this case schizophrenia.  This  diathesis-stress model has been applied in empirical investigations of the etiology, but not in the study of the course of schizophrenia. The literature review that follows will examine in detail some of the concepts described above. sphere,  In the psychosocial  for example, there are maior divergences in competing  notions of the key elements of social relationships.  As well,  there are several discrepant traditions for assessing the presence of stressors.  Unfortunately,  little prospective  evidence has accumulated on the psychosocial components of a diathesis—stress model of schizophrenia. Following the review of the psychosocial sphere,  I will  examine the literature pertaining to two biological variables that likely represent different etiological factors.  To the  degree that lateral ventricle size and smooth-pursuit eye movement dysfunction reflect liability for the initial onset of  Literature Review  schizophrenia,  9  they may also represent a continuum of liability  with respect to course and outcome of the disorder.  Social Factors in the Course of Schizophrenia One of the more prominent models in mental health,  the  stress—process model, brings together two sets of psychosocial factors to account for the etiology of distress.  As will be seen  below, the interaction of stressful life events and supportive social relationships has been repeatedly applied to the study of many domains of mental health.  Nonetheless, the relevant  literature on schizophrenia is relatively limited.  Stressful Life Events In the stress—process model, stressful life events  (LE5)  have been proposed as having an adverse impact, either to prompt the initial occurrence of distress, or to adversely affect the course of a disorder.  The notion is that the accumulation of  stressors at some point exceeds the individual’s capacity to deal with the multiple demands,  and at that point a person becomes  distressed. Definition and measurement of stressors.  Stress has been  used to describe both the cause of a decline in mental health, and the effect itself.  In the lay idiom, having “too much stress  in one’s life” can refer to the presence of unwanted and negative events, or it can refer to the distress that follows the  Literature Review  10  occurrence of those events. will be avoided.  Instead,  or other challenges, and  In this thesis, the term ‘stress’ ‘stressors’ will refer to life events  ‘distress’  to one’s subjective reaction  to the occurrence of life events. There are several concurrent traditions in the study of stressors.  One of these examines the effects of a single and  often catastrophic event, e.g., a natural disaster or injury following a motor vehicle accident  (Henderson & Brown,  1988).  A  second tradition examines the effects of ongoing strain on mental health, such as being physically disabled (Cf. Turner & McLean, 1989).  A third examines the effects of  described as  ‘daily hassles’  Most frequently, though,  ‘micro—events’, often  (e.g., Lazarus & Folkman,  1984).  investigators have assessed the  cumulative effects of more ordinary life events. effects of losing one’s job,  the birth of a child,  Here, the a change in  residence or the start of an intimate relationship, are assessed with respect to mental health.  In this tradition, studies have  repeatedly demonstrated that the accumulation of these stressful life events is associated with decrements in mental health (Finlay—Jones,  1988).  The measurement of life events typically occurs with the use of a structured checklist. Holmes and Rahe (1967)  The original instrument developed by  is relatively brief.  It includes  questions that ask about the occurrence of 34 LE5 in either the past 6 or 12 months.  Elaborations of this method often apply  linear weights in summing the LEs, which are based on the relative recency of each event, and on the subjective weighting  Literature Review  11  of the effect or the normative impact of each event (Thoits, 1983). A more exhaustive assessment of LEs has included a lengthy interview about each event Harris, & Birley,  1973).  (Brown & Harris,  1978;  Brown, Sklair,  Here, the participant tells the  detailed story of the occurrence of the event.  An interviewer  records the entire narrative, and then embarks upon an elaborate set of questions regarding each event, with reference to the extent to which it was related to current mood and was within the control of the respondent.  This protocol is used for the purpose  of interpreting the effects of stressors both at the cumulative level and at the level of the individual life event.  The choice  of this more thorough interview versus the more brief checklist (described above) depends on the objectives of the research investigation. In order to begin to argue for a causal role, the temporal priority of life events and symptoms must be established. nonchronic diseases,  In  it is usually clear whether life events  precede or follow the onset of symptoms.  In chronic or relapsing  disorders such as schizophrenia, disaggregating the events that are putative causes of symptoms from those that are the result of symptoms is accomplished by classifying events as independent vs. dependent, respectively. independent’,  A third category,  ‘possibly  is reserved for events that are not likely to have  been the result of pre-existing distress. Life events and the onset of schizophrenia.  Three studies  (summarized in Table 1) have assessed the role of LEs in  Literature Review  12  precipitating the initial onset of schizophrenia. Korten,  et al.  Day, Nielsen,  (1987) reported on a cross—national study of  first—admission patients in six countries.  In each site,  participants compared LEs in the three weeks prior to the first episode with those in previous three week periods.  In five of  the research centres, patients reported an excess of independent events,  i.e.  those outside of their control,  treatment contact.  Thus,  immediately prior to  it appears that an increased number of  LEs may have played a causal role in the onset, as opposed to being a result of early signs of illness. An earlier study by Jacobs and Myers  (1976) also examined  LE5 among first-admission schizophrenics compared to those reported by a matched, normal comparison group.  They inquired  about LE5 in the year prior to entry in the study.  They found  that schizophrenics had significantly more LE5 than did normals, although there was no significant difference when the events were restricted to those rated as independent.  Thus, this study  offers only equivocal support for a possible causal role for life events in the onset of schizophrenia. A third investigation, schizophrenics.  failed to find an excess of LEs among  Gureje and Adewumni  (1988) compared their first  episode group to a matched normal comparison group, and enquired about events in the previous six months.  There were no group  differences in the mean number of LEs, nor did the proportion of each group reporting at least one LE differ.  The negative  findings were also true for the three—month period prior to intake.  1. All LEs: schizophrenics > controls. 2. Independent LEa: no group differences. 1. No group diff. in mean number of LEa In either 3 or 6 month period prior to intake. 2. No difference between grps. in proportion who reported at least one LE.  Schiz. patients vs. matched controls. Schiz. patients vs. matched controls.  Jacobs & Meyers (1976)  Gujere & Adewumni (1988)  Relapsing patients have more Independent LEs in month prior to relapse, compared to: a) more remote periods in the study, and b) non-relapsing pta. in the same month. 1. Women only: pta. had more LEa than controls. 2. ClusterIng of LEa in 3 weeks prior to relapse only among married female patients. Relapse preceded by at least 1 LE in 56% of pta. In low EE homes, vs. 9% in high EE homes. 83% of relapsing pta. vs. 25% of pta. of nonrelapsing pts. have threatening LE.  “Recent onset” schiz. pta. interviewed every 8 wks. over 12 months Acute and chronic schiz. pta. in Saudi Arabia Unmedlc’d chronic schiz. pta. in high- vs. low—EE homes Medic’d chronic schiz. pta, all in high EE homes  Ventura et al. (1989)  Al Khani et al. (1986)  Leff et al. (1973)  Leff et al. (1983)  Life events and relapse  excess LEs in 3 wks. prior to 1. All sites: intake, compared to previous 3 wk. period. 2. More indep. LEs prior to intake in 5/6 sites.  Schiz. patients in six countries,  Results  Day et al. (1987)  Life events and onset  Sample  Stressful Life Events, Onset and Relapse  Author(s)  Table 1.  I(.‘)  m  14  Literature Review  Overall,  it seems that the key to detecting the hypothesized  role of life events in prompting the initial onset of schizophrenia may be to use a within-subjects design. et al.  (1987) used this strategy,  When Day  five of six centres found an  excess of independent events. Life events and relapse.  Four studies have provided direct  evidence to indicate that LEs play a role in relapse.  In a  prospective study, Ventura, Nuechterlein, Lukoff, and Pederson Hardesty (1989) described a group of recent-onset schizophrenic patients who had just finished a treatment program. the follow-up period,  Throughout  life events, symptoms and adaptive  functioning were documented.  Their results showed that relapsing  patients had more independent LEs in the month prior to the psychotic episodes, compared with non—relapsing patients.  The  same result was found when relapsing patients were used as their own controls:  There was an excess of LEs immediately prior to  the episode, compared to more remote periods. In an early study, Leff, Hirsch, Gaind, Rhode and Stevens (1973) studied a mixed group of acute and chronic patients, who were randomly assigned to long-term drug or placebo treatment. Their results showed that a larger number of patients who relapsed while on medication reported at least one life event in the previous five weeks, compared to relapsing patients on placebo.  The authors interpreted this to show that schizophrenic  patients who were on medication were “protected against the stresses implicit in uneventful social intercourse,  and were  15  Literature Review  unlikely to relapse unless exposed to some additional stress in the form of one or another life event” In a later study,  (p.  660).  the same research team (Leff, Kuipers,  Berkowitz, Vaughn, & Sturgeon,  1983) compared relapsing to  nonrelapsing patients, all of whom were living in high EE homes and were on medication.  Their results showed that life events  were more often present in the three weeks prior to relapse, compared either to previous periods or to nonrelapsing patients throughout the study.  The authors interpreted the results as  showing that acute stressors  (life events) had an additional  effect beyond that represented by chronic strain (high levels of expressed emotion). A fourth study also offers positive, but qualified, results. Al Khani,  Bebbington, Watson, and House  Arabian study,  (1986),  in a Saudi  found a higher frequency of LEs among a mixed  group of acute and chronic schizophrenics, compared to a matched normal comparison sample--but only among women.  They also found  a clustering of LEs in the three weeks prior to the index episode, but only among married female patients  (compared to the  corresponding normal control subgroup). Life events and course.  Host studies investigating LEs and  schizophrenia have examined the more-or-less immediate consequences of situational stressors.  That is, the role of LEs  has been considered as a precursor either to initial onset or to relapse, thereby taking a short-term view of outcome.  This usual  view of the role of LEs corresponds to a “triggering” model. That is, LEs precipitate onset or relapse, but they do not  Literature Review  16  substantially alter the probability that it will happen--only its In their model, Brown and Birley (1968)  timing.  even developed  formulae to estimate the degree to which LE5 bring forward in time the occurrence of a psychotic episode. A different model might test the effects of LEs on the cumulative course of illness.  If it is true that LEs affect the  timing of a psychotic episode, then a cumulative measure of course may show adverse effects.  For example, an episode may  begin sooner because of LES, but the timing of its resolution may be unaffected—-resulting in longer episodes.  Alternatively,  to  the degree that LEs “bring forward” the timing of an episode, but do not affect its length, the duration of symptom—free interepisodic periods would be shorter.  In either case,  schizophrenics with a relative abundance of LE5 should show a poorer course.  In this way,  life events assume the status of a  hazard, and not simply a precipitant.  This model has not yet  been empirically tested*. Limitations in the LE area.  There are two major issues that  are common to all of the LE studies, but they have yet to be addressed——either conceptually or empirically.  The first issue  relates to the frequent exclusion of slow-onset patients: studies  (Brown & Birley,  Adewumni,  1968; Day et al.,  1988; .Jacobs & Myers,  1987;  most  Gureje &  1976; Ventura et al.,  1989)  * Only recently has there has been an investigation of the Hirsch, cumulative effects of LES over a longer period of time. Bowen, Emami, Cramer, Jolly, and Haw (1993) found that LE5 acted cumulatively to increase the risk of relapse over a 12-month follow-up period. To date, their work has been reported only in abstract form.  Literature Review  17  included only those patients whose onset occurred within 3-6 months of referral to the study.  If it is true that acute (vs.  insidious) onset and the presence of precipitating events cluster together (Dohrenwend & Egri,  1981), then the practice of  excluding slow-onset patients may render the apparently substantive findings exclusion criterion.  (outlined above) an artifact of this Only a study that includes the full range  of insidious and acute onset patients would be able to resolve this issue. A second significant issue yet to be addressed relates to the conceptualization and measurement of the onset period.  In  it is now a fundamental requirement that  the LE literature,  events must be dated with respect to the onset of psychotic symptoms  (Dohrenwend & Egri,  1981;  Spring,  1981),  in order to  separate LEs that may cause onset from those that may be the result of the disorder.  However,  one further step is necessary.  There remains the possibility that some LE5 are the result of prodromal symptoms.  Such a possibility is particularly cogent  where the onset of the prodrome is not accompanied by profound occupational or social impairment.  For example, a working parent  may experience several weeks or even months of social withdrawal and flat affect.  Even without florid psychotic symptoms, one  would expect LBs that follow the onset of the prodromal period, such as reprimands in the workplace or arguments with a spouse. Thus, a proper account of onset must include the prodromal period.  Literature Review  18  Finally, studies of life events tend to be based on naturalistic,  longitudinal research designs.  Here, assertions  regarding causality rely on prospective correlations or temporally—based comparisons (e.g., the number of life events immediately before relapse vs.  other time periods).  However, the  definitive standards by which causality is demonstrated require evidence beyond prospective correlations, even if obvious confounding factors are controlled.  Other evidence might be  derived from the experimental manipulation of stressors to examine resultant distress.  In general, ethical requirements  prohibit the collection of this kind of evidence. Summary.  There is good evidence that an increase in the  number of life events may play a causal role in the onset of both initial and subsequent episodes.  In this triggering model, more  work is needed, based on research designs that include a broad sample of patients and an adequate assessment of the duration of onset.  Similar conditions are needed to assess an alternate  model that addresses LEs with respect to course and outcome, wherein the influence of possible mediating variables such as social relationships may be examined.  Social Relationships A vast literature has accumulated on the effects of social relationships in a wide array of physical and psychological disorders.  In physical medicine, the presence of supportive  others leads to increased survival rate after the detection of breast cancer and the occurrence of myocardial infarct  (Medalie &  19  Literature Review  Goldbourt,  1976),  chronic pain  reductions in the severity of arthritis and  (Davidson, Bowden,  & Tholen,  1979; Porrit,  1979),  and decreases in the rate of pregnancy and birth complications (Nuckolls, Cassel,  & Kaplan,  1972).  In the mental health sphere, greater amounts of supportive relationships have been shown to have a protective function: Among people experiencing stressful life events, rates of depressive and anxiety symptoms are substantially lower for those with more social support  (Mueller,  1980).  While there is an  impressive body of evidence for the role of social relationships in preventing the occurrence or mitigating the effects of anxiety and mood disorders  (Alloway & Bebbington,  1987), there is  relatively little empirical work on the role of social support in schizophrenia.  Before reviewing that evidence, however, some  clarification of the terminology and concepts is required. The “social support” literature suffers from problems of definition.  While most authors share the perspective that there  is something about social relationships that is valuable in terms of resistance to or recovery from distress, agreement stops there.  There is little consensus about the concept itself, and  even less about the critical elements of social relationships that are thought to be helpful.  Most approaches to the study of  social relationships fall into three categories:  social  networks, social resources, and the perception of social support. The social network approach involves the study of the structural characteristics of a person’s social relationships. The most obvious is the number of people named to the network  Literature Review  20  (Greenblatt, Becerra, & Serafetinides,  1982).  The composition of  the network is also associated with one’s mental health: presence of a confidant, often a spouse, & DeLongis,  the  is most important (Coyne  In general, greater numbers of both kin and  1986).  non-kin are important to the preservation of good mental health (Schradle & Dougher,  1985),  compared to acquaintances  as is the presence of close friends,  (Henderson,  1988).  The network  approach, then, has shown that a number of structural characteristics are related to good mental health. The transactional approach to social resources focuses only on those types of relationships that fulfill certain functions. Thus, theorists have tried to categorize the nature or the quality of the transactions that comprise social relationships, and focus on those that are deemed to enhance health status.  For  example, Caplan (1974) emphasizes mutuality and reciprocity in each of emotional support, cognitive support, and tangible assistance.  Cobb  (1976),  on the other hand, argues that the  essential feature is the provision of informational feedback. Henderson (1984,  1988) offers a more complex taxonomy based on  attachment theory (Weiss,  1974).  Turner (1983) argues that the transactional approach confounds the presence of social relationships with the perception of being supported.  In his view, social support is always positive:  It  is the degree to which one is satisfied with the current social resources.  Since the experience of being supported is also  affected by other factors, he argues that controlling for  Literature Review  21  constructs such as mastery and locus of control successfully captures the perception of social support per Se. There is no consensus in the area on the definition of healthenhancing interpersonal transactions.  Empirically, there is no  reason to select as best either the network, the transactional, or the perception-based approach.  The best approach appears to  be an inclusive one that incorporates the network, transactional, and perception-based elements.  For example,  it must recognize  the distinction between the support arising from a spouse, good friends, and acquaintances.  Moreover, an inclusive measure  should assess emotional support, cognitive or informational support, and some index of tangible assistance.  If data based on  all of these considerations are collected, then supportive social relationships have been well—measured. Social relationships and schizophrenia.  While the literature  on supportive social relationships and emotional health in general is substantial, that relating to schizophrenia is small. Most authors who have studied the psychotic disorders have stated that their ultimate interest is in the role of social relationships as they influence the course of the disorder, with possible implications for treatment.  In spite of this stated  intention, all but two papers have been based on a cross sectional design:  unable to comment on the direction of  causality, most of the published reports can provide no clear implications for treatment. starting point.  Nonetheless, they do provide a  Literature Review  22  Host investigations in the area (summarized in Table 2) have taken the network approach, based on cross—sectional research designs. Fleming,  Here, a number of studies & Lin,  Wood, Frazier, Geiger,  1989; Garrison, & Crowder,  (Erickson, Beiser,  lacono,  1978; Pattison, DeFrancisco,  1975; Sokolovsky,  1978; Westermeyer & Pattison,  Cohen, Berger,  &  1981) have reported that  the networks of schizophrenic persons are smaller than those of non-psychiatric comparison groups.  This reduced network size is  probably not specific to schizophrenia:  in an earlier report  from the MAP Project, we reported that a comparison group of affective psychotics also had smaller networks,  although this  group occupied a median position between the schizophrenic and normal groups  (Erickson et al.,  1989).  Evidence also suggests that the networks of schizophrenic patients become further reduced over the course of illness (Westermeyer & Pattison, (1978) showed that,  1981).  For example,  Sokolovsky et al.  over a two—year span, network size diminished  by 40%. Schizophrenics also appear to differ in the structural characteristics of their social networks.  One of the most  widely—cited findings is that family members are over represented, compared to the networks of both psychiatric and normal comparison groups  (Pattison et al.,  1975; Tolsdorf,  1976).  We reported that the relative over—representation of kin is not due to an increase in their absolute magnitude in the network, but rather to the decreased presence of friends and acquaintances (Erickson et al.,  1989).  In summary, most studies of social  1. Size: no significant difference. controls > schiz. 2. Multiplexity: kin in network: Schiz > controls. 3.  1. Networks of normative sample are larger in number and greater in multiplexity. 2. Schiz. pts. have one good friend at most.  Network characteristics: a) size--- normal controls> asymptomatic schiz. pts. > schiz. with residual sx’s. b) multiplexity—— same rank order as above but nonsignificant trend. network measures predicted Outcome: rehosp’zn, but only for mildly impaired. Psychotic villagers have smaller networks, & their social relationships are made up of smaller and fewer ‘clusters’. At intake: controls > aff.> schiz. a) Kin in network: b) Availability of support from confidants: schiz. < affectives, controls. C) Adequacy of support from acquaintances: schiz. < affectives, controls. Predicting outcome: a) More nonkin predicts good outcome (all pts.). more supportive acquaint— b) For schiz. pts.: ances predicts better outcome, but more kin in network predicts poorer outcome.  10 first—admission schiz. pts. vs. 10 medical controls.  Puerto Rican women in NYC: schiz. pts. vs. controls,  Chronic schiz. pts. in hotel, vs. occupants with no psych history.  schiz. vs. Laotian peasants: affective psychosis vs. normal controls First—episode schiz. vs. affective psychosis vs. normal controls.  Garrison (1978)  Sokolovsky et al. (1978); Cohen & Sokolovsky (1981)  Westermeyer & Pattlson (1981)  Erickson et al. (1989)  Tolsdorf  (1976)  Pattison et al. (1975)  normals > neurotics > 1. Network size: psychotics (mostly family).  Results  Neurotic vs. psychotic pts. vs. normative community sample.  Sample  Social Relationships and Schizophrenia  Author(s)  Table 2.  w  lb  lb  (b  Literature Review  24  relationships and schizophrenia have used a cross—sectional approach, where authors found differences between schizophrenic and normal comparison groups. Only two papers have directly assessed the role of social relationships on the course of the disorder. reports  (Cohen & Sokolovsky,  One of those  1978) concluded that, among chronic  patients, there is evidence that social relationships do have a positive effect on outcome but only among the mildly impaired. Among more severely impaired chronic patients, social relationships did not influence outcome. examined first-episode patients:  The second paper  Erickson et al.  (1989),  describing the MAP Project sample, reported that more numerous social relationships outside the family appear to have a positive effect on course.  By contrast, the presence of greater numbers  of family predicted a poorer short-term course for schizophrenics.  A somewhat different picture was true for the  affective psychotic patients:  greater numbers of both kin and  nonkin had a positive effect on short-term outcome.  Stress-Process Models If research documenting the unique and independent effects of stressful life events and supportive social relationships on mental health is the first step in building a stress—process model, then assessing their simultaneous effects is the next step.  That second step, building a two—factor stress-process  model, has been applied repeatedly to the study of many physical disorders  (Cohen & Syme,  1985).  In the mental health sphere, the  25  Literature Review  model has been validated in the study of clinical populations with affective and anxiety disorders, as well as in epidemiological studies of nonclinical populations (Mueller, 1980).  In each case, two kinds of effects are noted:  (1) a main  effect, where the presence of supportive social relationships has been associated with reductions in the number of LE5; and (2) an interactive or buffering effect, where social relationships are associated with diminished effects of LEs when they do happen. A three—factor stress—process model has also been applied to the study of many disorders and conditions.  Here, the concept of  coping is added to the consideration of social relationships as an additional intervening factor in the stressor—distress relationship (Lazarus & Folkman,  1984).  Both the stress-process model and the expanded stress-andcoping model have been the subject of empirical investigation for at least 20 years.  Despite hundreds of reports about dozens of  physical and mental conditions, neither model has ever resulted in a published report about schizophrenia.  Neither model has  been applied empirically to either the etiology or the course of schizophrenia.  The apparent absence in the literature of these  models is indeed curious.  It may be that empirical  investigations of these models vis-a—vis schizophrenia have come to nought, and the negative findings have simply not been published. In sum, the best information available on the stress-process model as applied to the psychotic disorders Involves separate and parallel results about its components.  Regarding the first  26  Literature Review  component, the stressors,  investigators of schizophrenia have  used the cumulative index of life events in most published work. There, the objective has been to document the proximal association between an increase in the number of stressors and either initial onset or a subsequent relapse.  No report has  assessed the long—term, cumulative effect on course. In the study of other domains of mental health the second component of the stress—process model, supportive social relationships, has frequently been shown to have a preventative and a buffering role in the face of stressors.  In schizophrenia,  however, most of the empirical reports of supportive social relationships have been limited to findings based on crosssectional research designs.  Here, the results do suggest that  schizophrenic people have smaller social networks,  fewer social  resources characterized by supportive transactions, and generally feel less supported than other people.  In examining the effect  of supportive social relationships on the course of schizophrenia, two longitudinal studies have shown that social relationships influence outcome. Even so, differences in samples and measures mean that replication is still needed.  Literature Review  27  Biological Factors:  I. Lateral Ventricle Size  In 1976, Eve Johnstone and her colleagues at the Institute of Psychiatry in London first described the use of a new X-ray imaging technique to study in vivo the structure of the brains of schizophrenic people (Johnstone, Crow, Frith, Husband, 1976).  & Kreel,  They found that, compared to a medical control group,  schizophrenics’  lateral ventricles were moderately enlarged.  Since that first study, the cumulative results from more than 100 reports using computerized tomography (CT) have provided in vivo evidence that schizophrenia is very likely,  in large part, a  ‘primary brain disease’. The lateral ventricles, also known as the first and second ventricles, are symmetric fluid-filled cavities on each side of the brain.  Surrounding these cavities are midbrain structures  such as the thalamus, hypothalamus, corpus callosum.  hippocampus,  fornix, and  Investigators hoped that any ventricular  enlargement detected on a CT film would be a direct clue to pathology in surrounding tissue.  Previously,  other lines of  investigation had found evidence of pathology in these structures.  Never before, however, had a technique been  available that would rule out the artifactual explanations of the previous results by offering in vivo evidence. Most of the published work has been devoted to establishing the presence of ventricular enlargement, with the aim of investigating the etiology of schizophrenia.  As will be seen  below, many technological and methodological issues have troubled  Literature Review  28  the area, but most authorities now believe that ventricular enlargement is present among at least a substantial portion of schizophrenics, and that it occurs at or before the onset of the first psychotic episode.  The magnitude and prevalence of  enlargement, as will be seen, are issues that await definitive When taken In the aggregate, though, there is  resolution.  widespread agreement that lateral ventricular enlargement ([NE) is related to etiology. Enlarged ventricles may reflect a relative prominence of nonfamilial etiological influences in schizophrenia. of the  ‘sporadic’ vs.  ‘familial’ distinction,  In support  four studies have  documented an association between increased VBR and an absence of family history of schizophrenia (Pearlson, Garbacz, Moberg, Ahn, & DePaulo,  1985; Reveley, Reveley, Clifford, & Murray,  1982;  Sacchetti, Caizaroni, Vita, Terzi, Pollastro, & Cazzullo, 1992; Turner, Toone, & Brett-Jones,  1986).  Other studies, however,  have failed to find this relationship (Farmer, Jackson, McGuffin, & Storey,  1987; Reddy, Mukherjee, Schnur, Chin,  & Degreef,  1990),  and one has documented an association in the opposite direction (Owen, Lewis,  & Murray,  1989).  Because of several limitations, the significance of LyE for the course of illness is not clear because,  for the most part,  possibly overlapping relationships between ventricle size and premorbid functioning, clinical state characteristics, and other indicators of biological vulnerability have not been controlled. To the degree that these other correlates of ventricle size are  Literature Review  29  examined, the significance of increased ventricle size on outcome may become clearer.  CT Imaging and Ventricle Size:  General Issues  Prior to the development of CT technology,  in vivo imaging  had been done with pneumoencephalography, whereby air was actually injected into the head via the lumbar subarachnoid space.  X-rays of the head would then allow imaging of the  ventricular system and cortex.  Between 1920 and 1960, many  investigations had reported enlargement of ventricles and atrophy of cortical tissue in schizophrenia (Haug,  1962).  Since the middle of the 1970s, the technology involved in computerized tomographic imaging has been the basis of an unprecedented advance in the study of schizophrenia.  The new  technique has offered a noninvasive in vivo view of brain anatomy, methods  free of artifact that had been associated with previous (Seidman,  1983).  While CT technology has provided a tremendous impetus to research into the etiology of schizophrenia,  its major limitation  is that CT data index relatively gross anatomy in the brain. This is primarily a function of the limited powers of resolution in the image.  For the past decade there has truly been a  revolution occurring in the neurosciences, but the revolution has not yet advanced to the stage of mapping in sufficient detail the brain structures--and their interconnections--that are likely anomalous in  schizophrenia.  Advances in technology that allow  in vivo study of fibres and tracts are rapidly developing,  e.g.,  Literature Review  30  new structural imaging  (magnetic resonance imaging; MRI), as well  as functional mapping (e.g., positron emission tomography [PET], regional cerebral blood flow [rCBF], dynamic MRI)*.  Overall,  the  limited resolution of the CT image puts a similar limit on the conceptual understanding of the specific mechanism(s) Measurement issues.  involved.  The changes in ventricle size that are  hypothesized to occur in schizophrenia are subtle, and require special attention to measurement and methodology.  Unfortunately,  many of the earlier studies were based on simple linear measurements, where the ventricle was measured at its widest point, and compared to the maximum width of the total brain. Methodological studies  (e.g., Penn,  Belanger, & Yasnoff,  1978)  soon showed low correlations between this one—dimensional approach and the actual volume of the ventricles.  A better  alternative is the measurement of area using a planimeter, whereby the outlines of the ventricles and the brain are traced. The area of the ventricle is then calculated as a percentage of the total brain area.  This ventricle-to-brain ratio (VBR)  sensitive to small differences, and correlates well three-dimensional assessments of total volume  is  (r>.95) with  (Penn et al.,  1978). Ventricle size and diagnostic specificity.  Enlargement of  the ventricles can be the result of many causes other than schizophrenia, and the implications of enlargement are not necessarily clear.  For example, LVE can be permanent,  following  * The new technologies have superseded the CT approach, but CT based data are still current with respect to the course of schizophrenia.  Literature Review  31  blockage of the flow of cerebrospinal fluid (CSF; TerBrugge & Rao,  or degeneration of brain tissue (Bird,  1983)  1982).  It can  also be reversible, as is the case with a marked change in electrolyte balance or poor nutritional status Winter,  & Wilson,  1978).  Dilation of a ventricle can be  for example following trauma,  localized,  inflammation, or  vascular hemorrhage, or it can be diffuse. case in schizophrenia (Seidman, disease (Bird, & Holgate,  (Benston, Reza,  The latter is the  1983), but also in Alzheimer’s  1982) and alcoholism (Carlen, Wilkinson, Wortzman,  1984).  Ventricular enlargement does not necessarily imply tissue atrophy.  For example,  in Alzheimer’s disease, a change in  ventricle size is invariably accompanied by tissue atrophy. Contrarily,  jp37  in Huntington’s disease and hydrocephalus,  ventricular enlargement occurs without atrophy (TerBrugge & Rao, 1983).  A third possibility is that brain tissue can be affected  by some conditions, ventricle size.  e.g., cerebral anoxia, without a change in  In summary, any change in ventricle size is not  specific to schizophrenia,  and does not necessarily imply atrophy  in brain tissue.  Lateral Ventricles in Schizophrenia Several measures of brain structure have been studied in schizophrenia.  The size of the third ventricle, the cortical  sulci, and the cerebellum have been examined in addition to the size of the lateral ventricles.  Only the lateral ventricles have  Literature Review  32  been selected for examination in this study,  for the following  reasons: ——  ——  -—  --  -—  due to the limited resolution of the CT image, structures are best suited for study  large  there is better agreement on measurement issues in lateral ventricles, as compared to measurement of third ventricles, sulci and tissue atrophy lateral ventricles are the most frequently studied, and hence best understood limited evidence from previous studies suggests a link between LV size and prognosis, unlike other measures of brain structure two other parts of brain structure, third ventricles and measures of cortical tissue atrophy, are small, resulting in limited variance, which in turn can be expected to attenuate any predictive relationship  For these reasons,  lateral ventricle size was the best  available index of changes in brain structure, at the time of the inception of the MAP Project. Group differences in ventricle size.  Raz and Raz (1990)  recently conducted a quantitative review of the more than 100 studies that have examined the CT scans of schizophrenic people. Regarding lateral ventricles, the authors found a mean effect size of 0.70 and controls.  (SD=.54), based on 53 comparisons between patients This effect size corresponds to a 5Th overlap  between the two groups. In their review, Raz and Raz  (1990) classified five of the  53 comparisons as outliers, because the effect sizes were large and discontinuous with those derived from the other 48 comparisons.  The distribution of 48 studies formed a smooth and  normal curve, suggesting that lateral ventricle size is not  Literature Review  33  bimodally distributed for schizophrenic and  ‘normal’  people.  Removal of the five outliers reduced the effect size to M=.57, which corresponds to a 63 overlap.  In Cohen’s  (1977)  nomenclature, this is roughly midway between a moderate (d=.70) and a small  (d=.40) effect size.  The method of measurement seemed to make a difference in the findings.  Where two-dimensional planimetry was used, the mean  effect size was significantly larger measures  (M=.43).  (M=.76) than with linear  This difference is even more dramatic if the  outliers are removed  (.63 vs  .23, respectively; Raz & Raz,  1990),  although the magnitude of the mean effect size is reduced.  Thus,  the best estimate from the meta—analysis would appear to be based on area measures with the  ‘outlier’  studies removed from  consideration: the effect size from such studies would appear to be approximately 0.60. Timing of ventricular enlargement.  Two strategies have been  used to examine the timing of enlargement, with the purpose of determining whether ventricular enlargement is a possible cause or result of the disorder.  The first examines ventricle size  among schizophrenic people at the time of their first lifetime episode of psychosis,  or very early in the course of illness.  Here, Weinberger, DeLisi, Perman, Targum, and Wyatt (1982) found that schizophreniform and chronic schizophrenic patients showed comparable lateral ventricle enlargement, compared to controls. A second study, comparing teenage schizophrenic and schizophreniform patients to both a psychiatric and to a normal  Literature Review  control group, al.,  34  found similar results  (Schulz, Koller, Kishore,  et  1983). A third report, however,  results:  failed to find the expected  In the MAP Project, schizophreniform and schizophrenic  patients were compared both to medical and normal control groups Smith, Moreau et al.,  (lacono,  1988;  Smith,  Unlike the  1986).  two other studies of early—phase patients, no significant group differences in lateral ventricle size were detected.  In fact,  the ventricle size in the two patient and two control groups was remarkably similar  (mean VBR ±. s.d.:  schizophrenics 6.7 ±.. 2.6;  schizophreniform 5.9 t 2.4; normal controls 6.4  2.8; medical  controls 5.8 + 2.6). This inconsistency is hard to resolve.  The negative  findings of Smith (1986) were not due to a lack of statistical power.  One possible explanation relates to the nature of the  sample:  Many authors interpret findings  in the field as  suggesting that a portion (but not all) of schizophrenic patients have LyE.  The MAP Project is the first to report on a reasonably  representative sample--which may include a much larger proportion of normal  (small) ventricle patients than have studies which draw  on hospital populations.  Without a representative sample, the  good-prognosis patients who are not currently in treatment and who are more likely to have normal—sized ventricles, are underrepresented.  If this is true,  then studies that recruit  patients from treatment facilities will derive larger mean ventricle sizes than studies based on  representative samples.  Literature Review  35  The second strategy regarding the timing of ventricle enlargement examines the stability of ventricle size over time to see if LVE is progressive.  Here, three groups of investigators  have re-examined schizophrenic patients either 2—5 or 7—9 years after their initial scan.  All found no change in cerebral  ventricular size, whether in chronic, continuously ill patients (Illowsky, Juliano, Bigelow, phase patients  & Weinberger,  (Sponheim, lacono,  Valvassori, & Cazzullo,  1988).  1988) or young, early-  & Beiser,  1991; Vita, Saccetti,  Together, the results from the  two strategies indicate that ventricular enlargement is present among at least some schizophrenics at onset, and is likely stable after the first psychotic episode. Magnitude of enlargement.  Several lines of evidence provide  estimates of the magnitude of enlargement.  The first is based on  the meta—analysis of Raz and Raz (1990), who found an effect size of approximately .60.  Based on a normal distribution with a mean  of six and a standard deviation of three VBR units, the typical enlargement will be about 1.8 VBR units. in another review, was based on raw data:  A second approach, used Smith (1986)  found  that the median difference between the mean of schizophrenic and control group ventricles was 1.7 VBR units  (SD=0.7).  Twin studies provide a third source of data for estimating the degree of enlargement. monozygotic  The ventricle size of normal  (HZ) twins Is virtually identical  (r=.97),  and  healthy dizygotic (DZ) twins showed substantially less similarity (r=.35; Reveley et al.,  1982; Suddath, Christison, Torrey,  Casanova,  1990).  & Weinberger,  Thus,  any difference in ventricle  Literature Review  36  size in MZ twins discordant for schizophrenia is probably an excellent estimate of LyE. In their magnetic resonance imaging (MRI) al.  (1990)  study, Suddath et  found that ventricles of the affected twin were  approximately 15-2O larger than those of the unaffected twin. In another twin study, the mean intrapair difference estimates for discordant and healthy MZ twin pairs were 2.16 and units, respectively (Reveley et al., between these coefficients  1982).  .36 VBR  The difference  (1.8 VBR units) can be taken as an  indication of the degree of enlargement.  Taken together, the  meta-analysis, the estimates from raw data,  and the twin studies  all derive approximately similar estimates of the magnitude of enlargement. Prevalence.  The nature of the distributions of lateral  ventricle size limit the ability to estimate the prevalence of LVE.  The substantial overlap in the distributions of LV size of  schizophrenics and control groups contributes to widely differing estimates of the prevalence of LV enlargement. There are several definitions of enlargement.  The most  frequently used definition is the mean plus two standard deviations, based on a control group within the same study (Pfefferbaum, colleagues  Zipursky, Lim,  et al.,  1988).  Weinberger and his  (Weinberger, DeLisi, Neophytides, & Wyatt,  1979) have  suggested that the cutoff be set at 10 VBR units--anything larger would be considered an enlarged ventricle.  A third method of  defining large and small ventricles is simply to divide a sample at the median,  i.e. the “median split” technique.  Literature Review  However,  37  in light of the absence of a bimodal distribution  (Daniel & Weinberger,  1991),  it is difficult to assert  unequivocally that any particular cut—off point be used (Crow, 1988).  Psychometrically,  as a continuous variable.  it would appear that VBR is best used Without a rational basis for the  selection of any specific cut—off point,  it seems impossible to  determine a meaningful estimate of prevalence of ventricular enlargement. There is, however, one current exception to this rule.  A  reasonably good estimate can be made in the case of monozygotic (HZ) twins who are discordant for schizophrenia, because the ventricles of healthy HZ twins show an extraordinarily high degree of similarity in size.  Here, the combined results of two  studies of discordant HZ twins found that the affected twin had larger ventricles in 19 of 22 Suddath et al.,  1990).  (86%) cases  (Reveley et al.,  1982;  Unfortunately, no equivalent estimates  are currently possible on the prevalence of LyE in broader samples of schizophrenics.  Correlates of Ventricle Size In an attempt to understand better the significance of LVE in the course of schizophrenia, research workers have studied a number of its correlates.  They include premorbid functioning,  clinical status, treatment response, [NE and Dremorbid adjustment.  and prognosis. If large ventricles are  associated both with poorer outcome and poorer premorbid adjustment, then any prognostic liability represented by LVE may  Literature Review  38  simply be due to deficits in premorbid ad:iustment.  Thus,  in  order to demonstrate any liability of large ventricles on course and outcome, one must control for the relationship between ventricle size and premorbid adiustment. As displayed in Table 3, most of the studies that examined this relationship were based on chronic patients. however,  Two studies,  represented a wide range of severity and functioning  among the chronic samples  (Pearlson et al.,  Reveley, Kolokowska, Ardern,  & Mandeibrote,  1985; Williams, 1985).  Both found  that large ventricles were associated with poor premorbid functioning, which included the presence of schizoid traits in childhood and adolescence.  Of the three studies comprised of  samples with a restricted range of chronic patients, two detected a relationship between poor premorbid functioning and larger ventricles  (Jeste, Kleinman, Potkin, Luchins,  & Weinberger,  Pandurangi, Bilder, Rieder, Mukherjee, & Hamer,  1982;  1988) and one  reported a trend in the opposite direction (Nasrallah et al., 1983).  The sixth study of schizophrenic siblings found no VBR  premorbid relationship (DeLisi, Goldin, Hamovit,  et al.,  1986).  In summary, about two—thirds of studies of LV size and premorbid status have documented a significant relationship, where larger ventricles have been associated with poorer premorbid status.  That LV size and premorbid status are  Schizoid traits more often present among patients with large ventricles.  Correlation between premprbld Index and VBR: r=.40 (larger ventricles = poorer functioning). Correlation between VBR and: a) total premorbid score: r=.13 (n.s.), b) social functioning: r=.34 (p.OG). No significant correlation between VBR and either of 2 premorbid scores.  Chronic schiz. patients (broad range of functioning)  Schiz. patients living in the community.  Chronic schizophrenics.  Schiz. siblings (26 patients from 12 families),  Chronic schizophrenic men.  Pearison et al. (1985)  Pandurangi et al. (1988)  DeLisi et al. (1986)  Nasrallah et al. (1983)  Authors suggest no premorbid differences in VBR groups, although re—analysis suggests otherwise (see text).  2.  Large ventricle group had poorer childhood adjustment. No relationship between adolescent premorbid ratings and ventricle size.  Williams et al. (1985)  1.  Chronic, treatment—refractory schizophrenic in-patients,  Results  Weinberger et al. (1980); Jeste et al. (1982)  Sample  Ventricle Size and Premorbid Adjustment  Author(s)  Table 3.  ‘.0  w  m  1  m  I-.  Literature Review  40  associated is consistent with the notion that ventricular enlargement occurs prior to the onset of schizophrenia, but only in some portion of afflicted people.  An alternative explanation  is that premorbid status is a mediating variable in etiology, e.g., people with poor premorbid functioning are more prone to ventricular enlargement.  Either way, the implication for  research projects that attempt to predict course and outcome is that premorbid adjustment should be included as a control variable. Severity of illness.  In order to interpret a relationship  between ventricle size and prognosis as possibly causal, one must also assess any concurrent presence of [NE with overall severity of illness. After all,  it may be that VBR simply reflects a more  severe form of illness. Unfortunately, data are sparse.  DeLisi et al.  (1986)  examined the relationship between VBR and a global measure of severity.  Their sample consisted of 26 schizophrenic siblings  from 12 families; other sample characteristics were not reported. The results from this study revealed a nonsignificant trend between VBR and an overall severity rating (r=.33,  p<.1O). Again,  the lack of description of the sample makes the findings difficult to interpret. In a meta-analysis, Raz and Raz (1990) assessed the hypothesis that severe forms of illness would have larger ventricles.  After controlling for patients’ age and length of  illness, the cumulative length of hospitalization accounted for 21% of the variance (partial  =.46)  in the effect size of  Literature Review  41  studies that compared the size of lateral ventricles in schizophrenics and control groups.  In other words, severity of  illness is associated with lateral ventricle size. In pursuing the severity-VBR association, much of the work has been directed at two more specific correlates:  The  positive/negative syndromes, and treatment response.  Regarding  the former, several overlapping but distinct concepts have been proposed,  including the Type I vs.  II distinction of Crow (1980),  and the deficit syndrome of Carpenter and Strauss Carpenter, Heinrichs, & Wagman,  1988).  (1987) has described three factors:  (e.g.,  More recently, Liddle  positive symptoms, a  disorganization cluster, and the usual negative symptom grouping  *  In specifying the nature of the severity, most investigators have found hypothesized correlations between large VBR and an increased number of negative symptoms et al.,  1982; Kemali, Mai, Galderisi, et al.,  Johnstone, Crow, et al., DePaulo, et al.,  (Andreasen, Olsen, Dennert,  1984; Takahashi,  1985,  1987; Owens,  1985; Pearison, Garbacz, Breakey, Ahn, manage, Kato, et al.,  &  1981; Williams  1985), although several studies have failed to detect  such an association (Nasrallah, Kuperman, Hamra, Whitters,  1983; Pandurangi et al.,  1988).  & McKalley—  The relationship  between increased ventricle size and negative symptoms has also been noted in most MRI studies (e.g., Andreasen, Nasrallah, Dunn, * Whatever the particular concept, it is important to recall that these are clusters of characteristics, akin to the notion of ‘fuzzy sets’, rather than discrete subgroups of patients who each have all of the characteristics.  Literature Review  et al.,  1986; Andreasen, Erhardt,  Besson, Cherryman, Scwartzkopf, al.,  42  et al,  Swayze, et al.,  1990; Besson,  1987; Olsen, Nasrallah, Coffman, &  1990; although not in Mathew, Parein, Prakesh, et  1985).  Treatment response.  If negative symptoms and large ventricles  are indeed associated, this begins to resemble Crow’s Type II subtype, where one characteristic is poor response to medication. A number of studies have prospectively assessed VBR and response to neuroleptic medication.  prospective studies have  In general,  found that patients with large ventricles show poorer treatment response  (Jeste et al.,  Smith & Maser, 1980).  1982; Luchins, Lewine,  & Meltzer,  1983; Weinberger, Cannon—Spoor, Potkin,  1984;  & Wyatt,  Retrospective investigations, on the other hand, have  found no relationship between VBR and medication response (e.g., Nasrallah et al.,  1983; Williams et al.,  1985).  LV size and ronosis in schizophrenia.  Several approaches  have been taken to attempt to link lateral ventricle size with general prognosis in schizophrenia,  rather than that suggested by  treatment response over a number of weeks.  Since this issue  bears directly on one of the aims of this thesis, the evidence here will be examined in some detail.  As will be seen below  (Table 4), the results from the studies reported to date show a modest but consistent relationship. One study found an equivocal relationship between VBR and outcome among three groups of schizophrenics selected to represent a good, poor,  or intermediate course of illness  Katsanis et al. (1991)  First—episode schizophrenics from MAP Project.  Using median split, large VBR group has poorer outcome at both 9 and 18 months.  chronic ward patients > acute ward patients.  Chronic schiz. pta. from acutecare or continuous—care ward.  Luchins & l4eltzer  Mean VBR:  Comparing extreme groups, large VBR more prevalent in the worst outcome group.  either good, Schiz. patients: intermediate, or poor outcome,  Williams et al. (1985)  (1986)  Mean VBR:  unemployed patients > employed patients, controls.  1. No difference in 2—year outcome between large large VBR patients and all others. the largest VBR group 2. Extreme groups: had lower scores on 2 outcome measures.  Schiz. patients, either contin— uously employed or unemployed vs. matched normal controls.  First—episode schizo— phreniform patients.  Results  Pearlson et al. (1984)  DeLisi et al. (1983)  Sample  Ventricle Size and Prognosis  Author(s)  Table 4.  LA)  1  I-,.  t-I  Literature Review  (Williams et al.,  44  1985).  When VBR was used as a continuous  dependent variable, an analysis of variance (ANOVA) three outcome groups was not significant.  among the  However, when VBR was  dichotomized and used as a grouping variable among the goodoutcome subset of cases, significant results were obtained--there were more patients in the good—outcome group with normal ventricles, than with large ventricles. Luchins and Meltzer  (1986) used a cross—sectional strategy  in comparing two groups of chronic schizophrenic patients, one from an acute-care ward and the other from a continuous-care ward, which were matched for total duration of illness.  Their  results showed that the mean VBR was higher and a greater portion of large ventricles were observed in the continuous-care, pooroutcome group. Pearison et al.  (1984) assessed the association VBR and  employment as a measure of outcome.  The authors found that the  persistently unemployed patients had significantly larger ventricles, compared to their continuously employed counterparts. Furthermore, the mean ventricle size of employed patient groups was strikingly similar to that of a normal control group. Two studies investigated prospectively the relationship between ventricle size and outcome among first—episode samples. The first (DeLisi, Schwartz, Targum,  et al.,  1983) divided their  sample into normal- and large-ventricle groups:  They found a  modest but nonsignificant trend for normal-ventricle patients to show better outcome.  Stronger and significant results were  obtained when the best and worst outcome quartiles were  Literature Review  contrasted:  45  People with large ventricles were overrepresented in  the worst outcome group.  In the other study of a first-episode  sample, schizophrenics with larger ventricles had significantly poorer short—term outcome (Katsanis,  lacono,  & Beiser,  1991). At  18-month follow-up, MAP Pro.iect patients with larger ventricle size had higher symptom checklist scores.  Nonsignificant trends  for large VBR patients to have poorer work functioning and poorer global rating scores were also apparent.  These findings were  maintained even after controlling for the level of premorbid functioning.  Moreover, the results were true for the  schizophrenic group but not for the affective psychosis group. These results,  if replicated, suggest that the pathophysiologic  process(es) responsible for LVE may have a modest effect on short-term  course and outcome, over and above any effect they  may have had on childhood and adolescent adaptive functioning. In sum,  every study that has addressed the issue has found a  relationship between larger ventricles and poorer outcome in schizophrenia, albeit two of the six studies reported mixed (positive and null)  findings.  The only study that has controlled  for obvious confounding factors is from the MAP Proiect team, where LV size was associated with 18-month outcome, after controlling for premorbid status.  Together, these findings point  to an association between LV size and prognosis that is modest but robust.  When the studies are limited to those assessing a  full range of patients, the findings are limited to the short term course of schizophrenia.  Whether the relationship continues  to be true for longer-term outcome remains to be seen.  Literature Review  46  Ventricle Size and Affective Disorders In CT studies of brain structures, people with Bipolar and Major Depressive Disorders have served as psychiatric comparison If ventricular enlargement is present in affective  groups.  disorders,  it may reflect a similar, nonspecific etiological  factor as is present in schizophrenia.  The hazard may be  independent and additive to the etiology of affective disorders, it may be etiologically related,  or it may be an epiphenomenon  secondary to major psychopathology in general. Table 5 summarizes the published reports of the investigation of LyE in affective disorders, grouped according to Bipolar, Depressed or those that have studied both types of affective patients. The VBR of Bipolar patients has been the subject of six comparisons: Calloway,  As shown in Table 5, three of those studies  & Mann,  1985; Nasrallah, McCalley-Whitters,  1982; Pearlson et al.,  (Dolan,  & 3acoby,  1984) reported finding larger ventricles  in the bipolar patients, compared to nonpsychiatric controls. fourth (Dewan,  Haldipur, Lane,  et al.,  1988)  A  found an excess  prevalence of large ventricles, although no difference between patients and controls in mean VBR. et al.,  1988; Nasrallah et al.,  Of the three studies  1982; Pearison et al.,  provided prevalence data, all noted that about one-  (Dewan  1984) that  Mean VBR: a) schiz., bipolar > controls; b) continuously unemployed > employed. 2. VBR unrelated to number of hospzns.  1. Mean VBR: no group differences. 2. % with large ventricles: bipolar > schiz. no difference (on any 3. Severity: index) between large and normal VBR.  1. Mean VBR: bipolar = depr. pts. 2. VBR unrelated to outcome.  1. Mean VBR: no group differences 2. Large/small VBR unrelated to outcome for both affective groups.  Mean VBR:  1. Mean VBR: 2. Outcome:  Mean VBR:  Delusional depressives > non Mean VBR: delusional depressives, controls.  a) all patients ) controls; Mean VBR; b) poor—outcome pta > good-outcome pts.  Bipolar patients, vs. medical controls.  Bipolar vs. depressive pts. vs. normal controls. Schiz., schizophreniform, bipolar, & depr. pta. vs. medical & normal controls.  Bipolar vs. depressed pts.  Elderly depressives vs. matched healthy controls.  Severely depressed pta. vs. normal controls. Delusional vs. nondeluded depressives vs. controls  Depressed pta. vs. medical controls.  Dewan et al. (1988)  Dolan et al. (1985)  lacono et al. (1983)  Roy—Byrne et al. (1988)  Jacoby & Levy (1980)  Scott et al. (1983)  Targum et al. (1983)  Shima et al. (1984>  Depr.  pta.  > medical controls.  no group difference. large VBR pts = normal VBR pta  depressives.  1.  Bipolar vs. schiz. pts. vs. normal controls.  Pearlson et al. (1985)  Bipolars  no group differences.  Mean VBR:  Bipolar vs. schiz. vs. schlzoaffective pts.  Rieder et al. (1983)  > controls.  1. Mean VBR: bipolar, schlz. > controls. 29% of bipolars; 34% schiz. 2. Large VBR:  Results  Bipolar vs. schiz. pts. vs. medical controls.  Sample  Ventricle Size and Affective Disorders  Nasrallah et al. (1982)  Author(s)  Table 5.  1-..  m  h  Literature Review  third  48  (range 29-35%) of bipolar patients had VBR scores that  exceeded the control group mean by two standard deviations. Taken together, the two indices of LyE, mean VBR and prevalence of large ventricles, suggest that a substantial minority of bipolar patients do have enlarged ventricles. Among people with Major Depression, the evidence regarding larger ventricles inconclusive.  Three of the studies  Golden, Ruedrich, al.,  (compared to nonpsychiatric controls)  & Bishop,  1983;  (Dolan et al.,  1985;  is Scott,  Shima, Shikano, Kitamura, et  1984) reported positive results, and a fourth found positive  results for delusional but not nondelusional depressives Rosen, DeLisi, Weinberger, & Citrim, depressive-control differences Levy,  1980).  1983).  Two others found no  (lacono et al.,  1988; Jacoby &  Three of the studies assessed prevalence:  and Levy (1980)  (Targum,  Jacoby  found that 29% of their elderly sample had large  ventricles, and Targum et al.  (1983) reported a similar figure  (25%) among their delusional patients.  None of their  nondelusional depressives had large ventricles. The conclusion to be drawn from the aggregated studies of affective disorder patients to date is similar to that based on studies of schizophrenics:  A majority, but not all,  of the  investigations have found evidence of LVE in both Bipolar and Major Depressive disorders,  using group mean VBR or the increased  prevalence of large ventricles.  Precise estimates of prevalence,  or descriptions of the subgroup of patients who can be expected to exhibit LyE, are currently not possible.  Literature Review  49  Correlates of ventricle size in affective disorder.  Research  examining the correlates of ventricle size among people with affective disorders is limited.  To date,  a small number of  papers have examined correlates such as premorbid status and prognosis. (1985) compared premorbid status among  Pearison et al.  subgroups of bipolar patients who did and did not exhibit ventriculomegaly.  Their global index assessed the entire  premorbid lifespan, employment history.  from perinatal complications to cumulative The results indicated that bipolar patients  with and without large ventricles were not different with respect to premorbid status. et al.  Similar null results were reported by Dewan  (1988), who used a different global rating to assess  premorbid functioning in their Bipolar sample. Four studies have attempted to relate LVE to prognosis in affective patients.  In one study of intermediate— or good-  outcome bipolar patients, the results indicated an equivocal relationship between ventricle size and outcome (Pearlson et al., 1984).  When treated as a continuous variable, VBR showed no  relationship to two measures of outcome.  When the authors  divided the bipolar patients into quartiles based on VBR scores, and contrasted the extreme groups, however, the largest—ventricle group had poorer outcome on both measures than did the group with the smallest ventricles.  Another study of depressed patients  found that those with poor short-term outcome had larger mean VBR compared to both good—outcome patients and controls, whose scores  Literature Review  did not differ  50  (Shima et al.,  apparent, however,  1984); these differences were  only among late—onset (>50 years) patients.  Two studies reported unequivocal negative results. and Levy (1980)  Jacoby  failed to find an association between short-term  outcome and ventricle size among elderly depressed patients. final study of this type is the report by Katsanis et al. on the MAP Proiect  samples.  As described above,  The  (1991)  they related  the LV size of unipolar and bipolar patients with psychotic features to three measures of outcome, while controlling for premorbid status.  Three measures of outcome, taken at both 9—  and 18-month follow—up, were not related to ventricle size in either affective group. In sum, neither of two reports dealing with premorbid status among bipolar patients found an association with ventricle size. Regarding the association between outcome and ventricle size, the results are mixed but largely negative.  Among depressed  patients, the conclusion regarding a VBR-outcome association is similarly negative.  Overall,  the weight of evidence suggests  that, while lateral ventricular enlargement is likely present among affective patients, there is no relationship between ventricle size and either premorbid status or prognosis.  Summary:  VBR as a Predictor of Outcome  In this section, we have seen that LV size is a good, imperfect,  if  index of structural changes in the schizophrenia  disease process.  Most, but not all, studies found that  schizophrenic people have larger ventricles than medical or  Literature Review  normal controls.  51  The magnitude of the group difference is  equivalent to 1.8 VBR units, size.  or approximately 30% of original  The timing of the enlargement is probably at or before  onset of the disorder,  and further changes are unlikely.  The crucial question of the prevalence of enlargement remains unanswered.  The distribution of ventricle size in schizophrenic  people is smooth;  it is not bimodal, as it would be if some  people had a substantial degree of enlargement and others had none.  In the special case of discordant MZ twins, almost all of  the probands have ventricles larger than the unaffected twin. The most likely explanation is that enlargement is itself continuously distributed--some patients have a small increase in LV size, most have a modest degree of enlargement, and some have strikingly larger ventricles. Of more direct relevance to this dissertation project are the correlates of large ventricles, which include poor premorbid functioning, severity of illness negative symptoms), In this light,  (particularly an excess of  and poor response to neuroleptic medication.  the correlate of poor prognosis is meaningful:  Most relevant to this thesis is that all of the previous studies have found either an equivocal or an unqualified significant relationship between LV size and prognosis.  Of these, the most  relevant is the paper describing short-term outcome of MAP Project participants  (Katsanis et al.,  1991), where ventricle  size was significantly related to three measures of short-term outcome, after controlling for premorbid functioning.  Literature Review  52  Previous reports that have assessed ventricle size and outcome in affective disorder have been nearly unanimous in failing to detect a relationship between the two variables. While most studies have detected group differences in LV size between affective patients and nonpatient controls, the increased ventricle size does not seem to have any apparent significance for outcome, nor for premorbid functioning. While there seems to be substantial empirical support for the hypothesis that LV size will predict outcome, the VBR-outcome relationship will likely be of moderate size.  The principal  reason relates to the distinction between ventricular size and enlargement:  The median extent of enlargement is small  (1.8 VBR  units), compared to the wide range of variability in schizophrenic and control samples  (4—11 VBR units).  This means  that ventricle size--even if measured at the onset of the disorder——is not a powerful strategy for assessing the presence or prevalence of enlargement.  More importantly, the magnitude of  correlations between outcome and ventricle size will be much smaller than between outcome and enlargement per se. Unfortunately, no measure of enlargement is currently available.  53  Literature Review  Biological Factors:  II.  Smooth Pursuit Eve Movements  If the research on brain structural attributes in schizophrenia has been empirically driven, due to the availability of the new imaging technology, then another avenue of investigation-—that of impaired smooth pursuit eye movements-— has been theoretically driven.  Because the research has been  guided by the conceptual requirements of genetic  (trait) markers,  the literature is more coherent, and the implications of the findings are clearer. This section is organized in a manner similar to preceding sections.  First,  I will present brief  introductions to the  mechanics of oculomotion, and to the requirements of the theory of genetic markers.  Second,  I will assess the putative trait of  impaired oculomotion in schizophrenia vis—a-vis these formal requirements.  In the third section,  pursuit eye movements vulnerability.  I propose impaired smooth  (SPEM) as an index of biological  To the extent that this argument is persuasive,  will finally review the literature consistent with the notion, viz,  the correlates of impaired eye movements.  Systems Governing Eve Movements There are three principal oculomotor systems, own function and anatomy.  each with its  The first is the smooth pursuit eye  movement system, which is concerned with keeping a moving object of interest on the fovea of the retina.  The SPEM system is  designed to track the kind of motion that is illustrated by a  I  54  Literature Review  smoothly oscillating pendulum,  i.e.  relatively slow (less than  forty degrees of arc per second) and mostly lateral movements.  A  second oculomotor system governs saccades, relatively fast eye movements, wherein vision during the change of eye position is unimportant.  Saccades provide for quick changes in eye position,  and also serve to correct errors generated in the SPEM tracking. The third visual system is the vestibulo-ocular reflex (VOR), which is used by a moving person to keep a stationary target in sight. Dysfunction in the SPEM system is not specific to schizophrenia.  It is present in disorders such as alcoholism and  multiple sclerosis, which are associated with diffuse neuropathology. disorders,  SPEM impairment is also noted in localized  like Parkinsonism (Johnston & Pirozzolo,  1988).  It is difficult to infer the neuroanatomical significance of abnormal SPEM, because the control systems for eye tracking are diffusely located in the human brain.  Failure of the smooth  pursuit system can be caused by a failure in a number of cortical areas, e.g.  frontal eye fields, posterior, parietal, or middle  temporal regions.  Similarly,  it can be a result of failure in  subcortical areas associated with the modulation of eye movements, such as the basal ganglia, cerebellum, (Thaker, Buchanan, Kirkpatrick, pursuit, therefore, dysfunction:  & Tamminga,  1989).  or pons Smooth  is only a general indicator of oculomotion  it lacks anatomic specificity.  55  Literature Review  Recording and Scoring SPEM In the early studies of SPEM, subjects visually tracked a While this method  small weight that was attached to a string. was easy to use,  it did not invite a lot of precision.  A more  precise instrument is an oscilloscope, where a dot moves back and forth on a screen tracing approximately twenty degrees of visual arc. When movements of the stimulus are recorded on moving graph paper,  a smooth sine wave is produced.  performance is good, smooth sine wave.  If eye-tracking  the movement of the eyes will replicate that  When tracking is poor, however,  the tracing of  the eye movements will be jerky and irregular. To measure lateral eye movements, two techniques have been widely used.  The first is infrared reflection, where a light  source is mounted near the eye.  Because the pupil and the iris  are not equal in their ability to reflect light, and because there is a sharp boundary between the two, photocells placed at the outer edge of the eye can measure reliably differences in the amount of reflected light.  The ratio of more to less reflected  light at various locations around the eye is the basis for determining eye position at any point in time. An alternative is the measurement of small electrical potentials generated from the eye itself.  Within the eyeball, a  small standing electrical charge is always present:  the retina,  because of the high density of neurons, generates about one millivolt of potential more than the cornea. are placed at the outsides of the eyes  If two electrodes  (the canthi), and the eyes  56  Literature Review  are looking straight ahead, there will be no differences between electrodes in the amount of current measured. laterally,  When the eyes move  the retinas, with their greater electrical charge,  move closer to one electrode.  The small difference in current,  relative to baseline (eyes forward),  is the index of the  deflection of the eyes. After sources of artifact such as eye blink or head movement are eliminated, the data are subject to either global ratings or to mathematical scoring.  The most frequent global rating system  is a dichotomous one (Holzman, Proctor,  & Hughes,  1973) of good  or poor tracking performance, although a five—point scale has also been used (Shagass, Roemer,  & Amadeo,  1976).  Mathematical  scoring has been done using a gain score or signal—to—noise measure  (Lindsay, Holzman, Haberman,  &  Yasillo,  1978), but is  more often subject to a root-mean—square (RMS) error analysis. Here, the actual tracking performance--after correcting for phase and amplitude differences-—is superimposed on to that generated by the stimulus, and then the average distance between the two tracings is computed.  All of these measures, whether globally  defined or mathematically computed, correlate well with each other.  The global method of scoring is less sensitive than the  mathematically-scored approaches, and it may over-estimate the prevalence of SPEll dysfunction (Clementz, Grove, Sweeney,  1992).  lacono,  &  57  Literature Review  Theoretical Requirements for a Genetic Marker To satisfy the requirements for a putative genetic marker, a characteristic must meet four general requirements. have trait properties,  i.e.  It should  it should be stable over time, and be It must show relative specificity  unaffected by clinical state.  It should also show increased  and sensitivity to a disorder.  prevalence in unaffected family members.  Finally, the proposed  marker should segregate with the disorder in family members. will be seen below,  As  impairment in smooth pursuit eye movements  largely fulfills those theoretical requirements. Until recently, the prevalence of impaired SPEM was estimated to be 60-80% among schizophrenic people, compared to 8% among normal control subjects  (Holzman, 1992).  More recently, however,  estimates of prevalence derived from mathematically-scored data from three independent samples have generated much lower prevalence estimates et al.,  1992).  (20-37% of schizophrenic patients; Clementz  SPEM dysfunction is present in mood disorders,  but only as a state measure:  it is not present,  for example,  among people diagnosed as Major Depression in Remission (Szymanski, Kane,  & Lieberman, 1991).  Whether most or only a  substantial minority of schizophrenic patients have impaired SPEM, does not alter the fact that SPEM dysfunction shows good (relative) specificity to schizophrenia.  Hence,  it meets this  criterion for a genetic marker. Until recently,  it also appeared that the related criterion  of high sensitivity had been met  (prevalence 51-85% of diagnosed  cases; Holzman, Kringlen, Matthyse,  et al.,  1988).  Using the  Literature Review  58  mathematical scoring procedures, relative sensitivity is still achieved, with 20—37% of schizophrenic patients so identified. However,  if the unit of analysis for sensitivity is the family of  a schizophrenic proband, then the sensitivity is approximately .75 (Clementz et al.,  1992).  Thus,  this requirement of a genetic  marker is substantially fulfilled. SPEll dysfunction in schizophrenics has good trait properties,  since it is present among both remitted and  symptomatic patients  (lacono,  1988).  It has been shown to be  stable over periods as long as two years  (lacono & Lykken,  SPEM functioning is unaffected by most medication:  1978).  tracking  performance by patients during drug—free periods is much the same as while on antipsychotic medication.  SPEM function is affected  by lithium carbonate and several CNS depressant medications, but these reflect state (rather than trait) variations.  Finally,  SPEM dysfunction in schizophrenic people does not appear to be a function of an inability to maintain voluntary attention, distractibility, a lack of motivation, or a generalized deficit (Szymanski et al.,  1991).  Another requirement for a genetic marker is that it show increased prevalence among family members, and in particular show evidence for substantial heritability.  Among studies of twins,  the results suggest that SPEM shows a substantial degree of heritability:  the intraclass correlations derived from healthy  monozygotic twins average 0.68, dizygotic twins average 0.35 SPEM shows  whereas those from healthy  (lacono,  1988).  Similarly,  impaired  increased prevalence among parents and adolescent  Literature Review  59  children of schizophrenic patients, compared to similar relatives of healthy control samples.  Using global ratings of eye-  tracking, between one-third and one-half of the first-degree relatives of schizophrenics showed SPEM dysfunction, compared to about 10% of the relatives of mood disorder patients Soloman, Levin, & Waternaux,  1984).  (Holzman,  Using RMS error or gain—  score approaches, eye-tracking dysfunction is present among 1922% of first—degree family members in schizophrenia. The presence of schizotypal or schizoid personality traits/disorder is often taken to represent the presence of a schizophrenic genotype. with a genotype,  If impaired SPEM is indeed associated  then poor tracking should be significantly  associated with schizoid or schizotypal traits in nonpsychiatric This was the reasoning used by Siever, Coursey,  samples.  Alterman, et al.  (1984), who screened 284 male college students  for eye—tracking abilities.  They administered a diagnostic  interview to the best and the worst 109o of eye-trackers in the sample.  They found that a significantly greater number of poor-  tracking students (54%) met DSM—III criteria for schizotypal personality disorder, compared to 11% of the good-tracking group. One final theoretical requirement of a genetic marker is that it should segregate with the disorder,  e.g. good-tracking  schizophrenics should have good-tracking (nonschizophrenic) relatives.  Some investigators (Holzman et al.,  Holzman, & Lange,  1984; Matthyse,  1986) have provided evidence, however,  of good  tracking patients with poor-tracking but otherwise healthy relatives.  The authors explained their findings with the concept  Literature Review  60  of a latent trait or pleiotropy:  the genotype can be expressed  as poor SPEM, as schizophrenia, or both.  Correlates of SPEM Dysfunction Impaired oculomotion is often considered an impairment in the ability to sustain involuntary attention.  Since smooth pursuit  impairment is present before and after a psychotic episode, then (in)ability to maintain basic attention would likely influence inter-episodic functioning. course and outcome, correlates.  To propose it as a predictor of  one would want to assess several of its  For example,  in order to understand its unique and  independent status as a predictor, one would wish to know the degree to which it co-occurs with other putative indices of biological liability.  Further,  if the putative genetic marker is  indeed an index of vulnerability,  it should be related  independently to premorbid status and clinical characteristics. Finally,  the vulnerability marker may be related to prognosis.  SPEM and lateral ventricular enlargement.  The research  relating SPEM dysfunction to other indices of vulnerability has been relatively sparse.  Of the papers that have been published,  it is the relationship between SPEM and ventricle size that has been the most frequent object of study. In a study of chronic, treatment-refractory patients, markedly disordered eye tracking was disproportionately prevalent among patients with large ventricles (64%), compared to those with normal ventricles (30%; Weinberger & Wyatt, 1982).  In  another study where patients represented a range of severity, a  Literature Review  61  nonsignificant correlation of r=.35 between eye movement dysfunction and LV size was detected Berggren, & Schalling,  1985).  (Bartfai, Levander, Nybeck,  The correlation may have been  attenuated due to the use of linear measures of ventricle size, and statistical power was limited due to the small sample size In a third study assessing ventricle size and eye  (n=18).  movements, correlations of r=.28—.35 were reported (Siever, van Kammen, Linnoila,  et al.,  1986).  Although these coefficients  were not statistically significant, the sample size was small (n=13) The only report of the relationship between ventricle size and SPEM to use relatively large and representative samples comes from Smith (1986), who reported on three diagnostic groups from the MAP Project.  Pearson correlations between VBR and SPEM for  schizophrenic and schizophreniform groups were trivial  (r=.02 and  r=-.13, respectively). Taken in the aggregate, these studies show a weak relationship between the two variables—-if one exists at all. The magnitude of the relationship can be expected to be increased if assessed among patients with more severe forms of illness, since both SPEM and VBR have been independently associated with severity. studies Wyatt,  Such an interpretation is possible in three of the  (Bartfai et al., 1982),  observed.  1985;  Siever et al.,  1986; Weinberger &  where coefficients in the order of 0.3 were  Only in the Smith (1986) report was the association  based on a more representative sample.  There, the correlations  were close to zero for the schizophreniform and schizophrenic  62  Literature Review  patients. zero.  Thus,  the true magnitude of association may be near  This suggests that the two indices of biological  vulnerability are virtually independent,  and may represent two  distinct pathophysiological processes. SPEll and premorbid ad1ustment.  If SPEll dysfunction indexes a  disorder of nonvoluntary attention, and if it is a trait that is unrelated to symptom status, then one might expect it to be present prior to the onset of schizophrenia.  Only two studies  have directly assessed the relationship between impaired SPEll and premorbid adjustment.  In one small study (n=22), there was no  significant relationship between eye tracking and adolescent social adjustment (Bartfai et al.,  1985).  The actual coefficient  was not provided. In another small study (n=14), premorbid social adjustment was correlated both with qualitative and quantitative ratings of SPEll.  Neither of the correlations achieved significance  (qualitative measure: r=.48, p<.lO; quantitative measure r=.31, p=n.s.; Siever, van Kammen, Linnoila, Alterman, Hare,  & Murray,  1986). The small sample size in each of these studies greatly hampers any interpretation of the aggregate results. conclusion that can safely be drawn is that,  The only  if a correlation  between SPEll dysfunction and premorbid adjustment exists,  it is  not of an overwhelming magnitude. SPEll and clinical characteristics.  In the literature, there  is a scattering of results in the area of the clinical correlates of SPEll.  Siever et al.  (1986) speculated that poor eye trackers  Literature Review  63  may represent a more severe subgroup of schizophrenia.  In  assessing a small sample of chronic, treatment—refractory schizophrenic patients, they found no significant relationship between severity of psychosis and SPEM (r=.17).  The low  reliability of their single-item measure of severity, however, may have contributed to the negative results. In another study,  Bartfai et al.  (1985) used a  psychometrically superior measure of global severity of illness. Despite this improvement,  there was no relation between SPEM and  the global severity rating.  Thus,  conclusions from these two studies,  if one can draw tentative it may be that eye-tracking  dysfunction does not imply a general liability such as a more severe type of illness. SPEM and prognosis. independent of others, Again, (1986),  If SPEM represents a specific liability it may be reflected in course and outcome.  only two studies have addressed this issue.  Siever et al.  in their study of 14 young patients, did not find a  significant relationship between the total number of lifetime psychiatric hospitalizations and SPEM measures poorer trackers having more admissions).  (r=—.31, with  Once again, the small  sample size limits the interpretation of the negative findings. The only other reference is in an edited chapter by lacono (1988).  There he briefly described previously unpublished data  from the MAP Project, based on 56 schizophrenic participants. significant relationship between SPEM and 18—month outcome was present, with 90% of poor trackers having a poor outcome, compared to only 30% of good trackers.  No definition of poor  A  Literature Review  64  outcome was provided in that report, nor has a more detailed version of the results been published.  lacono concluded that  biological vulnerability may play a contributing role not only in the etiology, but also the maintenance of schizophrenia.  Summary The literature regarding smooth pursuit eye movements is relatively tidy,  even if the precise mechanism that is  dysfunctional in schizophrenia is not understood.  We have seen  that the theoretical requirements for a genetic marker have been largely met. al.,  If one accepts the Holzman and Matthyse  1988) hypothesis of a latent trait, however,  (Holzman et  then SPEM  impairment stands well as a genetic marker for schizophrenia. Together with other, possible genetic markers, eventually serve as a means to  ‘triangulate’  SPEM may  ultimate  genetically—based etiological factors. Until such a time,  SPEM impairment may be used as an index of  one kind of biological vulnerability, which may useful in predicting the course of the disorder.  We have seen that little  research has been dedicated to this purpose:  From the literature  reviewed above, SPEM impairment is largely or completely independent of another leading candidate for biological vulnerability,  i.e.  enlarged ventricles.  If SPEM impairment is indeed a biological liability, then early signs may be evident. adjustment,  two studies  Using the level of premorbid social  (each with small samples) detected  correlations between SPEM and premorbid adjustment that were in  Literature Review  65  the expected direction, even though each was statistically nonsignificant. liability,  Similarly,  if SPEM is indeed a biological  it may presage a poorer course.  This was the case in  one brief report from the MAP Project regarding short—term outcome (lacono et al.,  1988),  and a nonsignificant trend was  noted in a small study reported by Siever et al. Overall,  (1986).  there may be good reason to hypothesize that SPEM  impairment will be associated with a poorer prognosis.  This  project is the first to assess that relationship on a large sample, and with a wide range of patients.  This project is also  the first to assess the unique and independent contribution of SPEM impairment on long—term outcome, while controlling for other factors that may simultaneously affect the course of schizophrenia.  Literature Review  66  General Issues in the Prediction of Course and Outcome  The tradition of identifying predictors of the course of schizophrenia has a long and illustrious history.  Of most  interest have been predictors that are evident prior to or at the time of the first episode, both because of the notion that they are untainted by the effects of lengthy illness, and because of the implications for early and presumably more effective intervention. outcome,  For more extensive reviews of predictors of  interested readers may consult Stoffelmayr, Dillavou,  and Hunter  (1983) or Beiser and lacono  (1990).  The variables  germane to the proposed study, sociodemographic factors, aspects of premorbid functioning, and characteristics of the mode of onset, are briefly reviewed below.  Sociodemographic and Premorbid Factors Demographic variables such as age at onset, marital status, and sex have been repeatedly associated with outcome.  For  example, an early age at onset is associated with a poorer outcome (=.15—.25), suggesting that younger people have a more serious form of illness, or that they do not have a chance to develop their social skills, occupational training, cognitive capacities Adams,  & Chapman,  (Bromet, Harrow,  & Kasi,  or even basic  1974; Burstein,  1974).  Marital status is slightly stronger as a predictor  l— 2 (L=.  .32), with people who have ever been married having a better prognosis, compared to those who have not.  Marital status as a  Literature Review  67  predictor is usually seen as reflecting better premorbid social skills and a later age of onset (Lorei & Guel,  1973;  Stephens,  Further, marital status among people with schizophrenia  1978).  is strongly related to sex, where women tend to be married or have been married by the time of illness onset. Sex is also related in other ways to outcome.  Compared to  men, women tend to be have briefer hospital stays and manage to stay out of hospital for longer periods of time (Angermeyer, Kuehn, & Goldstein,  1989).  They tend to have higher levels of  premorbid functioning and later onset Further,  (Goldstein,  1988).  the superior outcome of women over men appears to be  related to their greater representation among familial types of schizophrenia, where the nonfamilial cases tend to have a poorer outcome (Goldstein, Santangelo, Simpson, Finally,  & Tsuang,  schizophrenic women may have an advantage due to the  relative dopamine—inhibiting effects of estrogens Thus,  1990).  (Seeman,  1982).  in assessing the role of novel predictors of outcome, one  must account for several overlapping predictors that are already well-known.  Duration of Onset and the Course of Schizophrenia In order to assess the importance of any baseline variable (measured at the onset of the disorder) on the outcome of schizophrenia, one must control for the effects of varying lengths of onset.  For example, any differences in social  relationships observed at intake may simply be the result of a  Literature Review  68  longer and insidious onset and as such may be a feature of the disorder, and not a premorbid characteristic. The notion that acute (vs.  insidious) onset predicts good  prognosis in schizophrenia has achieved legendary status as a prognostic indicator Sins,  & Marder,  Carpenter,  (cf. Bland,  1982;  Docherty, van Kammen,  1978; Neale & Oltmanns,  1977).  1980;  Strauss &  This conclusion may be premature, because the  area suffers from a number of major limitations. The first limitation relates to  the failure to clearly  conceptualize——and hence assess——onset per se. empirical reports,  Of the six  illustrated in Table 6, only one has used the  concept of onset to characterize the illness itself.  Vaillant  (1964) used as a definition the duration from the first signs of the episode to the appearance of florid psychotic symptoms, the prodromal period.  i.e.  All others have used the concept of onset  to describe the time lag between the appearance of psychotic symptoms and treatment contact  (usually admission to hospital).  A second limitation is that the studies that have investigated the concept appear to be based, with one exception, on chronic or mixed groups of patients.  As such, the predictor  in question is not the onset of illness, but the onset of subsequent episodes.  Inasmuch as the characteristics of later  episodes are influenced by factors such as treatment and medication, the predictive role of later episodes is probably unlike that of the first episode,  i.e. the onset of the illness.  Moreover, the use of readmitted, chronic patients probably  1. Reactive subtype has better outcome. 2. Characteristics related to better outcome include acute (vs.insidious) onset.  Acute onset more frequent among patients who achieved full remission (82%) than among those who did not (42%). ‘Dorrelation between mode of onset and: a) 9 mo. hosp. status: r=.21 (p<.O5) b) 5 yr. hosp. status: r=.14 (n.s.) c) Days in hospital: r=.22 (p<.O5). Mode of onset unrelated to in/out of hospital at follow—up, or days in hosp Correlation between mode of onset and: a) hosp. days or occupl functg: n.s. b) social adjustment and global rating both r > .35 (p<.O ). 5  Acute onset correlated with outcome (r=.24, p<.05) Shorter duration of onset ( < 1 wk. and 1-4 wks. vs. > 4 wks.) related to reduced probability of relapse.  Retrospective follow—up of patients hospitalized between 1944—1954.  Consecutive admissions in 1947-50 and 1961—62.  Consecutive male admissions.  Mixed group of males; firstand multiple—admissions.  In—patients.  Retrospective follow-up of pts. first admitted 1948—59. Prospective 2—year follow—up of first—episode patients  Vaillant (1964)  Cancro & Sugerman (1968)  Burstein et al. (1974)  Bromet et al. (1974)  Stephens et al. (1978)  MacMillan et al. (1986)  Results  Stephens & Astrup (1963)  Sample  Onset Characteristics and Outcome in Schizophrenia  Author(s)  Table 6.  0-I  1  Literature Review  7  reduces the variance in outcome and thus any association with a predictor variable. A third, psychometric limitation is that the studies of onset have tended to dichotomize a continuous variable:  most  investigators used the simple dichotomy of acute/insidious, based on a six-month cut_off*.  Only one group of investigators,  seems, has gone beyond the dichotomy. MacMillan, Crow, Johnson,  & Johnstone  it  As shown in Table 6, (1986) described onset in  their sample by means of a three-point ordinal scale.  Whether  these three categories of onset or the more usual dichotomous approach was used, much predictive sensitivity was likely lost. Finally, despite its prominent status as a predictor of outcome,  evaluations of the type of onset have not controlled for  other well—known prognostic indicators.  One would expect, for  example, a strong degree of intercorrelation between type of onset, premorbid social functioning, and outcome (Cancro & Sugerman,  1968;  Stoffelmayr et al.,  1983).  Similarly, the type  of onset may gain its predictive power because it signals a more severe form of illness.  In general, any assessment of the unique  influence of onset on outcome would need to control for the influence of “third variables” such as age at onset, sex, and premorbid social or occupational functioning. To summarize,  research must incorporate a number of  methodological requirements in order to achieve a clear assessment of the influence of onset on the course of illness. To date, I have not been able to discover the origin of the tradition of using 6 months as the crucial point for defining acute vs. insidious onset. *  Literature Review  These include:  71  the use of a first-admission sample;  distinguishing between two stages of onset, the prodromal phase and the treatment lag—time; analyzing these intervals as continuous variables; and finally, controlling for other, overlapping predictors of outcome.  Predictors of Short— vs. Longer—Term Outcome The schizophrenic process,  even when disabling and chronic,  does not involve continuous and unremitting deterioration. McGlashan’s  (1988)  review of long—term follow-up studies of  schizophrenia concluded that, at some point, the decline in functioning appears to bottom out or reach a plateau.  He wrote  that the bottoming-out point varies widely between individuals and across studies, but he estimated that it occurs on average roughly 10 years after the onset of illness. In many domains, past behaviour.  the best predictor of future behaviour is  This axiom is no less true vis-a-vis  schizophrenia where,  for most measures of outcome, the strongest  predictive relationship is with the baseline value of the same measure of functioning.  Of these baseline values, variables  related to premorbid functioning,  socializing, occupational  functioning, and global ratings have been most frequently examined.  In general, the magnitude of the correlation between  baseline and outcome measures of functioning tends to be =.30.40  (Stoffelmayr et al.,  1983).  Social functioning, however, has  a unique role as a predictor in that the magnitude of the correlation with outcome is remarkably consistent (L=. ) 36 —. 32  Literature Review  72  across the domains of functioning that serve as measures of outcome.  Previous work functioning, on the other hand,  variable,  in that it correlates well with work functioning at  is more  follow—up (average =.4O), but more modestly with other domains of functioning at follow—up (average =.22 with hospitalization, and .24 with global assessment). McGlashan (1986) also offered evidence for the changing role of predictor variables.  He reported that, for the first decade  following diagnosis, variables relating to premorbid social adjustment and intimate relationships were the best predictors of outcome.  In the second decade, however, family environment and  early-illness characteristics attained predictive primacy.  In  particular, a history of overinvolvement between patient and family and symptoms such as paranoid ideation were the best prognostic indicators.  In the third decade and beyond, the best  predictors of outcome were a family history of schizophrenia and the premorbid acquisition of vocational skills. the study’s  Unfortunately,  retrospective design necessitated the use of a  sample that had at least two years of archival data available. This meant that McGlashan’s sample was biased towards chronicity at the outset, perhaps with a restricted range of outcome, and limited generalizability.  Nonetheless,  it shows phase-related  specificity in predicting the outcome of schizophrenia. Even if variables do retain their predictive power at different phases of the illness, one would not expect that similar magnitudes of association will be present over different predictive intervals.  Thus,  it may be important for follow—up  Literature Review  73  studies to identify the outcome period and exercise caution in generalizing the relevance of predictors to outcome “as a whole” (McGlashan,  1986).  That there are different predictive relationships at various points in the course of illness may have important implications for intervention.  For example, to demonstrate that the natural  history of the disorder includes a plateau may point to renewed treatment efforts--long after patients,  family and service-  providers have been resigned to a fate that appears to be more disabling than it would be otherwise.  Similarly, changes in the  relative strength of predictor variables at different points in time may have implications for the relative emphasis on work, family, or social life at various points in the patient’s “career”.  Finally, to demonstrate similar predictive  relationships at different points in time may have the simple function of providing a replication of earlier results, and hence serve to make the earlier conclusions more robust. Alternatively,  if a variable is associated with outcome at one  point in time but not another,  the possibility of phase—specific  predictive relationships is introduced.  Current Vulnerability-Stress Models  In the foregoing pages, ‘diathesis-stess’ and  frequent mention has been made of  ‘vulnerability—stress’  as if there were one  Literature Review  unitary concept.  74  It is worthwhile to review briefly two of the  more frequently cited models, to show that elaborations of the vulnerability-stress concept are anything but unitary. In 1977, Joseph Zubin and Bonnie Spring announced “A New View of Schizophrenia”, where a large number of people are endowed with a degree of vulnerability that under suitable circumstances might be expressed in an episode of schizophrenic illness. Different concepts of diathesis had previously been proposed, but (Gottesman &  only as a function of variable genetic contributions Shields,  1982; Meehl,  two new elements:  1962).  The Zubin and Spring model added  First, they argued that there could be  numerous contributions to a person’s vulnerability, above genetic inheritance. could be due to traumas, complications,  Any number of acquired propensities  specific diseases, perinatal  family experiences,  adolescent peer interactions,  and other life events that would affect homeostasis. way,  over and  In this  they proposed an integration of models of etiology derived  from sociological, developmental, genetic, neuropsychiatric, and other traditions.  Particular concepts or variables were not  proposed as elements of the model,  only that causal influences  could come from other than the genetic domain. Second,  Zubin and Spring argued that vulnerability was a  stable, enduring trait, whereas schizophrenia was a state that waxed and waned.  A person with a high level of enduring  vulnerability would have many ‘challenges’  in daily living that  would be sufficient to prompt a psychotic episode, whereas one with low enduring vulnerability would require nothing short of a  Literature Review  75  catastrophic event to precipitate psychosis.  The challenges that  disrupt adaptation may be endogenous, e.g. maturational changes within the organism,  inadequate nutrition,  responses to infection or stress.  or pathological  Alternatively, exogenous  challenges may be life events or disruptions in one’s social network. A more elaborate model is provided by Nuechterlein (1987; Nuechterlein & Dawson,  1984).  Figure 1 illustrates his “overall  heuristic schema of schizophrenic psychotic episodes” 306).  (1987, p.  It shows four classes of variables that contribute to  psychotic episodes:  personal vulnerability factors, personal  protective factors, environmental protective factors, and environmental potentiators and stressors.  On some occasions,  they interact to determine transient intermediate states  (e.g., a  prodromal period), which are then subjected to coping strategies. If coping is not successful,  a psychotic episode will ensue.  The Nuechterlein (1987) model is an advance over that proposed by Zubin and Spring (1977) detail.  in that it is specified in greater  Each node in Figure 1 corresponds to one or more  empirically demonstrated vulnerability factors.  Unfortunately,  none of the risk indicators are pathognomonic for schizophrenia; thus, Nuechterlein points to a need for several separate vulnerability-stress models to address subgroups of schizophrenia that likely have distinct  (but overlapping) etiologies.  A  related need is greater attention to the distinction between initial and later episodes, so that residual effects from the  Literature Review  Figure 1.  76  Nuechterlein’s  for Possible Factors  Personal Vulneraoflity Factors  (1987)  Vulnerability—Stress Framework  in the Development of Schizophrenic Episodes  Literature Review  77  first will not be confused with vulnerability ‘markers’ Magaziner,  & Steinhauer,  1983)  (Zubin,  for subsequent episodes.  The vulnerability models of Zubin and Spring (1977) and Nuechterlein (1987) are offered as conceptual frameworks. such, they have three prominent limitations. the  ‘black box’  of  As  The first is that  ‘Interaction’ shown in the Nuechterlein figure  is present for conceptual and not empirical reasons; the principal emphasis to date has been on delineating the properties of the various components.  As yet,  the interactions among those  components have received less attention, as have the bidirectional influences of the model’s components by the double—headed arrows).  (as indicated  A second limitation is that the  significance for a vulnerability model of some of the components is itself not completely clear,  e.g. the hyper- vs. hypo  responsive electrodermal patterns of schizophrenics.  Third, both  the Nuechterlein and Zubin and Spring models pertain to onset and relapse; they are not intended as vulnerability models of course and outcome.  SummarY and Discussion of Literature Review  For a field that is conceptually nested in a vulnerablity stress paradigm, the apparent absence in the schizophrenia literature of such models of course and outcome is surprising. Nonetheless, as has been described in the previous review, varying amounts of published data are available on the individual  Literature Review  78  psychosocial and biological components that may be combined into a diathesis—stress model.  Psychosocial components of a diathesis—stress model.  The  association between enhanced health status and supportive social relationships is well documented.  Research workers must address  the complexity of peoples’  involvement with others  DeLongis,  1985).  1986; Lieberman,  about social relationships,  (Coyne &  We need to know what it is  from what kind of people, and in what  kind of situations, that enhances health. Psychotic disorders, extensive scrutiny.  however, have not been the subject of  Of two studies that have examined directly  the effect of supportive social relationships on the course of schizophrenia, both reported a positive relationship with outcome.  Only one of those reports  (Erickson et al.,  1989),  based on MAP Project data, assessed the various components of social relationships as they pertained to course and outcome. That study awaits replication. A proper understanding of the impact of social relationships on the outcome of schizophrenia in particular has some unique requirements.  Among the most important are controlling for the  effects of onset, using a broad and representative sample of first—episode schizophrenics, and controlling for the effects of other influences on outcome.  Assessment of social relationships  should include elements from the network, transactional, and perception-based approaches.  Literature Review  79  Other elements particular to schizophrenia also need to be addressed.  The role of social relationships must be compared  both with normal controls and a psychiatric comparison group. Moreover,  if it is true that McGlashan’s  (1986) assertion is  true that predictive relationships vary over time, then the role of social relationships in predicting outcome should be assessed with at least two follow-up assessments. In the study of life events  (LEs), the most common research  paradigm is to assess events that are proximal to a psychotic episode.  Results from recent studies indicate that increased  numbers of life events have an Immediate, adverse effect, but no investigation has yet assessed the cumulative effect of LE5 over a longer period of time. Such an assessment must occur in the context of a high degree of statistical or methodological control.  One must control for  factors that influence outcome, such as baseline functioning, the nature of onset, and the usual demographic characteristics.  A  representative sample must be used to avoid selecting only acute onset patients or some unknown portion of patients whose onset Is easily described.  Without a sample based on the full range of  onset characteristics, we cannot know that the existing findings are not an artifact of research design. An assessment of the effects of LEs, as they are moderated by social relationships, has not yet been reported.  This may be  due to the recency of the evidence regarding the effect of LEs. An assessment of such a two—factor stress-process would be a significant contribution to the literature.  Literature Review  80  Biological influences on the course of schizophrenia.  Of  many possible biological indices that may be related to the course of schizophrenia, two prominent candidates are lateral ventricle  (LV) size and smooth-pursuit eye movements  (SPEM).  Although the precise mechanism that may link these anomalies to a detrimental influence on course is unclear,  linking prognosis and  LV size or SPEM dysfunction may provide a valuable clue to the underlying pathophysiology. Lateral ventricular enlargement is probably present in a substantial minority of schizophrenic patients.  It is also  present in mood disorders, but it lacks the significance for premorbid functioning and prognosis seen in schizophrenia.  The  modest degree of ventricular enlargement and the substantial overlap with nonschizophrenic samples will limit the magnitude of any association between LV size and outcome.  Such an expectation  is consistent with the literature reviewed above,  including the  report from the MAP Project based on short—term outcome.  The  current project provides an opportunity for partial replication, based on longer-term outcome and a greater degree of control over extraneous influences. In contrast to LV size, smooth pursuit eye movements are under partial genetic control, and dysfunction seems to be relatively specific to schizophrenia.  Previous efforts to  investigate the possible association between smooth-pursuit eye movements and outcome indicated a modest, nonsignificant relationship with lifetime hospitalizations and poorer short—term outcome.  The current project has an opportunity to examine that  Literature Review  81  association in greater detail, based on longer-term outcome. Once again,  one would expect an association to be modest--since  many neural mechanisms are involved in the control of SPEM. A vulnerablity—stress model of course and outcome.  A  vulnerability—stress model as it pertains to course of schizophrenia has not been reported.  Reports of the interplay  between biological and social factors with respect to treatment interventions have appeared, but only to demonstrate in an immediate and experimental fashion that interpersonal stressors (family member rated as high on Expressed Emotion) have a demonstrable effect on a patient’s psychophysiological status (Sturgeon, Turpin, Kuipers, Berkowitz & Leff,  1984).  To demonstrate a diathesis-stress model of course and outcome, social factors must be assessed at the onset of the disorder,  in order to argue that they may have a causal role.  Supportive social relationships and stressful life events are conceptually well-suited as components of a diathesis-stress model.  Similarly, the choice of these measures is supported  empirically,  even if the body of literature pertaining to  schizophrenia is limited. Among a host of candidates, the biological factors reviewed above are among the most robust findings with respect to schizophrenia.  Lateral ventricular enlargement is present among  both schizophrenic and affective patients, but it seems to have significance for premorbid and post-morbid functioning only among schizophrenics due to its presence in the “Type II” syndrome of Crow (1980).  Smooth pursuit eye movement dysfunction, on the  Literature Review  other hand,  82  is not present among patients with mood disorders.  Even though recent estimates of reduced prevalence seem to indicate that only a substantial minority of schizophrenic patients have impaired SPEM, such a dysfunction may be associated with inferior course and outcome due to impairment in abilities related to nonvoluntary attention. social relationships,  To the degree that each of  life events, VBR, and SPEM are related to  prognosis, their concurrent and interactive influences may be combined in a vulnerability-stress model of course and outcome.  83  IV.  METHODS:  THE M.A.P. PROJECT  The arkers d Eredictors of Schizophrenia (MAP) Project at the University of British Columbia was initiated in 1981 through the joint efforts of Drs. Morton Beiser and William lacono.  The  principal interests of lacono were the psychophysiological markers of schizophrenia, whereas Beiser was primarily concerned with the psychosocial predictors of the course of schizophrenia. Data collection began in 1982, and the last of the five-year follow-up interviews were held in the late summer of 1989.  The  methodological attributes of this study include a representative schizophrenic sample, two comparison groups, reliable diagnostic procedures,  incorporation of collateral sources of information,  and a careful selection of measures.  Samples The full MAP Project includes three samples:  a group with  first—episode schizophrenia, a comparison group with first— episode affective psychosis, and a matched comparison group of normal volunteers.  Since the objectives of this dissertation  involve within-group predictions, and issues regarding specificity between diagnostic groups, the normal comparison group will not be discussed further.  When recruited, the index  participant from each of the three groups was asked to nominate a family member or close friend to provide collateral information. Members of the schizophrenic and affective psychosis comparison group were recruited using the same broad set of  Methods  criteria.  84  Chosen to ensure an overinclusive bias at the level of  the referring agencies, the criteria were: a) at least one clearly psychotic symptom (e.g. hallucination, delusion, thought disorder, marked change in behaviour), Q marked reduction or loss of b) two or more of: interest in usual activities; reduced initiative and drive; deterioration in performance; social withdrawal; persistent self—neglect; c) the ‘a’ or ‘b’ criterion occurred within three months of the referral to the research project; d)  age 15-54 years; and  e) residence in the Greater Vancouver area for six months. Exclusion criteria included:  receipt of a prescription for  antipsychotic or antidepressant medication prior to contact with the referring agency;  evidence of organic disorders  epilepsy); psychosis due to alcohol or drugs;  (e.g.  and mental  retardation. The use of the broad inclusion criteria had two advantages. The first relates to minimizing potential selection bias at the level of the referring agent:  the liberal criteria ensured that  the full range of schizophrenic patients were brought to the attention of project staff, rather than just the narrow and stereotypic type of chronic,  impaired schizophrenics.  A second  advantage was the “instant” creation of a psychiatric comparison group. A further effort at minimizing selection bias involved the use of a community-wide referral system.  Frequently, recruitment is  done entirely from in—patient populations--resulting in samples that tend to have disproportionate numbers of patients with  85  Methods  severe forms of the disorder.  To avoid this bias towards  severity, the MAP Project “cast the net” to a wide range of treatment services.  In addition to all in—patient facilities,  every out-patient service and all psychiatrists in private practice in Greater Vancouver were regularly contacted.  Referral  sources also included all university and college counselling centres,  immigration counselling centres, and a 1-in-S sample of  all general practice physicians.  Short of knocking on doors in a  residential survey, we covered as comprehensively as possible all of the Greater Vancouver area. A total of 300 psychiatric patients were identified. these,  175 signed a consent form to participate.  the referrals 31  Of  Ninety-four of  (31.3%) refused to participate, and an additional  (10.3%) either moved out of the area or disappeared from the  treatment system before they could be contacted.  Analyses  comparing the 175 participants with the 125 referrals lost to the study revealed no age or sex differences*.  Procedures Data relevant to this thesis were collected at four points in time.  Intake interviews were conducted within three months of  * Due to ethical constraints, the effects of this “pre—inclusion We have age and sex data only attrition” are largely unknown. because research assistants were able to observe these characteristics in approaching the referred patients for consent. Ethical considerations prohibited the collection of any other The effect of pre-inclusion attrition seems to be information. virtually ignored by investigators engaged in predicting the of dozens of studies in the area, none course of schizophrenia: has addressed this problem.  86  Methods  the first treatment contact.  Follow—up interviews were held nine  and eighteen months later, as well as five years after intake to the study.  At intake,  each subject was asked to nominate a  family member or close friend who would serve as a collateral informant; these “significant others”  (SOS) were interviewed only  at intake and at 18—month follow-up.  The time—line for the  various interviews held with index participants and their SOS is illustrated in Table 7. Once potential subjects had been identified by a referring mental health agency, a member of the project staff confirmed that the inclusion and exclusion criteria had been met.  The  potential participant was then approached to gain consent for the diagnostic and psychosocial interviews.  In the case of acutely  psychotic patients, this occurred only when the attending physician (or service provider)  indicated that the person was  capable of providing informed consent.  To qualify for the “first  admission” criterion, patients must have had their first treatment contact not earlier than three months prior to their referral to the study. At intake to the study,  a psychiatrist or registered  psychologist administered the Present State Examination (PSE, 9th Ed.; Wing, Cooper,  & Sartorius,  1974) to all patients.  At this  time, Master’s—level interviewers administered questionnaires regarding various psychosocial variables  (described below).  The  interviewers also obtained collateral information from the SO and reviewed clinic or hospital records.  Methods  87 Table 7.  Summary of Measures  Intake interviews * * * * * * * * * * *  Demographics Onset characteristics Present State Examination (PSE; 9th Edition) DSM—III Axis I diagnosis Supportive social relationships Life events in year prior to first treatment contact Smooth pursuit eye movements CT scan: brain lateral ventricle size Collateral information from significant other (SO) Review of clinical records highest level of Global rating (DSM-III Axis V): adaptive functioning in year prior to intake  Nine-month follow-up * *  Axis I diagnosis Life events since discharge from first hospitalization, or since intake interview  Eighteen-month follow-up * * * *  Axis I diagnosis re 18—months & ‘Best Overall’ diagnosis Life events since previous (9-month) interview Collateral information from SO Global rating (DSM—III Axis V): highest level of adaptive functioning since 9-month interview  Five-year follow-up * * * * * *  Occupational functioning Social activities (where full interview was conducted; see text) Type of residence Medication status Treatment status Global rating of adaptive functioning (DSM-III Axis V) in the past month  88  Methods  Upon completion of the intake diagnostic and psychosocial interviews, the subjects were asked if they would participate in another component of the project.  They were offered the option  of returning to lacono’s laboratory where several psychophysiological variables,  including smooth pursuit eye  movement function, would be assessed.  A total of 134  (77%) of  the 175 psychiatric participants volunteered for the second stage of the study.  This second stage took about 90 minutes.  Upon completion of the psychophysiological protocol, participants were asked if they wished to participate in a third stage, the CT scan.  This session lasted an average of 40  A total of 90 participants, or 51% of the original  minutes.  psychiatric group, consented to the CT scan. Nine months after their first treatment contact, participants were contacted to take part in a 45-minute follow-up interview.  The interview included an abbreviated Diagnostic  Interview Schedule  (DIS; Robins, Helzer, Croughan,  & Ratcliff,  1981), which was later used to determine diagnostic status at the time of the nine—month follow-up interview.  Stressful life  events were also assessed at this stage. At 18 months, second,  participants were contacted to take part in a  longer follow-up interview.  At this time, a psychiatrist  or registered psychologist repeated the PSE diagnostic interview. A lay interviewer obtained from the participant extensive data regarding treatment, medication, work or school performance, living situation, and the extent of socializing during the follow-up interval.  Interviews with SOs also provided follow—up  Methods  89  data.  The diagnostic data and information regarding outcome  status  (collected from the index participant, the SO, and a  review of medical records) were used to determine diagnostic status and global ratings at 18-month follow—up.  A total of 127  (73%) of the psychiatric group were successfully interviewed at 18 months. Five years after their inclusion in the study, summers of 1987,  1988, and 1989,  in the  participants were again  On these occasions, due to scarce resources, the full  contacted.  PSE diagnostic interview was not used.  In its place, the trained  lay interviewers administered an abbreviated version of the DIS. Other than this change, the full battery of questions from the previous follow-up interviews was used to  document the various  aspects of adaptive functioning and treatment utilization that together comprise outcome.  The measure of social support,  previously administered at intake, was also repeated at the fiveyear follow—up.  The SO was not contacted at this time.  Attrition is a problem for all longitudinal studies.  In the  MAP Project, two factors made the potential for attrition even more pronounced. 25 years,  First, because the sample was young  SD=7 years) and predominantly male  (mean age  (68%), we expected a  lot of geographical mobility over the course of the study. A second factor interfering with the response rate at follow-up was that, during  the early stages of the project, British Columbia  was gripped by an economic recession: the unemployment rate for the young male population hovered near 25% (Shaw,  1985).  At this  Methods  90  time, many young people had to travel great distances to find work. In light of the usual influences on attrition, as well as the two that were especially relevant to the MAP Project, project staff mailed out newsletters and Christmas cards over the course of the follow-up period, and each time included change of address cards. To further aid in relocating participants,  both the index  subject and the SO were asked, at intake and at 18 months, to provide the names of two other persons who would likely know their whereabouts,  in case they lost contact with the project.  A  final effort at minimizing attrition was undertaken in the summers of 1987 and 1988, when we attempted to locate by telephone those participants who would be due for a five-year follow-up in subsequent years. In planning the five-year follow-up we thought that, due to the greater passage of time, we might not achieve a sufficient response rate if we depended solely on face—to—face interviews. With this in mind, we developed alternate procedures that would serve as proxies, and could be used in several situations.  One  procedure was for subjects who refused to participate in the full interview, but would agree to answer a few questions over the telephone.  Second, when we could not locate the subject, we  contacted the SO or another family member with whom we had had previous contact.  Where the contact person had extensive  knowledge of the subject’s status, and where the subject was unavailable, the contact person served as the informant. Finally, where we could not locate the subject, and where the SO  91  Methods  was not able to provide these few data, we relied on archival sources, viz, clinical or hospital records.  The data collected  in this proxy manner related to current status in four domains: treatment, medication,  and living  occupational functioning,  All of the “live” interviews for the five—year  situation.  follow-up were completed by August 1989; the response rate is described below.  Measures The measures, described below, are also summarized in Table 7.  The descriptions of the instruments are divided into  sections concerning those used at intake, at nine- and 18—month follow-up, and finally at five-year follow—up.  Samples of the  questionnaire and interview measures that were used in the data analyses are included as Appendix I. Intake.  The measures at intake pertain to demographic  characteristics, baseline functioning, diagnoses, as well as to the predictors of  interest: social relationships,  life events,  eye movements, and brain ventricle size. *  Demographic questionnaire.  This instrument collected the  standard range of demographic information, occupation, age, and sex.  including  Since the psychiatric group were  young, and manywere still in high school or university, ratings of socioeconomic status parents’  occupation.  Blishen & Carroll,  (SES) were based on their  Using the Blishen socioeconomic index  1978;  Blishen & McRoberts,  rating reflects a combination of education,  1976),  the SES  income, and  92  Methods  prestige.  In all analyses,  the SES variable for any  participant was taken as the higher of the two parental ratings. *  Present State Examination (PSE, 9th Edition).  (Wing et al.,  1974)  The PSE  is a semi—structured diagnostic  interview, designed to yield reliable diagnoses.  It was  administered at intake and at 18—month follow-up by either a psychiatrist or a registered clinical psychologist. *  DSM-III Diagnoses.  Case conferences, based on the PSE  interview and collateral information, were held to determine DSM-III  (American Psychiatric Association [APAI,  diagnoses. conferences.  1980)  At least two clinicians were present for case Structured checklists for each diagnostic  category were used in order to maximize adherence to the DSM LII Axis I criteria.  Following completion of the checklists,  diagnoses were generated on all five axes of the DSM—III for each member of the psychiatric group.  Table 8 shows the  sample size for the various diagnostic groups for all Axis I categories. The DSM—III Axis V rating represents the highest level of adaptive functioning in the year prior to intake.  It  assesses both social and occupational functioning; ratings were based on data provided both by the index participant and the SO. Assessments of the formal properties of the DSM—III axes that were conducted for previous stages in the MAP Project are satisfactory.  For the Axis I diagnoses, we  93  Methods  Table 8.  Sample Size at Three Stages of Intake, by Diagnosis  Stage of intake  Diagnosis  Psychosocial (n)  Psychophysiological (n)  CT scan (n)  Schizophrenic  72  60  45  Bipolar  44  35  23  Major Depression Schizophreniform Schizoaffective Paranoid  30  26  11  9  8  4  7  2  3  7  1  3  6  2  1  175  134  90  Other&  TOTAL SAMPLE  Note. This table displays “best overall” diagnoses, which At combine data from intake and two follow-up interviews. intake, two cases who received a functional psychosis were later diagnosed as organic psychoses (viz, drug-induced, and hyperthyroidism). aAtypical, Brief Reactive, or organic psychosis).  Methods  94  measured agreement among pairs of project clinicians based on the first 15 cases entered into the study. coefficients ranged from .74 to  .96.  The Kappa  Inter—rater  reliabilities for Axis V 1 the highest level of adaptive functioning in the year prior to intake, were  .89  (based on  Pearson correlations). *  Interview Schedule for Social Interaction (1851).  This  questionnaire assesses both the structural characteristics of the social network and the nature of and satisfaction with supportive social transactions.  It was designed to query  eight areas of social relationships as theorized by Weiss (1974), and was standardized on a neurotic and a representative community sample Henderson, 1980).  (Duncan—Jones,  1981,  1980; Henderson, Duncan—Jones, Byrne,  1981b;  & Scott,  Unfortunately, the results of the ISSI with the MAP  Project psychotic and normal control samples did not support the use of those eight scales: were far from satisfactory. own scales from the 1551  Cronbach’s alpha coefficients  As a result, we generated our  items.  In our revision, social  support as a transactional concept is measured with two scales:  the availability of either close and confiding  relationships, or acquaintances. the perception of support,  i.e.  Two other scales measure the degree of satisfaction  with each of confidants and acquaintances. shown in Table 9  (headings  ‘B’  through  ‘E’):  coefficients are quite satisfactory (.70 to sample).  The scales are the alpha .89 in our  The structural characteristics of the respondent’s  Methods Table 9.  95 Interview Schedule for Social Interaction:  Scale Items  Kin and Nonkin Social Network Size: Number of “family and close friends” named in response to the Persons named following questions (divided into kin vs. nonkin). in response to multiple items are counted only once: * Person with whom you feel most comfortable talking frankly * All persons with whom you are close, fond of, attached to * People you count on for emotional or material support * A single, lasting intimate relationship? * An important person with whom you’re no longer in contact? * Person close to you who died recently? * One who knows you very well as a person? * A person you can lean on? * Someone with whom you share important feelings? * A person who comforts you in his/her arms? * People who regularly depend on you for help, guidance or advice  A.  B. Availability of Close and Confiding Relationships * No. of friends who could come to your home at any time (unannounced), and still be welcome? * No. of friends whose home you could visit at any time * No. of family and friends with whom you talk frankly * No. of friends or family who comfort you in their arms * No. of people outside your home who really appreciate what you do for them? * No. of people who tell you that you are good at things C. Adeuacv of Close and Confiding Relationships For each of the items listed in ‘B’, respondants are asked “Is this about right for you, or would you like more or less of this kind of relationship?” (scored on 1-3 Likert-type scale). D. Acquaintances: Availability Introduction: “Acquaintances are people you know a little but Items include: who are not close friends”. * No. of people who you know just a little, e.g. you may not know their names but you greet each other when you pass by * No. of people not close to you from whom you easily ask small favours * No. of acquaintances for whom y. do small favours * No. of acquaintances from whom you could expect help in times of trouble * No. of acquaintances who can expect practical or material help from you in times of trouble * No. of people who, when you get upset, you can tell them just how you feel E. Adequacy of Acauaintances Ratings of adequacy of acquaintances, as per  ‘C’.  96  Methods  social network are assessed with two indices,  reflecting the  number of kin and nonkin in the network. *  Duration of onset.  were determined:  Three stages of the onset of psychosis  the appearance of first noticeable signs,  the first appearance of florid psychotic symptoms, and the initiation of treatment-seeking.  The duration of the  intervals between the first noticeable signs and florid psychotic symptoms corresponds to what is often called the prodromal period.  The interval between the first appearance  of florid symptoms and treatment—seeking we have called treatment lag—time.  Derivation of the three stages of onset  is described in Beiser, Erickson, Fleming, and lacono (1991). Interrater reliabilities, based on judgements about the patient’s age at each of the three stages of onset, range from satisfactory to high. (1993),  As described in Beiser et al.  interrater agreement on the age at the onset of first  noticeable signs, at first psychotic symptoms, and at treatment—seeking for the schizophrenic cases yielded Pearson correlations of  .66,  .97 and  .98, respectively.  Coefficients  of agreement for the affective cases exceeded .90 for all three stages. *  Life Events.  To assess the cumulative effects of  stressors, we used the Holmes and Rahe (1967) approach.  At  intake, 9 months, and 18 months, subjects were interviewed with a list of 33 possible events that comprised the Social Readjustment Rating Scale intake,  (SRRS; Holmes & Rahe,  1967).  participants were asked to provide a brief  At  97  Methods  description of events that had occurred within the year prior to admission, and to indicate the month in which each event occurred.  had  The patient was also asked to provide a  rating of the impact of each event, on a 7—point Likert—type (—3=very negative;  scale  O=no effect;  +3=very positive).  The  occurrence of life events was queried again at the 9- and 18month interview, covering the period of time since the previous interview. At each of intake, nine, and 18 months, the events were classified as to their temporal relationship with any major psychiatric disturbance.  The classification scheme  determined if a reported event was independent or possibly independent of, symptomatology. al.,  or was likely caused by, pre—existing As is the custom in the literature (Day et  1987; Ventura et al.,  ‘possibly independent’  1989),  only the  ‘independent’ and  categories were subjected to  statistical analyses. *  Smooth pursuit eye movement (SPEll) function.  SPEM function  was assessed using the procedures described by lacono and Lykken  (1981), wherein eye movements were recorded while the  participant watched a luminous dot on an oscilloscope screen. The target moved in a sinusoidal fashion, covering twenty degrees of visual angle, at a frequency of 0.4 cycles per second for 20 cycles. use of chin,  Head movement was minimized with the  forehead, and temporal bone rests.  Silver  silver chloride electrodes applied at the outer canthi of both eyes recorded the position of the eyes compared to a  Methods  98  reference electrode attached to the right earlobe.  Scoring  of SPEM accuracy used the root-mean-square error method, whereby a computer program aligns the target and eye position, adjusts for phase and amplitude differences, and then calculates the difference between the two signals. root-mean—square error  (RMSE)  The  is then expressed in standard  deviation units. The RMSE approach has good formal properties. global index of SPEM function,  As a  it has good retest reliability  over two years, shows evidence of heritability, and discriminates schizophrenic from other psychiatric and nonpsychiatric comparison groups Lieberman, *  (Szymanski, Kane, &  1991).  CT assessments.  The CT scans were obtained in the  Radiology Department of the Health Sciences Centre of the University of British Columbia, scanner.  A total of 13—16 slices were taken,  millimetres thick, plane.  using a Siemens Somatom each 8  at an angle parallel to the canthomeatal  Based on noncontrast scans, all measurements were  blind with respect to diagnostic status.  The CT film that  showed the lateral ventricles at their largest was chosen for measurement, and enlarged to 6O of life size.  The areas of  the total brain and the ventricles were determined by tracing with a planimeter the perimeter of the structures. of repeated tracings was used for analyses.  The inter-rater  agreement on the VBR measures was high (intraclass correlation 0.92; Katsanis et al.,  1991).  The mean  99  Methods  Nine—month follow-up measures.  This follow—up interview is  relevant to this thesis to the extent that it was used to derive “cross-sectional”  diagnoses on DSM-III Axis I, and to assess the  occurrence of life events. *  The abbreviated Diagnostic Interview Schedule  (DIS).  The  DIS is a highly structured interview, designed to be administered by lay interviewers Ratcliff,  1981).  (Robins,  Helzer, Croughan, &  For the nine-month follow-up,  only the  sections relating to mania, depression, anxiety, schizophrenia, and paranoia were used. *  Life events. The Holmes and Rahe  (1967)  Social Readjustment  Rating Scale was repeated at the nine-month follow-up, where the occurrence of life events since discharge from the index hospitalization was queried. Eighteen—month follow-up.  This interview replicated the  earlier follow-up interview, with two exceptions. the DIS,  Rather than  a diagnostic interview was conducted by a project  clinician using the PSE.  Supplementary data, regarding treatment  and medication status, occupational and social functioning, and residential status,  as well as a symptom checklist, were also  collected from the index participants. provided collateral information.  At this time,  SOs  Once again, all of the  supplementary data helped to form diagnostic decisions. At this time, case conferences were again held to determine the 9- and 18-month diagnoses.  In addition to deriving these  “cross—sectional” diagnoses, all of the data from the intake and  Methods  100  two short—term follow-up interviews were combined in case conferences to produce a longitudinal, diagnosis.  “best-estimate” Axis I  It is these “best overall” Axis I diagnoses, based on  consensus among project clinicians, that were used to define groups in this project. Five-year follow-up.  Five-year data were collected using  either the full interview or abbreviated procedures.  As  described below, data from each procedure were combined to constitute the dependent measure,  the DSM-III Axis V (highest  level of social and occupational functioning). *  Full five-year interview.  For about 50 of the original  sample,  the full five—year interview was administered in  person.  Similar to the nine-month follow-up assessment,  it  included the abbreviated DIS and the SCL-90 symptom checklist.  Occupational functioning was measured using the  amount of time worked in the past month.  Information about  social functioning was collected by posing three questions regarding socializing with each of friend and family. Specifically, each participant was asked to consider an average week in the past month and indicate:  a) the number  of occasions, b) the number of people, and  the amount of  time spent socializing in that week. residential status,  C)  Data regarding  involvement with treatment services, and  medication status over the previous year were also collected. These variables were also coded with respect to the month prior to the interview.  101  Methods  *  Abbreviated procedures.  Where it was not possible to  administer the full interview, proxy procedures were used. As described earlier, one of the four substitute procedure was comprised of a long-distance telephone call with the participant.  Here, the data reflecting occupational and  social functioning was the same as that derived from the full,  in-person interview.  included:  The other three proxy procedures  (a) a shorter telephone call with a participant  who refused the full interview;  (b) an interview with a  Significant Other who had previously participated in the study; and (c) a review of clinical records.  In the  substitute procedures, the minimum data collected from patients or significant others reflected four dimensions of current functioning  (i.e., the past month):  1. Are you working?  (Please describe.]  2. What is your living situation?  [Please describe.]  3. Are you seeing a doctor or anyone else for emotional problems? 4. *  [Please describe.]  Are you taking any medication?  Five-year Axis V rating.  [Please describe.]  Ratings regarding the highest  level of adaptive functioning at five years reflect occupational and social functioning in the past month. As a global measure, the Axis V ratings as described in 0511III do not disaggregate qualitative from quantitative aspects of either social and occupational functioning. Thus,  there is an assumption that there will be at least a  moderate degree of correspondance between quantitative and  102  Methods  qualitative components within each domain, and between the domains themselves. Because the proxy procedures at five years did not always gather data regarding social functioning, a small modification in the use of the Axis V scale was necessary. In the original 7-point scale, a person with moderate impairment in occupational functioning would merit a rating of either 4  (Fair) or 5 (Poor), depending on the concurrent  social functioning.  Similarly, a person with marked  impairment in occupational functioning would warrant a rating of either 5 social functioning.  (Poor)  or 6  (Very Poor), depending on the  For the cases whose proxy procedures  did not provide data on social functioning intermediate ratings were used,  45),  e.g.,  (approximately 4.5 for  participants with moderate impairment in occupational functioning, and 5.5 for those with marked occupational impairment. The formal properties of the Axis V scale, based on intake and 18-month interviews, are good:  earlier estimates  of interrater agreement resulted in Pearson correlations of 0.89.  Estimates of inter—rater reliability using the  modified Axis V at 5 years are provided below, Results chapter.  in the  103  V.  HYPOTHESES AND ANALYSES  The aim of this dissertation project is to construct an integrated model of the course of schizophrenia. MAP Project sample and research design,  Based on the  this project has the  potential to assess empirically four components of a vulnerability—stress model as it applies to course and outcome. To the degree that the individual components are related to fiveyear outcome, and if the distributional properties of the data are suitable, the components can be combined for simultaneous consideration in predicting outcome.  Hypotheses The specific hypotheses in this project follow from the general aims described in the introductory chapter.  They  include: 1. Social relationships will have a unique and independent role in predicting five-year outcome. 2. Increased numbers of stressful life events will have a modest, negative effect on five-year outcome. 3. Larger ventricles (ventricle to brain ratio; VBR) and smooth pursuit eye movement (SPEll) dysfunction will be associated with poorer five-year outcome for people with schizophrenia, but not affective psychosis. 4. Life events and social relationships will concurrently predict five-year outcome, as will the interaction For social relationships, this is between the two. equivalent to hypothesizing a main and a buffering effect as they interact with life events. 5. Social relationships, life events, VBR, and SPEll dysfunction will make concurrent contributions to the prediction of five-year outcome.  Hypotheses & Analyses  104  Analyses Before outlining specific analyses, some general considerations are in order.  In the prediction of five—year  outcome, the previous literature review has elaborated the bases for expecting relationships between and among the components of the proposed model:  the control variables, the psychosocial and  biological predictor variables, and outcome.  In general, I do  not expect that the relationship between each of the independent measures and five-year outcome will be strong.  Rather,  I expect  that the control and predictor variables will have an important but modest relationship with outcome, since a vast array of factors influence the course of the disorder.  Empirically,  I  expect that each of the modest relationships will be further attenuated by imperfect measurement, small variations in procedure, within-group heterogeneity, and potentially confounding factors that have not been assessed in this study. The general strategy is to use regression techniques to identify unique and independent relationships between predictor variables and outcome. V rating,  The dependent measure is the DSM-III Axis  the highest level of adaptive functioning in the month  prior to the five—year follow—up interview.  The independent  measures are comprised of a set of control variables (which includes the baseline Axis V rating, collected at intake to the study) and the set of predictor variables that represents the domain of interest.  Hypotheses  c  Analyses  Table 10.  105 Summary of Proposed Analyses  Phase 1. Methodological and Descriptive Analyses response rate and differential attrition choosing from multiple data sources at 5-year follow—up formal properties of measures treatment of missing values demographics, premorbid sample characteristics: characteristics, and adaptive functioning at intake and five years, for two groups -  —  -  —  -  Phase 2. The Independent Role of Predictor Variables Three separate hierarchical regression equations, each with For all the 5-year Axis V rating as the dependent measure. equations, Step 1 is comprised of main effects for the predictor Step 2 includes diagnosis and control variables of interest. variables previously identified as predicting outcome Step 3 is comprised of the interactions between (L<.lO). diagnosis and the predictor(s) of interest. Equation 1. Social relationships (N=100) Equation 2. Stressful life events (N=100) Equation 3. Eye movement dysfunction & brain ventricle size (N=63) Phase 3. Data Reduction principal components social relationships (6 variables): factors or two reduce to one analysis to sum the no. of independent & life events (6 variables): at intake, 9— & 18-months independent events possibly sex, SES, length of prodromal onset, control variables (age at multiple regression of these, period, baseline Axis V): and diagnosis, to identify strongest predictors —  —  -  Phase 4. A Stress-Process Model for Each Diagnostic GrouD One hierarchical regression equation (N=100): Step 1. Main effects for social relationships, life events, and baseline functioning Step 2. Interactions between each social relationship and life events variable Step 3. Main effects for diagnosis and baseline Axis V Step 4. Interactions between diagnosis and each of social relationship and life events variables social Step 5. Second-order interactions: diagnosis events by life relationships by Phase 5. Concurrent Influences of Social and Biological Factors One hierarchical regression (N=63): Step 1. Main effects of social relationships, life events, VBR and SPEM Step 2. Diagnosis diagnosis with social relationships, Step 3. Interactions: life events, VBR and SPEM  Hypotheses £ Analyses  106  The data analysis is based on a sequential strategy: summary of the five phases is included as Table 10.  a  In the first  phase, methodological and descriptive results are presented. Next, the hypotheses regarding the unique and independent contributions of the each of the predictor variables are assessed.  The third phase,  reduction.  The fourth,  if required,  is devoted to data  if required, assesses a stress—process  model within the psychosocial domain.  Finally,  if warranted, the  psychosocial and biological variables will be concurrently assessed in a combined vulnerability-stress model.  Phase One:  Methodological and Descriptive Analyses  The methodological issues considered in this thesis pertain to the aspects of the project that have not previously been examined in the MAP Project.  They include assessment of the formal  properties of measures,  the response rate at five years, and a  consideration of the sources of data used in the five-year follow—up. In previous studies within the MAP Project, schizophrenic and affective psychosis patients showed dramatic differences in functioning,  on a number of measures.  Since both groups will be  combined for some of the predictive analyses, descriptive data will be presented to provide a sense of group differences in adaptive functioning at intake to the project and five years later.  In this way, the different steps in the hierarchical  regressions that include diagnostic status will be more comprehensible.  Hypotheses  Phase Two:  c  Analyses  107  Establishing the Independent Role of Predictors  This phase of analysis consists of several hierarchical regression equations, where the first three hypotheses  (stated  below) are evaluated. As preliminary steps, descriptive data are presented for the two diagnostic groups for each predictor of interest. Demographic characteristics are also presented.  In so doing, the  reader will have a broad understanding of the manner in which these will affect multivariate predictions. The selection of control variables for entry in the regression equation was determined by further preliminary analyses.  Here,  any control variable that was correlated with five-year outcome was retained for the regression.  A liberal alpha level  (p<.10)  was used to select control variables. For each of the hypothesized predictors,  the five-year Axis V  rating serves as the dependent measure in a hierarchical regression equation.  In general, the regressions are based on  both diagnostic groups combined. Social relationships will have a unique Hypothesis One. and independent role in predicting five-year outcome. The first of the preliminary analyses pertaining to this hypothesis was an assessment of diagnostic differences in the social relationships at intake.  Next, correlations provided an  indication of the extent to which social networks,  social  resources and the perception of social support are related to five-year outcome, by diagnosis and for both groups combined.  A  Hypotheses  Analyses  108  further preliminary analysis assessed the ability of the six social relationship variables to predict outcome, without controlling for potentially confounding factors. the relative importance of social network,  In this way,  social resources, and  perceived social support can be assessed, prior to the addition of control variables. In assessing the hypothesis that social relationships will have a unique and independent role in predicting five-year outcome, the six social relationship variables were entered in a hierarchical regression as Step One.  Diagnostic status,  accompanied by control variables identified in preliminary analyses, comprised Step Two.  The interaction between diagnosis  and each of the six social relationship variables will comprise Step Three.  If preliminary analyses suggest that any correlation  between a control variable and outcome is substantially different for the two diagnostic groups, then a further interaction term may be warranted. Increased numbers of life events will Hypothesis Two. have a modest, negative effect on five-year outcome. The preliminary analyses again dealt with diagnostic differences in the predictor(s) of interest,  in this case the six  variables reflecting independent and possibly independent events at each of intake, nine, and 18 months.  Preliminary correlations  between the six life event measures and five—year Axis V ratings were computed for each diagnostic group and for both groups combined.  Hypotheses £ Analyses  If significant predictive relationships are apparent in these preliminary analyses, then the life events variables can be entered into a regression equation where the DSM III Axis V rating would again serve as the dependent measure. be comprised of six life event terms.  Step 1 would  Step 2 would again add  diagnostic status and control variables identified in the preliminary analyses.  Step 3 would include six interaction  terms: diagnosis by each of the life events variables. Larger brain ventricles and SPEll Hypothesis Three. dysfunction will be associated with poorer outcome for people with schizophrenia, but not affective psychosis. The psychosocial factors were the subject of two separate equations, because of the large number of variables within each domain.  Such is not the case in the biological sphere:  the vulnerability factors is represented by one variable.  each of As  such, the unique and independent role of the two variables can be assessed simultaneously,  in one regression equation.  Preliminary bivariate analyses would assess any relationship between outcome and each of the two biological vulnerability indicators.  If the bivariate analyses indicate that VBR and SPEM  predict outcome, a hierarchical regression equation can be used where VBR and SPEM measures comprise Step 1.  Here, the SPEM  variable may be dichotomized into poor— and good-tracking groups. Step 2 would include diagnosis and control variables identified in preliminary analyses.  The results from Step 3, comprised of  terms representing interactions with diagnosis, would determine if the diagnosis-specific aspect of the hypothesis is correct.  Hypotheses & Analyses  Phase Three:  110  Data Reduction  Although the sample size is reasonably large, there are nonetheless a limited number of variables that can be considered in a multiple regression equation.  If the analyses in Phase Two  indicate a larger number of significant predictor variables from the psychosocial and biological domains than can be simultaneously assessed in an hierarchical regression, then the following data reduction techniques may be required. There are six measures of social relationships.  If  necessary, they may be subjected to principal component analyses (PCA) to identify a small number of dimensions to use in (later) predictive equations.  If the six variables show similar  interrelationships among the two diagnostic groups, then the one or two components capturing the most variance would be retained for use in subsequent predictor equations.  If the factor  structure for the two diagnostic groups is not similar,  then  separate PCA analyses can be conducted. In the life events domain, data reduction is a simpler task. If required, the six measures of life events can be summed to assess their joint relationship with five-year outcome. A third strategy may be necessary for control variables.  If  preliminary bivariate analyses indicate that a large number of control variables are significantly associated with five-year outcome,  then they may be entered into a separate regression  equation to identify those that make independent contributions vs.  those that are redundant.  The control variables that achieve  Hypotheses & Analyses  ill  status as independent predictors of five-year outcome may be retained for inclusion in Phase Three or Four.  Phase Four:  Assessing a Stress—Process Model  Analyses in Phases Three and Four are contingent on results that support the earlier hypotheses.  In particular,  the presence  of a significant relationship between life events and outcome is crucial to both stress-process and diathesis—stress models. Life events and social relationships Hypothesis Four. will concurrently predict five-year outcome, as will the For social relationships, interaction between the two. a main and a to hypothesizing equivalent this is with life events. interact effect as they buffering The test of this hypothesis requires three sets of predictor variables:  one representing social relationships, one for the  cumulative LE rating, and one for the interaction term.  The  exact number of variables in each set cannot be specified in advance of the results from the Data Reduction phase. Previous regression analyses were based on all cases combined, and included first—order interaction terms involving diagnostic status by diagnosis).  (life events by diagnosis; social relationships  Since the assessment of a stress—process model  requires terms representing the interaction between life events and social relationships,  any diagnosis—specific tests requires  second-order interaction terms  (life events by social  relationships by diagnosis). With these considerations in mind,  the hypothesis concerning a  stress—process model would again use a hierarchical regression  Hypotheses  Lc  Analyses  112  strategy, with five-year Axis V as the dependent measure.  Step 1  would include social relationship and life events variables, Step 2 would consider the psychosocial elements. baseline Axis V.  (first-order)  and  interaction between the  Step 3 would add diagnostic status and  Step 4 would be comprised of terms representing  the interaction between diagnosis and each of the social relationships and life events terms.  Finally,  Step 5 would be  comprised of second—order interaction terms.  Phase Five:  Assessing Concurrently Social and Biological Factors  It is not possible to anticipate the specific analyses at this stage, since the components are largely influenced by the significance of earlier results.  As such, what appears below is  but one example whereby the final hypothesis may be tested. Social relationships, life events, VBR, Hypothesis Five. and SPEW dysfunction will make concurrent contributions to the prediction of five—year outcome. If results from Phase Two indicate significant relationships between each of the four predictor elements and outcome,  then  Step 1 of an hierarchical regression might include five elements representing main effects:  social relationships (two variables,  from data reduction phase),  life events, VBR, and SPEM.  would include diagnosis,  and Step 3 might include interactions  between diagnosis and the five terms from Step 1. predictors,  Step 2  These  together with the dependent measure, suggest 12  variables in the Phase Three prediction equation——prior to the addition of any control variables.  Since the number of variables  Hypotheses & Analyses  113  is near the upper limit (63 cases:  12 variables), no control  variables can be added in this analytic scenario.  Formal Statistical Considerations Two prominent statistical considerations are in order.  The  first relates to the inflation of the potential for Type I error, due to the presence of multiple parametric procedures, viz, multiple regression equations.  five  To control for Type I error in  the predictive analyses, the Bonferroni adjustment is used to set the alpha level at  .01.  The second statistical consideration relates to power. Since the sample size is already set,  I have determined the  probability of detecting an effect size of =.30.  This modest  magnitude of effect is chosen because it is consistent with many of the correlations described in the relevant literatures. on one-tailed correlations and an alpha of  Based  .01, the power in the  proposed analyses varies as a function of the number of available cases.  For equations involving social relationships and life  events variables, the number of cases will be about 100: power is  .76.  here,  For equations involving eye-tracking function,  where approximately 95 cases are present, power is  .74.  Finally,  in the case of equations involving CT scan data (N=65),  the power  is approximately .55.  These calculations show that,  the four sets of predictors,  the level of statistical power is  adequate to detect correlations of power is less than optimal.  for three of  .30.  For the fourth, VBR,  114  VI.  RESULTS  Before describing the predictive role of social relationships,  life events, and biological vulnerability, some  attention must be devoted to methodological issues and preliminary, descriptive results. addresses several methodological reports descriptive results,  Phase One, issues.  immediately below,  This section also  including the demographic  characteristics of the two diagnostic groups,  as well as their  adaptive functioning at baseline and over the five-year course of illness.  Results in Phase Two consider the predictive role of  each of social relationships, vulnerability indices.  life events, and biological  As will become apparent, the lack of  support for hypotheses regarding the independent role of some of the predictors precludes the need for progression to Phases Three, Four, and Five, where stress—process and vulnerablity stress models would have been considered. To control for the inflation of Type I error, the Bonferroni correction was applied to the analyses used to assess the five hypotheses pertaining to the prediction of outcome.  Here,  five  tests are used; thus, the alpha level for statistical significance is  .01.  In the predictive analyses,  identified only if they meet the  trends are  .05 alpha criterion.  When  considering preliminary and descriptive issues, no correction is applied:  the usual alpha level of  .05 is used.  115  Results  Phase One: Methodological and Descriptive Results Response Rate Data were considered eligible for the five year follow—up if they pertained to the period between 3.5 and 7.0 years after intake to the study.  This eligibility period was selected to be  within two years of the fifth anniversary of the patients’ psychotic episode;  first  3.5 years was used rather than 3.0 years so as  to be nearer the fifth anniversary than the 18-month follow-up. Of the total MAP Project sample other psychoses), participants,  (schizophrenic, affective, and  outcome is known for 133 of the surviving 168  for a response rate of 79.2%.  Seven participants  died during the five years following onset of illness: considers death an analyzable outcome, was 80.0% (140/175). available,  11  if one  the total response rate  Of the 35 cases about whom data were not  (31%) refused to respond and 24  (69%) could not be  located. For the two diagnostic groups relevant to this project, there was no difference in response rate: available on 54  (79.4%)  and 55  (76.4%)  outcome data are  of surviving schizophrenic  and affective psychosis participants, respectively.  There was no  differential attrition based on age at onset of illness, duration of the prodromal period,  or socioeconomic status.  Finally, there  were no differences in baseline adaptive functioning, viz, the Axis V rating prior to intake.  Overall,  it appears that the  schizophrenic and affective disorder patients for whom follow-up data were successfully collected can be taken as representative of the entire  (surviving) sample.  116  Results  As described in the preceding Methods chapter,  four avenues  were available to assess five—year outcome status: a) the full interview; b) an abbreviated interview by telephone; c) collateral informants  (i.e. a significant other); and  d) medical records, which could reflect either psychiatric or nonpsychiatric care. Inherent in the use of archival sources is a a potential for bias towards severity:  Poor-outcome cases present to clinics and  in-patient wards more frequently than do good-outcome cases.  In  this study, however, there was an opportunity to minimize this bias.  When the only follow-up data available for this study came  from such biased sources there was little choice but to use them, understanding their limitations.  For this study, the  ‘biased’  designation is used when medical records described care given for mental health purposes. nonpsychiatric care,  e.g.  Medical records reflecting emergency room visit for sprained  ankle, are not considered biased. Frequently, however, data from multiple sources are available. In such cases  (n=38,  or 29), a set of rules is needed to  determine a priori which source to use for data analysis. considerations are relevant in generating the rules:  Two  the date of  data collection, and whether data came from biased or unbiased sources.  First, data are designated as  ‘on due-date’  pertained to a period within three months of the fifth anniversary of intake to the study;  if riot, they are  if they  117  Results  Table 11.  Use of Rules in Selecting for Analyses Data From Multiple Sources Reflecting Five—Year Outcome.  Rule 1. On before off due date. 2. Unbiased before biased source. 3. Same time period: combine. 4. Use closest to due date.  TOTAL  % of cases with multiple n 5-yr. data  % of all cases with 5-yr. data  15  38.5  10  25.6  5  12.8  3.8  9  23.1  6.8  39  100.0  11.3 7.5  29.4  118  Results  designated as  ‘off due-date’.  as explained above.  The second consideration is bias,  The following rules, then,  form a hierarchy:  1. Use data on due-date before off due-date; 2.  If all data are off due-date, use arbitrary (unbiased) before biased data;  3.  If all data are off due-date, and both arbitrary and biased sources describe same time period, combine data;  4.  If all data are off due—date, and multiple data sets come from arbitrary or biased sources, choose data closest to due date.  Table 11 shows the use of these rules in the 39 cases with multiple sources of follow-up data.  The use of these rules  resulted in 65.4 of follow-up data coming from the index participant:  71 people completed the full interview,  14 the  abbreviated interview, and two patients had their in—person data supplemented by other sources. used in 216 of the cases minimum.  Psychiatric clinical records were  (n=28), keeping potential bias to a  A full description of the sources of five-year follow-  up data used for analysis is shown in Table 12. The issue of potential vs.  real bias as a function of the  source of follow—up data can be addressed by contrasting those cases whose outcome status was determined exclusively through the use of psychiatric records with those whose status was determined from other sources of data.  Here, the extent of bias cannot be  known fully; rather, we can only assess empirically indicators that may reflect either bias or true differences between patients who are more frequently in hospital.  For example,  if there are  119  Resul ts  Table 12.  Sources of Five—year Follow-up Data Used for Analyses.  Source of data 1. Full interview (index participant) 2. Abbreviated interview (index participant) 3. Significant other 4. Psychiatric records 5. Nonpsychiatric records 6. Multiple sourcesa re same time period Deceased 7. 8. No information TOTAL  n  96 of surviving 5-year cases  96 of intake cases  71  53.4  40.6  14  10.5  8.0  7 28 6 7  5.3 21.1 4.5 5.3  4.0 16.0 3.4 4.0  7 35  N.A. N.A.  175  100.0  4.0 20.0 100.0  f the seven cases with multiple sets of data describing five— 0 a year outcome, two included abbreviated data from the patient.  120  Results  archival/non-archival group differences in outcome, but none in terms of baseline functioning,  illness characteristics,  or  premorbid status, then one could infer that bias indeed may be present.  On the other hand,  if differences in outcome were  parallel to pre—existing baseline differences, then one could infer that those receiving care were the more disturbed subgroup. In the combined schizophrenic and affective psychosis sample, the five-year outcome status of 21 cases psychiatric records alone. ratings at five years  These patients have lower Axis V s.d.:  (mean  remaining 88 “non—archival” cases p<.OO1).  is based on  (19.3%)  3.3 + 1.3), compared to the  (4.6 j 1.3; =4.41, df=107,  But important to the assessment of bias is that they  also had lower baseline levels of adaptive functioning (mean intake Axis V ratings ÷ s.d.: =3.04, df=107, p<.OO ). 5  3.33 ±. 1.24 vs.  4.20 j. 1.17,  The lower baseline functioning is not  because the “archival” cases are more severely ill at intake: illness characteristics such as symptom severity, the length of the prodromal phase,  and duration of the initial hospitalization  are similar for the two groups.  Neither have the “archival”  cases always been labouring at a disadvantage:  educational  achievement and premorbid social functioning are strikingly similar for the archival and non-archival groups. A second way to assess bias is to examine the proportion of participants whose outcome is assessed by means other than psychiatric records.  Of these  ‘nonarchival’  patients, 66% were  receiving psychiatric treatment at the time of the five—year follow-up.  This suggests that,  if the 21  ‘archival’ cases had  121  Results  been ascertained by other means, been in treatment.  66% or 14 patients would have  Thus, at most,  the use of psychiatric records  added seven patients, or 6.4% of the combined schizophrenic and affective sample, who otherwise would not have been located. Overall, these results indicate that a strong bias is not present in the use of psychiatric records as an indicator of outcome.  Formal Properties of Measures Inter—rater reliability estimates for outcome measures were based on 20 cases.  Ratings were made,  in the spring of 1992, by  the writer and a psychiatrist* at the Clarke Institute of Psychiatry’s Division of Community, Culture and Health Studies. Since the global rating was derived using a 7-point ordinal-level scale with intermediate ranks possible (e.g. intraclass correlation coefficient reliability.  2.5,  4.5), an  (ICC) was used to estimate  The result is an ICC of 0.95,  indicating a high  level of agreement between the raters. Reliability of the classification of stressful life events was assessed by this writer and a doctoral candidate** in Clinical Psychology at the University of Ottawa. period,  Following a training  17 cases were randomly selected for inter—rater  reliability estimates.  Intraclass correlation coefficients were  based on the number of “independent” events per case, combining events across intake, nine—month, and 18-month interviews.  * **  Jiahui Zhang, M.B., M.Sc. Dianne E.  Chappell, N.A.  The  122  Results  ICC derived from this procedure was 0.73,  indicating reasonably  good agreement between raters. In a previous paper acceptable levels of  (Erickson et al.,  1989), we demonstrated  internal consistency for the four social  relationships scales.  There, Cronbach’s alpha coefficients were  calculated for the Availability and Adequacy of each of the Acquaintances and Confidants measures.  At that time, we did not  offer evidence for the social network measures.  Upon reflection,  it seems that further assessment of the formal properties of the six social relationships indicators members of the network)  is possible.  (four scales, two tallies of Specifically, one can  assess the convergent and discriminant validity by conducting a principal components factor analysis of the six indicators.  In  doing so, we can begin with an inspection of the correlation matrices for the two diagnostic groups, shown in Table 13. Table 13 shows that the magnitude of the inter-relationships among the social relationship measures is moderate. some convergence,  of the kind that one would expect:  correlations between the two groups are  .44 and  .52.  ‘adequacy’  There is The  measures for the two  There is less convergence than one would  expect in the way in which the number of nonkin is correlated to either the availability of friends or the availability of acquaintances among the schizophrenic group. the affective psychosis group the strongly correlated with the  ‘nonkin’  Conversely, among  variable is more  ‘availability of confidants’ scale.  Overall, the moderate degree of convergence among the six measures is adequate for factor analysis.  123  Results  Table 13. Intercorrelations Among Social Relationship Measures for Two Diagnostic Groups (schizophrenic patients below diagonal; affective psychosis patients above). No. of kin  No. of Kin  No. of nonkin  .21  Avail.of acquaint,  Adeq.of acquaint.  Adeq.of confid.  .13  .06  .28  .23  .13  —.10  .27  .00  —.25  .50  —.07  .00  .44  No. of Nonkin  .09  Avail, of Acquaint.  .10  .11  Adeq. of Acquaint.  —.01  .30  .19  Avail.of Confid.  .25  .53  .25  .33  Adeq. of Confid.  .01  .25  .17  .52  Note.  Avail.of confid.  .01 .20  Sample sizes: schizophrenic and affective groups each N=66.  124  Results  Other indicators of suitability for factor analysis are evident in the correlation matrices shown in Table 15:  There is  reasonable similarity in the two matrices, there is a reasonable proportion of coefficients in the L=. —. 50 range, 25  and there is  at least one coefficient of a reasonable magnitude for each variable. Whether the diagnostic groups are combined or considered separately, other indicators suggest that these data are acceptable for factor analysis.  The partial correlations are  small, the Kaiser-Meyer-Olkin measures of sampling adequacy (Kaiser,  1974) for the two groups are in the 0.60—.70 range, and  communality estimates are good. method of extraction,  Using the principal components  a two-factor solution is optimal  (either  for all cases combined, or separately for the two groups).  The  factor loadings displayed in Table 14 are based on oblique rotation procedures.  The results of the factor analysis, whether  based on all cases combined or separately for the two diagnostic groups,  indicate convergent and discriminant validity for the six  social relationship indicators. The results of the principal component analyses indicate two clear factors, whether derived from all cases combined or from diagnosis-specific analyses.  In each case, the “meta—factors”  are clear, and account for 59-60% of the variance.  Although the  factors were not constrained to be orthogonal, there is little correlation between the two factors in either patient group (schizophrenic sample =-.16; affective psychosis sample =.05).  24.8%  34.7% ——  19.6% ——  39.2%  22.2%  37.8%  ——  .85  —.08 .67  —.77 .25  .71  .84  .05  .72  .85  .04 .72 —.81  .07  67  .88  —.04  .75  —.08 .80  .63 —.13  • 61  .42  —.08  .80  .63  .87  .75 —.08  .83  .72  .04  .87  .78  .04  -.15 .58  .35  —.22  .68  .55  -.02  .65  .44  .31  Qual Ity  .57  Quantity  LOADINGS  .44  Qual ity  COMMUNALITY  Affective psychosis pts.  .45  .58  .45  .05  .52  .28  Quantity  LOADINGS  Qual— ity  COMMUNALITY  Quantity  LOADINGS  Schizophrenic patients  schizophrenic and affective groups each Sample sizes: Notes. Factor analysis based on principal components extraction N=66. and oblique rotation.  % variance  No. of kin in network No. of nonkin In network Avail, of confidants Avail, of acquaint. Adeq. of confidants Adeq. of acquaint.  Social r’ships indicator  COMMUNALITY  All cases  Factor Analyses of Social Relationships Measures: Table 14. Communality Estimates and Loadings for ‘Quantity’ and ‘Quality’ Factors for All Cases Combined and by Diagnosis.  U’  Results  126  As indicated in bold type in Table 14, the Quantity dimension is comprised of four composite indicators: of kin and nonkin in the social network, confidants and acquaintances. of the two adequacy scales.  and the availability of  The Quality dimension is comprised The only perfidious variable is the  ‘number of kin in the social network’: diagnosis,  the number  When considered by  it loads on both the Quantity and Quality dimensions  for each patient group. Within the Quality dimension,  the number of family variable  has a diagnosis—specific characteristic.  For schizophrenics, the  number of family in the network is negatively associated with other quality indicators.  In other words, schizophrenic people  seem to be either satisfied with family and dissatisfied with people outside the family,  or the reverse.  Affective psychosis  patients, on the other hand, report that they are generally satisfied with both kin and nonkin or generally dissatisfied with both. Overall, these principal component analyses are evidence for the validity of the six social relationship measures used in this study.  They indicate the convergence and divergence of  quantitative and qualitative aspects of social relationships, supporting the conceptual distinction between social resources and social support.  In particular, they offer support for  validity of the two social network variables, whose formal properties have not been previously assessed.  127  Results  Treatment of Missing Values In studies such as this one, a small proportion of missing data for each of a large number of variables can have a considerable effect on analyses.  One way of minimizing this  disadvantage is to use mean substitutions or similar procedures, provided only a small proportion of values are affected. In the first of the three broad categories of predictors, social relationships, data are missing for two reasons.  The  first reason includes instances where a respondent failed to answer a small number of questions on the ISSI, resulting in missing values for one or two scales.  Here, no more than 4.5?6 of  cases are missing on any of the six 1551 variables.  It is these  values which were replaced by the respective mean for the diagnostic group. participants  The second reason for missing data is that  (14 of 146,  questionnaire.  or 9.696)  failed to complete the ISSI  Because of the relatively large number of cases,  the ISSI data were left missing on these 14 cases. A similar approach was used in replacing missing values for the stressful life events data.  Missing data were not replaced  if all of the three life events interviews 18 months) were missed.  (intake, nine months,  For a number of cases, however,  two sets of life events data were available:  Here, the partial  missing data can be replaced using z—score equivalents. example,  one or  For  if the intake life events total is missing, we can  calculate the z—score equivalent for the number of life events at 9 and 18 months, intake value  and use the average z—score to calculate the  (based on mean and standard deviation of intake  128  Results  values for the particular diagnostic group). used to replace missing data in 10 (5.696) at nine months, and 15 In the biological domain, indicated.  This procedure was  (6.9%) cases at intake,  (10.4%)  8  at 18 months.  one minor substitution procedure is  Three of the participants who completed the CT  assessment are missing eye—tracking data;  for these cases the  mean RMS error value for the respective diagnostic group is used. The final considerations with respect to replacement of missing values refer to two control variables, the age at onset of psychotic symptoms, and the duration of the prodromal period. For the first, onset value:  14 of 146 cases  (9.6%) are missing the age at  The replacement strategy here is to use the  participant’s age at first treatment—seeking minus the diagnostically-appropriate median treatment lag-time.  For the  patients whose date of onset we were able to determine, the lag— time is the period between onset of prominent psychotic symptoms and the first contact with a treatment facility.  For the second  control variable, the duration of the prodromal period, (9.6%) are missing;  14 cases  the substitution uses the median for the  relevant diagnostic group.  Descriptive Results:  Baseline and Outcome  Table 15 describes the groups in terms of their demographic characteristics and their baseline level of functioning.  Results  describing social relationships, stressful life events, and biological vulnerability will be presented later, the predictive results.  just prior to  129  Results  Table 15. Sample Characteristics: Functioning by Diagnosis.  Age at intake  (yrs.)  Age at onset of psychotic symptoms Duration of prodromal phaseb (weeks) Socioeconomic status (Blishen scale) Baseline Axis V (1=grossly impaired; 7=superior)  Demographics and Baseline  Schizophr. Mean (S.D.)  Affective psychosis Mean (S.D.)  22.7  (5.3)  26.1  (8.0)  3.02  .005  21.6  (5.1)  25.9  (8.0)  3.84  .001  111  (141)  104  (177)  .04  n.s.  46.6  (16.1)  43.5  (13.7)  1.28  n.s.  4.5  (1.1)  5.54  3.5  (1.2)  ta  prob  .001  bFor the ‘duration of prodrome’ aAll t tests based on 144 d.f. variable, the mean and standard deviation displayed are the actual values, although the test of group differences is based on Van der Waerden scores (normalized ranks).  130  Results  Considering all the baseline characteristics simultaneously, a preliminary multivariate analysis of variance (MANOVA) indicates that the two diagnostic groups are different (Hotelling’s T =9.44, df=5,140, p<.O0l). 2 results from subsequent univariate  Table 15 shows the  tests.  It shows that the  affective psychosis participants are slightly but significantly older than the schizophrenic patients at intake to the study and at the onset of psychotic symptoms.  The proportion of men and  women is also different in the two groups:  men comprise 77.8€  of  the schizophrenia group, compared to 60.86 of the affective psychosis group (Chi—square=4.93, df=1, p<.O5).  There is no  (data not shown) or  difference in the level of education attained the socioeconomic background.  The higher level of functioning in the affective psychosis group at intake is also apparent at five years.  Figure 2  suggests that there is a parallelism in the Axis V ratings of the two diagnostic groups over time.  This observation is upheld by a  repeated measures analysis of variance. subjects factor  With one between-  (diagnosis) and one within-subjects factor  levels of time), there is a main effect for diagnosis df=1,107 p<.OO1),  a main effect for time  (two  (F=44.1l,  (F=5.90, df=1,107,  ), and no diagnosis by time interaction 2 p=.O  (F=.05, df=1,107,  p=n.s.). A more clinically-oriented way to describe global functioning is in terms of three categories. ‘Impaired’  The category of  includes persons with marked impairment in either work  or social functioning (Axis V ratings of 1-3).  The second  -o  N  U) C)  tI)  -o  CD  1<  CD C)  >  CD  3  H  (1)  -‘  (V  -n  CD  0 1’)  0) C 1 CD  rt  01  ()  CD  (a  DSM-III Axis V  132  Results  category,  ‘Compromised’, describes patients who, despite a  moderate degree of impairment in their social and occupational functioning, are managing to continue with their activities V=3.5-4.5).  Finally,  the category of  (Axis  ‘Functioning’ describes  people whose Axis V ratings of the quanitity of either social or occupational functioning is 5 or greater.  The proportion of  schizophrenic and affective psychosis patients falling into these categories in the periods leading up to the intake and five-year interviews are shown in Table 16. Use of the three categories indicates that approximately one—fifth of people with schizophrenia are either of the assessment periods. one—third and one-half are in the proportions are  ‘Compromised’.  ‘Functioning’ at  At the opposite end, between ‘Impaired’ category.  Similar  Among affective patients almost  the reverse seems to be the case, with a small minority (15—26%) in the  ‘Impaired’ category, and the majority (57-60%)  in the  ‘Functioning’ group.  Phase Two:  The Independent Role of Predictor Variables  Prior to assessing the hypotheses, Table 17 illustrates the zero-order correlations between baseline characteristics and five-year outcome, correlations that may overlap with the predictor variables of interest.  133  Results  Proportions of Patients in Three Categories of Table 16.. Adaptive Functioning Over Time, by Diagnosis.  Affective Psychosis  Schizophrenia Impaired  Compromised  Funct— ioning  Impaired  Compromised  Funct ioning  ()  48  32  20  16  27  57  18 mos.(%)  63  20  16  26  28  46  5 yrs.  32  46  22  15  25  60  Intake  ()  for Sample sizes at intake and five years: Note. affective (respectively); for schizophrenics, Ns=72 and 54 (respectively). patients, Ns=74 and 55  134  Results  Correlations Between Five-Year Outcome and Demographic Table 17. Characteristics and Baseline Functioning. Correlation with five-year Axis V rating Affective Psychosis All cases Schizophrenics (n=109) (n=55) (n=54) onset of (yrs.): psychotic symptoms Sex (1=male;2=female)  .03  Duration of prodromal phasea (weeks) Duration of treatment lag—time (weeks) Socioeconomic status (Blishen scale) Education (10—point ordinal scale)  .13  Baseline Axis V (1=grossly impaired; 7=super ior)  .20k  Age  .22  .16  **  —.04  .11 .17k **  .l4 —.23  .08  -.34  .12  —.15  —.05  .16  —.07  .08  .21k  **  37**  &For the ‘duration of prodrome’ variable, the correlation is computedbased oVan der Waerden scores (normalized ranks). p<.O5 p<.Ol p<.10  135  Results  Table 17 shows that characteristics observable at intake to the proiect have at best a modest relationship to long-term outcome.  Indeed, within the schizophrenic group,  two factors——  age at onset and duration of the prodromal period--that have achieved prominence as robust predictors of outcome in schizophrenia  (e.g.  Stoffelmayr et al.,  1983) show little  relationship to five-year outcome in this sample (L=.O 3 and =.13, respectively).  In the affective group,  the only  correlation coefficient that clearly indicates a significant association with outcome is sex (L=—. , 32  p<.O1), where men are  functioning better than women at five years*. In Table 17, two baseline characteristics predict outcome for all cases combined:  age at onset and intake Axis V ratings.  both correlations, however,  two concomitant effects are present.  For each of the two variables there are: intake; and b)  For  a) group differences at  lower correlation coefficients with outcome when  each group is considered separately.  These observations render  the higher “all cases combined” correlations a product of the differences between the two groups under consideration. does not compromise validity; rather,  *  This  it shows that diagnostic  When the affective psychosis group is subdivided by diagnosis, the men in the Bipolar group are functioning at a higher level 5.3 j. 1.0 vs. 4.2 16, respectively). than the women (H ± SD: In the Major Depressive group, however, the men and women are SD: men 5.3 ±. 1.2; women 5.0 functioning at similar levels (H Thus, the negative correlation between sex and 5-year 1.2). outcome among the entire affective psychosis group is likely the result of the lower functioning of the Bipolar women.  136  Results  status must be considered when analyses are based on all cases combined.  Social Relationships In this section, group differences are presented first.  They  are followed by simple predictive correlations, and then complex predictive analyses. As described elsewhere  (Erickson et al.,  1989), substantial  group differences in supportive social relationships were apparent at intake to the study. pertaining to social networks,  Table 18 displays results  resources, and perceived support,  but first a reminder about scale construction is in order. As described previously, values for the each of the four scales pertaining to social resources and perceived support were derived from the sum of standardized item scores. Standardization, based on the distribution from an age- and sex— matched normal comparison group, was necessary because the items within each scale were expressed in different metrics. Table 18 the  Thus,  in  ‘adequacy of close and confiding relationships’  entry for schizophrenic patients indicates a mean of 1.4 standard units below that of the normal controls. The data that appear in Table 18 for the two social network variables are not transformed:  they are the actual number of  people in the networks of each patient group.  For comparative  purposes, the social network of the normal comparison group was comprised of  (mean ± s.d)  2.9 family members  6.3 + 4.6 nonfamily members, and 4.6 ±.  (Erickson et al.,  1989).  Values for the  137  Results  Social Networks, Social Resources, and Perceived Table 18. Support at Intake for Two Groups.  Schizophrenic subjects (n=66)  Social network a) No. of kin b) No. of nonkin Social resourcesC a) Avail, of close & confiding rships b) Avail, of acquaintances Perceived a) Adeq. of close & confiding rships b) Adeq. of acquaintances  Affective psychosis subjects (n=66)  Fa  prob  mean  s.d.  mean  s.d.  3.8 3.8  2.6 3.0  4.6 5.7  2.9 4.9  2.92 7.39  —4.7  2.8  —1.4  4.2  27.48  .001  —3.2  2.4  —1.0  4.0  14.80  .001  —1.4  3.5  —0.8  4.6  .59  ns  —1.8  3.7  —0.3  3.5  5.81  .02  .09 .01  aAll univariate F—tests based on 1,130 degrees of freedom. bsocial network values displayed in the table are the number of CData displayed are standardized scale people in the network. values, based on a matched normal comparison group (see text).  138  Results  normal comparison group on the social resources and perceived support scales are as follows  (M  s.d.):  and confiding relationships 0.0 j. 3.9;  availability of close  availability of  acquaintances 0.1 ± 4.4; adequacy of close and confiding relationships 0.0 j 3.8; adequacy of acquaintances 0.1 ±. 3.4. Considering all six measures of social relationships simultaneously, significant differences were apparent between the two patient groups (Hotelling’s T =7.19, df=6,125, p<.OO1). 2 Subsequent univariate analyses of variance  (ANOVA5)  for three  quantitative indicators, number of nonkin in the network, availability of confidants, and availability of acquaintances, indicate that the schizophrenic group had fewer supportive social relationships, compared to the affective psychosis group, prior A similar, though nonsignificant,  to intake.  trend was evident  for the number of family members in the social network. On the qualitative side, schizophrenic patients are less satisfied with their more distal relationships.  No significant  difference is apparent in the level of satisfaction with confidants. Predicting outcome.  The foregoing results are informative  but are not directly related to the hypothesis at hand, viz, that supportive social relationships will make a difference in outcome.  Table 19 displays correlations between the five-year  Axis V rating and each of the six social relationship measures. Table 19  indicates that the zero—order correlations between  social relationship ratings and five—year outcome are, at best,  139  Results  Correlations Between Five-Year Outcome and Six Ratings Table 19. Relationships, for Two Groups and All Cases Combined. of Social  Schizophrenics (n=50) Social networka No. of kin No. of nonkin Social resource? a) Avail, of close & confiding rships b) Avail, of acquaintances Perceived supportb a) Adeq. of close & confiding rships b) Adeq. of acquaintances  Affective Psychosis (n=50)  All cases (n=100)  —.19k .31  .13 .02  .08  .20  .19  .08  .16  .10  .20  .11  .07  .17  .O6 .21  * *  aSocial network values displayed in the table are the number of bata displayed under Social Resources people in the network. and Perceived Support categories are standardized scale values, based on a matched normal comparison group (see text). *  p<.05  **  p<.Ol  140  Resul ts  modest in magnitude. magnitude,  A number of others are marginal in  but nonetheless are in the expected direction.  Specific significant findings indicate that, combined,  the availability of both confidants and acquaintances  is associated with five-year outcome. reflect,  for all patients  These correlations likely  in part, diagnostic differences  in the two variables,  since affective psychosis patients have both more social relationships and better outcome.  Within the schizophrenia  group, there is a nonsignificant trend for a larger number of nonfamily members of the social network to be associated with better outcome (=.31, p<.O5). The collective role of social relationships.  What about  the Joint role of the various indicators of social relationships? There is, after all,  overlap among the six measures.  For  example, the correlation between the two “quality” measures (Adequacy of Acquaintances, and Adequacy of Close and Confiding Relationships)  is L=. . 47  Similarly, the mean correlation among  the four “quantity” measures  (Number of Kin, Number of Nonkin,  Availability of Acquaintances, Availability of Close and Confiding relationships)  is =.35.  Beyond the issue of overlap,  it would be instructive to assess the magnitude of the collective association between outcome and the domain of supportive social relationships.  The multiple regression equation in Table 20  illustrates the results of such an assessment. The results in Table 20 show that, without considering any other influences, social relationships collectively account for about 86 of the variance in five—year outcome (Step 1. AdJusted  141  Results  Hierarchical Regression: The Collective Role of Table 20. Social Relationships in Predicting Five-Year Outcome.  Step 1. Main Effectsa No. kin in network No. nonkin in network Avail, of confidants Avail, of acquaint. Adeq. of confidants Adeq. of acquaint. Step 2. Diagnosisb (Schiz=1; Aff=2) Step 3. Interactionsc No. kin X dx. No. nonkin X dx. Avail. confid. X dx. Avail, acquaint. X dx. Adeq. confidants X dx. Adeq. acquaint. X dx.  Step  One  Step  Two  Beta  T  Beta  T  —.055 .070 .094 .108 .094 .070  —.54 .62 .81 .79 .81 .62  —.059 .065 .106 .017 .078 .028  —.61 .64 .77 .13 .71 .26  .394  Step Three  3.86***  Beta  T  —.363 —1.08 .716 1.48 —.484 —.97 .443 .74 .241 .67 —.097 —.28  .501  1.48  .409 1.03 —.736 -1.41 .561 1.20 —.452 —.80 191 -.54 .106 .31 —.  Notes. Sample size: N=100 (50 in schizophrenic group, 50 in Outcome is measured by Axis V ratings, the affective group). highest level of social and occupational functioning in the past month (higher scores indicate better outcome). 240 (df693) aMult.R.37; Adj R p=.O3. R = .13; chg =.08; Chg 2 2 chg_. (df=7,92), p<..OOl; Adj 2 R = .20; Chg R =.12; 1493 2 CMult.R.55; Adj R =.04; 2 =.19; Chg R 2 (df=13,86), p=ri.s. P<.001  142  Results  R—square=.08, p=.03).  No single social relationships measure  accounts for a significant amount of outcome variance.  After  entering the main effect for diagnosis in Step 2, the subsequent addition in Step 3 of the interaction between social relationships and diagnosis does not add to the prediction of outcome. The near-significant collective contribution of social relationships may be due to the stronger contribution of diagnosis to the prediction of outcome.  There is, after all,  considerable shared variance between diagnostic status and social relationships, on the one hand, and diagnosis and outcome, on the other.  The same argument may be made for other characteristics  observable at intake to the study, relationships were assessed.  i.e. the time when social  The issue of whether social  relationships have any unique and independent ability to predict outcome--over and above diagnosis and other baseline characteristics—-is addressed below. The unique role of social relationships.  Results from zero—  order correlations presented to this point indicate that several of the “control” variables predict five—year outcome, as do several of the social relationship measures.  Moreover, these  correlations are often of differing magnitude for the two diagnostic groups.  Finally, the “control” variables and the  social relationship measures are themselves inter—related. To assess the unique and independent role of these many variables, the following two—step multiple regression equation uses the five-year Axis V rating as the dependent variable.  Step  143  Resultrs  1 is comprised of main effects: the six measures of social relationships are accompanied by those “control” variables that show a significant or near—significant (p<.lO) correlation with five—year Axis V ratings.  zero—order  Step 2 is comprised  of interactions between diagnosis and all six social relationship ratings,  and interactions between diagnosis and control However, prior to the addition of diagnosis/control  variables.  variable interactions, equation.  Thus,  there are 18 terms in the regression  of all the potential interactions between  control variables and diagnosis, only that between sex and diagnosis has been added to the regression equation.  The  interaction with sex is selected because the difference between the outcome-sex correlations for the two diagnostic groups is strongest ; 32 L=—.  (schizophrenic group y.11; affective psychosis group  ) of all the zero-order correlations between 43 rff=.  control variables and outcome. Table 21 displays the effect of social relationships on five— year outcome, while controlling for concurrent influences from age, sex, diagnosis, the duration of the prodromal period, and baseline Axis V ratings.  In Step 1, no social relationship  variable achieves significance:  Where five of six ISSI variables  had a significant zero-order correlation with five—year Axis V (Table 19), role.  all have diminished to a nonsignificant independent  This appears to be due to their shared relationship with  several control variables:  Diagnostic status, duration of the  prodromal phase, and baseline Axis V ratings each make a  144  Results  Table 21. Hierarchical Regression: The Unique and Independent Role of Social Relationships in Predicting Five-Year Outcome.  STEP ONE: MAIN EFFECTSa. Social relationships No. of kin in network No. nonkin in network Avail, of confidants Avail, of acquaint. Adeq. of confidants Adeq. of acquaint. Control variables Diagnosis (1=sz,2=aff) Age at onset Dur’n of prodrome Intake Axis V Sex (1=male,2=female)  Step  One  Step  Two  Beta  T  Beta  T  —.067 .044 .056 .004 .034 .054  —.69 .43 .40 .03 .31 .51  —.637 —1.97 .510 1.09 —.174 —.35 .326 .58 .164 .48 —.243 —.74 **  3.05 .330 .070 .238 2.48k .242 2.10 —.152 —1.61  INTERACTIONSb. STEP TWO: Social rships X diagnosis No. of kin in network No. nonkin in network Avail, of confidants Avail, of acquaint. Adeq. of confidants Adeq. of acquaint. Control variables by dx Sex  .769 .083 .234 .251 .459  2.01 .79 2.45 2.17 1.45  .787 2.05 —.528 —1.05 .203 .43 -.351 —.66 .145 -.43 .301 .93  *  -  —1.014 —2.24  *  Note. Sample comprised of 50 patients with schizophrenia and 50 with affective psychosis. aMult.R.57; Adi R =.25; Chg R 2 =.33; =394 2 bMult.R.65; Adj 2 R = .10; R = .30; Chg 2 *  05 p<.  **  p<.O1  (df=11,88), p<.OO1 (df=18,81), p=.O8.  145  Results  significant independent contribution to the prediction of This shared variance does not negate any association  outcome.  between outcome and social relationships; rather, covariance,  in a manner where the  ‘control’  it indicates  predictors more  extensively account for variance in outcome that is already explained by social relationships predictors. Noteworthy among the significant main effects is the direction of the sign for the “duration of prodromal phase” Beta weight. that,  The positive weight  (Beta=.243, T=2.53, P<.O1)  for all patients combined,  predicts better outcome.  a longer prodromal period  When considered separately for the  schizophrenic and affective psychosis groups, the correlations is positive but marginal n.s.  and y.17,  indicates  the direction of  (from Table 17:  =.13,  p<..1O, respectively).  Considering results regarding the interaction between diagnosis and predictor variables, two terms approach significance.  The first, representing diagnosis by the number of  kin in the social network,  indicates that there is a tendency for  kin have a different effect on outcome for the two groups. Whereas the zero—order correlation between kin and outcome for each of the two diagnostic groups is nonsignificant when taken on its own (Table 17), Table 21 indicates that the difference between the two correlations approaches significance  (p<.05).  A somewhat different interaction effect is seen for sex as a predictor.  Table 17 showed that, among schizophrenics, sex is  not associated with outcome (=.11,  p=n.s.).  Contrarily,  for  affective patients, sex is significantly associated with five—  146  Results  year outcome  (=—.32, p<.O1;  i.e. men have a better outcome).  Step 2 of the current regression equation assesses the difference between =.11 and =-.32 and, after controlling for other variables,  finds that it approaches significance.  In Table 21, the lack of significance for the nonkin—by diagnosis interaction term is notable.  The zero-order  correlation between the number of nonkin and outcome was r=.31 (p<.05)  for schizophrenics,  affectives.  but only L=.O 2  (p=n.s.)  for  By considering only these two correlations, one  would expect the interaction term in Table 21 to be significant. However, was =.21  for all cases combined, the nonkin—outcome correlation (p<.O5); thus,  the absence of an interaction in the  regression equation may reflect a significant main effect, which in turn may have been encompassed by the relationship between outcome and baseline Axis V, diagnosis,  or duration of prodromal  period. To summarize, the results indicate that, combined,  for all patients  the social relationships that are present at onset are  related in a modest and nonsignificant way to five—year outcome, although this variance is shared with diagnosis and several baseline characteristics. variables account for about  Together, the social relationships of the variance in five-year  outcome of both diagnostic groups combined. schizophrenic group,  Within the  there is a nonsignificant tendency for more  nonfamily members in the social network to predict better outcome.  147  Results  Stressful Life Events The second hypothesis in this thesis was that increased numbers of stressful life events will have a modest, negative influence on five—year outcome. this hypothesis,  In presenting results related to first, descriptive  the order will be familiar:  results are presented regarding group differences in the occurrence of life events in the time periods leading up to intake,  9-month, and 18-month interviews.  Thereafter, simple and  then complex predictive relationships are presented in order to assess the role of life events on outcome. Table 22 displays the number of events reported by each group in the periods leading up to the intake, month interviews. Table 22,  9—month and 18-  Tests of the group differences displayed in  as well as other parametric analyses of the life events  (n+1) transformations 10 variables reported below, are based on 1og due to the strong positive skew for each of the life events variables.  Nonetheless,  for illustrative purposes, the mean and  standard deviations displayed in Table 22 are the actual numbers of life events. A preliminary MANOVA, with the number of independent and “possibly independent” events at the three points in time as the dependent measures,  indicates that the schizophrenic group  experienced fewer life events than the affective psychosis group (Hotelling’s T =2..18, df=6,117, p<.05). 2  The result is  attributable to differences in the number of independent events at intake and 18 months.  In the “possibly independent” category  at intake, a nonsignificant trend is present  ), where 6 (p=.O  148  Results  Table 22. Number of Life Events at Three Points in Time, by Diagnosis.  Number of life events  Schizophrenic subjects (n=64) mean  s.d.  Affective psychosis subjects (n=60) mean  prob  s.d.  Intake Independent Possibly indep.  2.6 1.6  2.8 1.8  3.9 2.2  3.2 2.2  2.44 1.74  9 month follow-up Independent Possibly indep.  2.3 1.5  2.5 1.8  2.5 1.5  2.6 1.8  .29 .14  18 month follow-up Independent Possibly indep.  2.8 1.5  2.5 1.6  3.6 1.6  2.6 1.8  1.84 .34  .02 .06  .02 ——  aThe tests, based on independent samples and pooled variance (n+1) transformations of the 10 estimates, use for analysis the 1og All tests have 122 d.f.. number of life events.  149  Results  schizophrenics experienced fewer events.  The group differences  are not due to concomitant differences in baseline functioning or the duration of the prodromal period:  a multivariate analysis of  covariance (MANCOVA) with these as covariates did not alter the results described above. Predicting five-year outcome.  Regardless of group  differences, does the number of events predict outcome?  Table 23  shows zero—order correlations between five—year Axis V ratings and each of the life event indices. There is a complete absence of expected effects.  Whereas  correlations of approximately =.3O had been hypothesized, the actual results, based on all cases combined, are remarkably close to zero  (range =—.O9 to =+.12).  For schizophrenic patients,  the correlations range from =-.O2 to =+.12.  For this group,  there is no single coefficient that departs significantly from zero. The range of correlations among the affective psychosis patients  is =-.O1 to  =—.  predict five-year outcome.  31,  indicating that life events do not  The same conclusion is reached when  the six life event variables are considered simultaneously in a hierarchical regression. outcome  As main effects, they do not predict  =-.04, p=n.s). 2 (Mult. R=.l6, AdL R  After adding  =.15, p<.OO1), 2 diagnosis in a second step (Mult R.=.46, Adi. R life events do not make a significant contribution when considered in interaction with diagnosis =.14, p=n.s.). 2 R  (Mult. R.=.52, Adj.  When the correlations and regression results  are considered together,  it seems that there is no linear  150  Results  Correlations Between Five-Year Outcome and the Log of Table 23. the Number of Life Events, by Time and Diagnosis.  Log of no. of life events  Schizophrenics (n=52) r  Intake Independent Possibly independent  prob.  Affective Psychos is (n=49) r  .10 .02  —.02 —.12  9—month follow-up Independent Possibly independent  .04 .08  —.01 —.31  18-month follow-up Independent Possibly independent  .09 —.02  —  ——  ——  —.16 —.18  prob.  All cases (n=101) r  prob.  .12 .02  ——  .01  ——  ——  —.01 —.09  .05 —.06  ——  ——  151  Results  relationship between life events and five—year outcome for either group. These results assess possible linear associations between life events and long—term outcome.  Nevertheless,  it may be that  a curvilinear association would be more appropriate, particular an inverted “U” relationship.  If true,  in  it would  indicate that there may be an optimum amount of stressors, whereby either too few or too many events would be associated with a poorer outcome.  To test such a notion,  scatterplots and the residual vs.  the bivariate  predicted plots were examined  for each of the log—life events by outcome combinations.  The  results indicated an absence of nonlinear relationships. Taken as a whole, the results derived here fall short of those hypothesized:  the number of life events does not predict  five-year outcome for either group.  Among schizophrenics, no  zero—order correlation with five—year Axis V exceeds =.12. Among affective psychosis patients, only one life events variable (“possibly independent” at nine months) achieves significance, but it is not accompanied by an association between the “independent” life events variable at nine months and five—year outcome.  Overall,  it appears that,  for both diagnostic groups,  life events in the 2.5 years surrounding intake to the study do not predict five-year outcome.  Biological Vulnerability and the Course of Illness If social relationships and life events comprise two components of a diathesis—stress model of the course of illness,  152  Results  biological vulnerability indicators comprise the third and fourth.  As described in the introductory chapter, previous  studies have failed to document correlations between outcome and ventricle—to-brain ratio (VBR) or smooth pursuit eye movements (SPEM) among affective disorder patients.  As such, the  hypotheses here are that VBR and SPEM dysfunction predict outcome for the schizophrenic group only. In presenting the descriptive results, a preliminary comment about the distribution of the data is in order. tracking error scores are left—censored  Because eye-  (skew=+1.62),  this  variable was transformed using the log(n+1) convention.  Thus,  the group means and standard deviations for RMS error shown below are actual scores,  but the parametric analyses are based on log-  transformed data.  VBR data are displayed and analysed in their  original, raw form.  The results in Table 24 show that there are  no group differences in either the mean eye—tracking error or ventricle—to—brain ratio. As discussed in lacono, Moreau, Beiser, Fleming, and Lin (1992),  there is a dramatic discontinuity in the distribution of  eye—tracking RMS error scores.  Such a discontinuity may indicate  a dichotomous phenomenon, and hence a dichotomous vulnerability factor in predicting the course of a disorder.  The distribution  of the raw RMSE scores is illustrated in Figure 3 and shows there are no cases between 304 and 353 units of error.  Thus,  this  point of discontinuity is used to define good and poor eye tracking groups.  When considered by diagnosis,  (15.O’o) schizophrenic cases, and 5 of 61  there are 9 of 60  (8.2%) affective cases  153  Results  Table 24. Diagnosis.  Eye-Tracking Errora and Ventricle—to-Brain Ratio by  Eye—tracking error Schizophrenics (n=61) Affectives (n=63)  Mean  Std.Dev.  (df)  prob.  185.38 178.36  121.63 108.32  (122)  n.s.  (77)  n.s.  Ventricle-to-brain ratio Schizophrenics (n=45) 6.60 Affectives (n=34) 6.04  2.62 2.40  .98  aEye.tracking error using root-mean-squared-error approach. bThe test for eye—tracking RMS error is based on log(n+1) transformed data.  154  Results  Smoothed Distribution of Eye-Tracking RMS  Figure 3.  Error Scores for 124 First—Episode Psychosis Patients.  *  36  +  33  ÷  30  +  27  4-  24  +  21  +  18  +  15  +  12  +  No. of pts.  *  9  ÷ *  6  +  3  +  0  +  +  +  100  200  300  Eye—tracking root-mean—squared  >480  400  (RMS)  error  155  Results  with poor eye_tracking*. significant  The difference in proportions is not  (Chi—square=l.36, df=l, p=n.sJ.  Unlike eye-tracking error scores, the VBR data are normally distributed  (skew=-.05; kurtosis=—.24).  discontinuity,  There is no Thus,  only a smooth progression of data values.  VBR is well suited to consideration in its continuous, raw form. The hypotheses are that VBR and eye-tracking error will predict outcome only for schizophrenics. VBR do not support the hypothesis:  The results regarding  larger ventricles do not  presage poorer outcome in schizophrenia  , p=n.s.). (L=. O 2  Similarly, the hypothesis of a significant correlation between eye—tracking error and five—year outcome for schizophrenics is not supported  (L=.ll, p=n.s.).  Since the distribution of the smooth pursuit error scores suggests a “natural” dichotomy,  the outcome of normal and poor-  tracking subgroups can be compared.  Using a univariate  ,  test,  the results indicate that eye-tracking performance is not associated with five—year outcome in schizophrenia.  In fact,  Axis V ratings are virtually identical for the good-tracking group (mean ±. sd: (3.9 £ 0.6; n=6;  3.8 ±. 1.3; n=42) =—.29, df=46,  and the poor-tracking group  p=n.s.).  In sum, the results  fail to support Hypothesis Three, that biological vulnerability factors will predict long—term outcome.  *  These figures are slightly different than those appearing in In their paper, they used lacono, !loreau, et al. (1992): compared to the current data, intake diagnoses based only on from intake, 9-month, and data on overall based diagnoses” “best 18-month follow-up interviews.  156  Results  Building Models  Both the stress—process and diathesis—stress models of the course of illness require the hypothesized relationship between stressful life events and outcome.  The results based on five—  year Axis V ratings described above indicate a virtual lack of correlation between outcome and the number of events. Phase Three, devoted to data reduction techniques, required.  Hence,  is not  Similarly, Phases Four and Five, where the predictors  would have been combined into stress-process and vulnerabilitystress models  (respectively) are not required.  results pertaining to social relationships,  Nonetheless, the  life events, and the  two biological vulnerability indicators can be discussed in the next chapter with respect to their individual influence on outcome.  Summary of Results  The results regarding methodology suggest that procedures involved in collecting the five-year follow-up data are sound. The response rate is good, with no detectable differential attrition, and with minimal opportunity for bias due to the source of follow-up data used in analyses.  Preliminary analyses  also show substantial differences in the level of functioning at each point in time for the diagnostic groups.  Nonetheless, the  157  Results  chances in the level of social and occupational functioning over time, as reflected by DSM—III Axis V ratings, are similar for schizophrenic and affective psychosis patients. In terms of the predictive results, none of the four sets of predictors is significantly associated with five—year outcome. Only one nonsignificant trend is apparent:  Social relationships  collectively account for about 8? of the variance in five-year outcome, although this variance is shared by several baseline characteristics  (including diagnosis).  When diagnostic groups  are considered separately, the hypothesized results are present only for schizophrenics, and only in the sense that the number of nonkin shows a nonsignificant tendency to predict outcome. The hypothesis regarding the predictive role of life events is not supported.  Group differences are present, where affective  psychosis patients experience more independent events than do schizophrenics. Negative findings also apply to the biological vulnerability factors.  Contrary to the hypotheses, neither biological measure  predicts five—year outcome. The null results do not lend themselves to either a stress— process or a diathesis—stress model:  These two models require  that life events have a modest negative impact on course and outcome.  In this project,  life events have no apparent  relationship with five—year Axis V ratings.  As such,  there seems  to be no point in trying to construct stress—process and diathesis—stress models by integrating results that were found  158  Results  here.  Thus,  the next step is to discuss in the next chapter the  results that have been presented——regardless of their place in an integrated model.  159  VII.  DISCUSSION  In considering the foregoing results, the organization of the discussion will follow the same order as the previous chapter.  The first section will review methodological and  preliminary results.  The second will examine descriptive  results, and predictive results vis—a—vis each of the three domains. project,  Finally,  I will assess the significance of this  its limitations, and discuss future research.  Methodology Several methodological features may have affected the substantive findings.  They include the definition of the  grouping variable, response rate and attrition,  instrumentation,  and formal statistical issues.  Diagnoses In the MAP Project, procedures.  the diagnoses are based on thorough  Information from semi—structured diagnostic  interviews, clinical records, and a collateral informant at both intake and 18- month follow-up were combined at case conferences to derive a “best overall” DSM—III diagnosis.  The result is  diagnostic groupings that are as homogeneous as current nosological standards allow.  160  Discussion  However pure, the DSM—III criteria provide for a narrow or By defining only the  restrictive definition of schizophrenia. ‘nuclear’  or  ‘core’  group of afflicted persons, the effect is to  circumscribe severity. ICD—9  Less restrictive criteria such as the  (World Health Organization [WHO],  1978) tend to include  patients with more affective symptoms, with psychotic symptoms that are shorter in duration, at all  or even with no psychotic symptoms  (e.g. simple schizophrenia).  Compared to DSM—III, one  might expect that the effect of less restrictive criteria would be a broader range of outcome, particularly at the more favourable end.  A related result might be to increase the  magnitude of any predictive correlations. This was the rationale behind an extensive set of supplementary analyses, using ICD-9 diagnoses to define schizophrenic and manic-depressive groups.  Use of the broad  definition of schizophrenia resulted in a group comprised of 91 cases; the manic-depressive group numbered 55.  Nineteen of the  DSM—III affective psychosis cases were designated schizophrenic in the lCD system.  The results indicate that, descriptively, the  change in nosology has little effect on sample characteristics such as age at onset or baseline levels of adaptive functioning. Moreover,  in assessing the predictive relationships between five  year outcome and each of social relationships, and SPEM,  life events, VBR  the correlations were substantially unaltered.  Thus,  the notion that the restrictive definition of DSM-III schizophrenia has restricted the range of adaptive functioning, and thus attenuated any prospective correlations that would have  161  Discussion  been present with a broader diagnostic definition of schizophrenia,  is unlikely.  Using DSM—III criteria, the affective psychosis group is diagnostically heterogeneous.  As such,  it is comprised of  approximately equal numbers of patients with Bipolar or Major Depressive Disorders, all of whom had psychotic symptoms in the index episode.  This heterogeneity is not an issue in this study,  as the primary purpose of the affective patients was to act as a psychotic comparison group.  In this thesis, analyses have  examined the Bipolar and Major Depression patients separately only when the pooled results appeared puzzling.  This is the case  in interpreting the role of sex as a predictor of five—year outcome for the combined affective group (cf. Table 17,  and  below). A detailed examination of the relationship between the outcome of Major Depression and Bipolar Disorder per se with social relationships,  life events, and biological vulnerability  is beyond the scope of this thesis.  Procedure The procedures in the intake and 18—month follow-up phases of the MAP Project have been reviewed in the preceding chapters. Several aspects of the five—year follow—up, however, are worthy of discussion here. The response rate was good:  five-year outcome status is  known for approximately four-fifths of the patients.  This  compares favourably to the earlier 9— and 18—month follow-up rates of 72-73.  The five—year response rate also compares  162  Discussion  favourably with other naturalistic studies of first—episode cohorts  (e.g.  Strauss & Carpenter,  1977),  although follow—up  investigations that also provide treatment have higher response rates  (e.g. Ventura et al.,  1989).  In this study, the epidemiological approach to recruitment was aimed at obtaining a representative sample.  In the follow-up  portions of the study, a high response rate is necessary at all stages to maintain that representativeness.  With four-fifths of  patients described at five years, and with numerous procedures failing to detect differential attrition,  it seems that the  sample described in this thesis can be taken as fairly representative of first-episode DSM-III schizophrenia. Although several procedures were undertaken to maintain contact with the cohort, with increased resources.  two improvements may have been possible Had we been able to expend more effort  in the period between 18—month and five—year follow—up dates, fewer patients may have been lost to follow—up.  Secondly, among  patients whose outcome was determined, there may have been a change in the nature of the data used:  Some of the patients who  provided abbreviated data would have completed the full procedure (e.g.,  if funds were available for lengthy long-distance phone  calls).  Instrumentation Formal properties of the various instruments were developed as part of earlier studies within the MAP Project, as well as specifically for this thesis.  Of the former, the internal  163  Discussion  consistency coefficients for the social relationships variables have the most bearing on this project.  In Erickson et al.  (1989), we described acceptable Cronbach’s alpha coefficients for four of the ISSI scales.  Although they are satisfactory, and  consistent with formal properties of other measures of supportive social transactions in the literature Byrne,  & Scott,  1980>,  (Henderson, Duncan—Jones,  it may be that the true role of social  relationships on the course of illness was attenuated due to less than perfect reliability, and thus is stronger than was demonstrated here. Indeed, the four ISSI scales were first developed for the earlier study assessing the role of social relationships on short—term outcome  (Erickson et al.,  original ISSI scales  1989) because Henderson’s  (Henderson et al.,  1980) did not cohere for  either our psychotic sample or the normal comparison group. Thus,  the four new scales developed in the MAP Project need to be  validated on another sample.  It will also be important to re  assess the evidence for convergent and discriminant validity of the six social relationship measures, using the principal components analysis described earlier or other techniques. A similar point may be made for the formal properties of the life events measures, where there is a satisfactory but not outstanding level of agreement between raters.  One way to raise  the inter—rater agreement about the life events data would be to date onset at the time of the original intake interviews, so that life event data collection could be anchored to onset.  For  example, any events that followed the onset of the prodrome would  164  Discussion  automatically be deleted from consideration as independent of the illness. The dependent measure in the prediction equations, DSM—III Axis V ratings,  has reasonably good formal properties.  The  inter—rater reliability estimates at five years and at 18 months indicate a high level of agreement. the Axis V scale is a global rating.  Despite good reliabilities, Even though it is  restricted to the domain of adaptive functioning,  it nonetheless  combines two separate but related dimensions of outcome: occupational and social functioning.  Ratings of any particular  sample may achieve excellent interrater reliability, but the nature of a global rating is that correlations between it and a domain-specific predictor such as occupational or social functioning alone will be attenuated compared to the use of a corresponding domain—specific measure of outcome al.,  (Stoffelmayr et  1983). Use of psychiatric records as the source of five—year  outcome data seems to have had little effect on the results.  The  patients whose adaptive functioning was described by these records indeed had lower Axis V ratings than those ascertained by other means.  Nonetheless, these patients also had lower levels  of functioning at intake.  Thus,  it seems that the use of a  potentially biased source of data has not been translated into actual bias.  Statistical Issues The primary statistical consideration was controlling for  165  Discussion  inflation of Type I error. from those proposed,  The actual analyses depart somewhat  but the number of statistical tests used to  assess the hypotheses remained the same.  The five analyses  include: —-  --  —-  --  two hierarchical regressions where social relationships predict outcome (Tables 20 and 21); a regression equation where life events predict outcome; a correlation between VBR and outcome; and a chi-square test assessing the effect of good and poor eye—tracking on outcome.  In light of the five statistical tests,  the Bonferroni  correction dictated the use of an alpha level of  .01.  One of the four predictors of interest meets the conventional criterion.  .05  This one result, where the joint and uncontrolled  effects of social relationships account for 86 of the variance, is rendered nonsignificant with the use of the  .01 criterion.  In  short, the use of a .01 alpha level reduces a somewhat mixed set of reliable results to a modest set of nonsignificant trends. Since the onset of the present study, in the analytic strategy.  I have added one step  Earlier, the objective was to show  that a predictor had a unique and independent influence on outcome.  Here, the strategy was to show that social  relationships or life events or biological risk added to the ability to predict outcome,  over and above that explained by any  traditional predictors such as age, sex, diagnosis, and level of premorbid functioning.  Computationally, that meant that control  variables would comprise the first step in hierarchical  166  Discussion  regression,  and the predictor variable(s)  of interest would  comprise a second step. Subsequently,  I decided to examine more carefully the  variance that the hypothesized predictors life events, biological vulnerability)  (social relationships,  share with more  traditional predictors such as age at onset, sex, and premorbid status.  This change in strategy is based on the notion that  shared variance is no less important as variance that is uniquely attributed to a domain of interest.  The result is an increased  emphasis on the zero—order correlations and, where warranted, a preliminary multiple regression equation where the joint influence of all indicators of the predictor domain of interest is considered  (cf. Table 20).  The change in strategy allows  greater consideration of clinical implications.  For example,  if  social relationships or life events share variance with sex or age in causing variability in outcome, the former but not the latter.  intervention can influence  This shift in analytic strategy  increases the ability to identify factors that are amenable to intervention.  Descriitive and Preliminary Results  Adaptive Functioning Over Time Approximately one—fifth (16—22?6)  of our schizophrenic sample  is healthy at any point in time, a proportion similar to those described in the literature (e.g., Ciompi,  1988).  This fact will  167  Discussion  be of some comfort to patients and their families, showing that prognosis,  even by DSM—III criteria,  is not universally dismal.  Another large portion of afflicted persons, approximately onehalf, can look forward at the five—year mark to keeping their disorder at bay while they continue their daily activities. Members of this subgroup will need courage to meet the challenge, and help in their struggle.  At the same time, many will be  heartened by the fact that maintaining one’s activities and responsibilities throughout the illness is not only possible, but probable. To aid in the interpretation of the five-year results, supplementary analyses predicting 18-month outcome are presented below.  Figure 4 illustrates the Axis V ratings of the two  diagnostic groups at 18 months, as well as at intake and five years. Both groups declined in functioning at 18 months, relative to intake, and then increased to a level above both intake and 18 months.  Moreover,  the parallel trajectory of adaptive  functioning in the two patient groups over time, as seen earlier in Figure 2, continues to be apparent when 18—month status is included. The parallel course of the two diagnostic groups is not limited to functioning measured by DSM-III Axis V ratings. Within the MAP Project, we  (Beiser,  Bean, Erickson,  et al.,  1993)  have already shown that the parallel course phenomenon holds true for occupational functioning between intake and 18 months. that study, the amount of time devoted to work or school  In  168  Discussion  Figure 4. Axis V Ratings at Three Points in Time, by Diagnosis.  6  5 F+-.13  4 > Cl,  3 Cl) 0  2  1  -.  0  Intake  —  Eighteen Months Five Years Time  z ts. 1 Schi illAffective Pts. P  169  Discussion  was measured using a 95-point scale, where the major contribution to a patient’s score was determined by a tally of the number of days at a job or in school in the past 9-12 months.  Thus, the  parallelism was also detected with a domain—specific and cumulative measure, as well as a global outcome  ‘snapshot’ measure of  (Axis V ratings).  Control Variables and Outcome It has been necessary to examine the relationship between outcome and several demographic and baseline characteristics, so as to anticipate any “third variable” explanations for any relationships between outcome and predictor domains of interest. Two baseline characteristics of particular interest are the duration of the prodromal phase, and sex—related differences in diagnostic status. Sex differences.  In the literature on course and outcome in  schizophrenia, one of the most robust findings is that women do better than men (e.g., Goldstein, described above, however,  1988).  In the results  this sex difference is absent:  at five  years, the sex difference in ratings of current adaptive functioning is marginal and nonsignificant.  Similarly, at 18  months there is an absence of sex diflerences in outcome. Both findings, however, may be due to the index used to measure outcome.  Axis V ratings are based on a  definition of outcome.  ‘snapshot’  If a cumulative measure is used to  measure short-term outcome, viz,  occupational functioning  170  Discussion  aggregated over the nine months leading up to the 18-month follow-up, the schizophrenic women in this sample show the usual advantage over their male counterparts et al.,  1993).  In the same domain, a  (Erickson,  unemployed vs.  in—hospital)  and men show comparable outcome.  Bean,  ‘snapshot’ measure of  current occupational functioning at five years part-time vs.  Beiser,  (full-time vs.  indicates that women  Taken together, this may  indicate that periods of relapse and impairment are briefer for women, compared to men. Mode of illness onset.  In the schizophrenia literature, the  duration of the prodromal phase has been robust as a prognostic indicator  (Stoffelmayr et al.,  1983), where a lengthy,  onset is a sign of poor prognosis.  insidious  In our sample of  schizophrenic patients, however, duration of the prodrome shows no significant relationship with either five-year or 18—month outcome.  The same is true for the affective cases.  When  considered in a multiple regression equation based on all cases combined (Table 21), however, the duration of the prodrome becomes significant (p<.Ol)  in predicting five—year outcome.  This indicates that the magnitude of the relationships detected with correlations  (=.13 for schizophrenics; =.l7 for affective  patients; all cases combined L=.l ) 4  is significant when other  variables are assessed simultaneously.  The surprising result is  the positive sign of the Beta coefficient,  indicating that a  longer prodrome predicts better five—year functioning. The positive relationship is difficult to understand.  If it  was a null result, there could be several explanations for the  171  Di5cussion  discrepancy with extant literature.  One relates to the nature of  the samples described in the literature:  With one exception,  samples are comprised of multiple-episode patients.  This means  that poor—outcome patients will be over-represented, thus exaggerating any true association between mode of onset and outcome.  In the literature, the mean correlation between outcome  and mode of onset is a modest =.25  (s.d.=.08;  cf. Table 6):  If  this is exaggerated due to sampling bias, then the true relationship may indeed be close to zero. however,  is not what these data suggest:  A null relationship, they point to a small  but significant positive relationship. A related issue is that much of this literature has used the concept of treatment lag-time as a proxy for mode of onset, instead of the duration of the prodromal period.  If one uses  lag—time in our samples, we still get results that disagree with the literature:  Among the schizophrenic sample, correlations  between lag—time prior to first treatment contact and either 18month or five—year outcome are virtually zero  (L=.OB and L=. , 09  respectively). Many of the previous studies used a six-month cut—off point to dichotomize either the prodromal period or the treatment lag— time, thereby defining acute- and insidious-onset subtypes Literature Review section above).  (see  When applied to our sample,  this strategy provides consistent negative results.  For example,  dichotomizing the prodromal period, there is no difference in five-year outcome for the acute vs.  insidious subgroups of either  the schizophrenic or affective patients (data not shown).  A  172  Discussion  similar absence of a significant relationship is found when the lag—time is dichotomized.  Thus,  the use of this less sensitive  analytic strategy washes out the positive relationship between the length of the prodromal period and five-year outcome.  Predictive Results  Social Relationships Within the domain of social relationships, the most important new finding is a nonsignificant trend  ) toward a 5 (p<.O  diagnosis—specific effect for the number of kin in the social network on five-year outcome.  When all patients are combined for  analysis, the increased presence of family tends to have a more negative influence on course for schizophrenic patients compared to a more positive influence on affective patients.  Although  this five—year multiple correlation is attenuated to  ‘trend’  magnitude, the result replicates the pattern shown at 18 months. For the schizophrenic patients, there is also a suggestion that the number of nonfamily members of the social network at intake continues to have an influence on five-year outcome. nonsignificant  This  , p<.O5) relationship is similar to the (L=. l 3  influence shown at 18-month outcome  (=.37, p<.O1), although the  five-year association is encompassed by other factors in a multiple regression paradigm.  173  Discussion  Overall, the magnitude of the association between social relationships and outcome has diminished between 18 months and five years.  Specifically, the correlations displayed in Table 19  are attenuated compared to those reported in our paper describing 18—month outcome  (Erickson et al.,  1989).  A similar conclusion  is derived from the multiple regressions predicting 18—month and five—year outcome:  Social relationship variables and their  interaction with diagnosis accounted for 156 of the variance in short—term outcome, but only 86 in longer—term outcome. In the literature, the only other study (Cohen & Sokolovsky, 1978) assessing the effect of social relationships on outcome found an effect among mildly but not severely impaired schizophrenic patients.  Results from this sample suggest that  the finding is not limited to mildly impaired patients. Together,  the 18—month and five-year data are consistent with the  notion that social relationships do influence the course of illness for schizophrenic patients with a broad range of impairment.  Specifically,  in light of the positive association  between nonkin and 18—month outcome for schizophrenic patients, and the positive association between kin and 18-month outcome for affective psychosis patients,  this may indicate that  schizophrenia implies a limited capacity to benefit from the kind of support that families members provide. Several other interpretations of these results are also possible.  One is that family members of schizophrenic patients  may be less able to provide constructive support in the face of a difficult burden.  A third possibility is an interpretation based  174  Discussion  on an artifact:  More relatives are associated with poorer  outcome for schizophrenia because more relatives are needed to care for patients with more severe disorders.  This  interpretation may be less likely than the others, since a similar correlation was not apparent for the affective psychosis patients.  In fact,  the 18—month outcome/social relationships  association was in the opposite direction for the affective group.  Life Events In the periods leading up to the intake and 18—month interviews, affective patients reported experiencing more independent events than did schizophrenic patients.  Since the  group difference remains after controlling for the length of the prodromal period and the level of baseline functioning, this appears to indicate that it takes fewer stressors,  on average, to  precipitate the onset of schizophrenia than affective psychosis. Alternatively,  it may be that the nature of self—report is  subject to differential bias in schizophrenia, affective psychosis.  compared to  For example, the general sensitivity to  people and events after onset may affect the retrospective reports of events prior to onset.  Here, people with  schizophrenia may count as an event something experienced as trivial by a person with affective disorder. It is likely that the lack of hypothesized association between five-year outcome and the number of life events is  175  Discussion  attributable to the 3.5 years between the latest assessment of life events and the assessment of outcome itself.  This lengthy  span of time may have reduced to zero what started as an effect of a modest magnitude.  If true,  it may be due to a true  diminution in the effects of life events,  or because a host of  other factors had opportunities to influence outcome.  In either  case, a stronger test of the role of stressful life events may be vis—a-vis short—term outcome. Life events and short-term outcome.  Ratings of 18-month  outcome differ from those of five—year outcome.  The latter  describes the highest level of adaptive functioning in the past month, whereas the former selects and rates the best functioning over the past nine months.  This raises an additional  consideration regarding the temporal priority of life events and outcome.  If,  for example, the best functioning during the  follow—up period was in month 10 or 11, then many life events will occur after the period reflected in the “18 month” Axis V rating.  Thus, only the intake and nine-month life events can be  included in the prediction of 18-month outcome.  Table 25  displays the associations between 18-month outcome and life events at intake and nine-month follow-up.  Also shown are  correlations between 18-month outcome and selected baseline characteristics. The table shows that, among schizophrenic participants, the number of independent life events occurring in the year prior to intake predicts 18-month Axis V ratings  (=.38,  p<.O1).  While  176  Discuss ion  Correlations Between 18—Month Outcome’ and the Number Table 25. 2 or Baseline Characteristics, by Diagnosis of Life Events Affective Psychosis (n=54)  Schizo— phrenics (n=55) r NO. OF LIFE EVENTS Intake Independent Possibly indep. 9—month follow-up Independent Possibly indep. BASELINE CHARACTERISTICS Intake Axis V Sex (1=male,2=female) Age at onset 3 Dur’n of prodrome 3 Treatment lag-time  .38 .00  .20 —.15  .005 ——  ——  ——  ——  r  prob.  .34 .14  .01 .08  .16 —.01  .05  prob.  .14 .10  .01  .16 .11  .37 .07 .12 —.06 .09  r  prob.  All cases (n=109)  —  .50 .28 .27 .05 .22  .08  .001 .02 .02 ——  .08  —  .54 .24 .29 .04 .00  ——  .001 .01 .001 ——  ——  Eighteen-month outcome is measured by Axis V ratings, the Notes. highest level of social and occupational functioning in the nine months leading up to the 18—month follow-up (higher scores 10 Correlations are based on log indicate better outcome). The duration of the prodromal period and transformations. treatment lag-time are based on Van der Waerden scores (normalized ranks).  177  Discussion  the magnitude of the correlation is as predicted,  more events are associated with  of the association is not: better outcome,  the direction  rather than less.  No other life events variable  was predictive of short—term outcome for schizophrenic patients. The correlations between life events at intake and 18-month outcome in the schizophrenic group may reflect a shared relationship with a third variable.  The second portion of Table  25 displays four of the leading candidates for explanations: prodrome.  intake Axis V, sex, age at onset, and duration of  Among the four,  month Axis V ratings Therefore,  ‘third variable’  only the Intake Axis V predicts 18—  (=.37,  p<.OO5)  it remains possible that,  for the schizophrenic group. for schizophrenics, the  positive correlation between the number of life events and 18— month outcome is an artifact of baseline functioning,  i.e.  previous impairment causes both fewer events at intake and poorer 18-month outcome for the schizophrenic patients. A somewhat different picture is present among affective psychosis patients.  As shown in Table 25, no single life event  rating is associated with 18—month outcome, although the ‘independent’ classification at nine months shows a nonsignificant  (=.2O,  ) trend. 8 p=.O  Again, the direction of the  trend is opposite to that hypothesized.  No other correlation  between short-term outcome and a life events variable approaches even the “trend” magnitude. Among baseline characteristics, too, a different picture is evident in the affective psychosis group.  Three of the four  178  Discussion  Stressful Life Events’ and Two Regression Equations: Table 26. 2 Separately Outcome Predicting 18-Month Characteristics Baseline Groups. Two Diagnostic for  SCHIZOPHRENIC GROUP  (n=55)  Intake: # events Independent Poss. indep. 9 months: # events Independent Poss. indep. Baseline characteristics Intake Axis V  T  Beta  .308 -.079  2.22 —.61  —.001 .015  —.01 .10  .285  2.12  prob.  .03 -—  .04  Overall statistics for schizophrenic group: Mult. R.=.47; Adj R—Square=.15; F= 2.85 (d.f.=5,49), p=.O2.  AFFECTIVE GROUP  (n=54)  # of events Independent Poss. indep. 9 months: # of events Independent Poss. indep. Baseline characteristics Intake Axis V Sex (1=male, 2=female) Age at onset  Beta  T  prob.  Intake:  -.043 .168  -.35 1.42  .091 —.227  .78 —1.85  .482 .273 .164  3.98 2.34 1.38  Overall statistics for affective group: Mult. R.=.65; Adi R—Sguare=.34; F= 4.89  ——  .07 .001 .02 -—  (d.f.=7,46), p=.OO . 1  Number of life events is based on log transformations. Notes. Eighteen month outcome is measured by Axis V ratings, the highest level of social and occupational functioning in the 9 months prior to the 18-month follow-up (higher scores indicate better Intake Axis V is based on the highest level of social outcome). and occupational functioning in the year prior to intake to the study.  179  Discussion  variables, outcome.  Intake Axis V, sex,  and age at onset, predict 18-month  The finding that sex predicts outcome  better outcome)  (i.e. women have  at 18 months is notable, given the opposite  finding at five years. In Table 26, two regression equations assess the unique and independent role of life events as they predict 18-month outcome for each diagnostic group.  In each case, the dependent measure  is the Axis V rating at 18 months,  reflecting the highest level  of social and occupational functioning since the nine month follow-up.  In each case, the predictor variables are the four  life event variables and the baseline characteristics shown earlier to be correlated with short—term outcome. The most striking result in Table 26 is that concerning the schizophrenic group, where the predictive role of the number of independent life events at intake is diminished only slightly. As before, the direction of the relationship is opposite to that which was hypothesized:  the sign of the Beta-weight indicates a  strong tendency for a greater number of predict better outcome, functioning.  Thus,  (log)  life events to  over and above the baseline level of  the artifactual explanation mentioned above  (e.g. that higher functioning participants get actively involved in their world, welcoming the challenge of a greater number of life events)  is unlikely.  Similarly, this relationship is not an  artifact of a greater duration of illness prior to intake to the study, since Table 25 showed that the duration of the prodromal period was unrelated to 18-month outcome.  As before,  the life  events in the period leading up to the nine—month follow—up do  180  Discussion  not predict 18-month outcome for schizophrenic participants. Together with Intake Axis V, the life events variables account for 15 of the variance in outcome (Mult R.=.47, Adjusted R— square=..15,  p=.O2).  Among affective psychosis participants, no life events variable predicts outcome in the regression equation.  Whereas  Table 25 showed a nonsignificant trend toward a correlation between the nine—month “independent” rating and short-term outcome  O, p=.08), there is no such result in the multiple 2 (L=.  regression analysis.  The absence of either a nonsignificant  trend or a significant effect in Table 26 for the nine-month “independent” rating likely indicates that being female and/or having a higher level of functioning at intake encompasses the (nonsignificant) relationship between life events at nine months and 18-month outcome. In short, predicting 18-month outcome that is proximal to the life events that were assessed in this study may be a more reasonable task than predicting distal,  five-year outcome.  the number of life events does predict 18—month outcome, differently for the two groups.  For schizophrenics,  in the year prior to intake do predict outcome:  Here,  albeit  life events  a greater number  of events is associated with better 18-month functioning, even after controlling for possible “third variable” explanations. Among affective psychosis patients, a nonsignificant trend is evident:  The number of independent events at nine months show a  tendency to predict 18-month outcome  O, p=.O8), 2 (L=.  although the  variance in this relationship overlaps with the variance between  181  Discus5ion  “control variables” and outcome.  Again, the direction of this  trend is opposite to that hypothesized.  Overall,  it appears that  even this stronger, proximal test of the role of life events does not support a stress—process model of the course of illness for either schizophrenia or affective psychosis. For schizophrenics, the meaning of a result opposite to that which was hypothesized is unclear:  one might speculate that  (currently unknown) subtypes have varying thresholds for original onset.  Thus, high-threshold schizophrenic patients would require  more stressors to precipitate illness onset, and subsequently would be less likely to relapse and more likely to have higher adaptive functioning throughout the course of the disorder. If this line of reasoning is valid,  it would change the  meaning of the terms in the regression equations in Table 26. true,  If  the life events at intake would serve as an index of the  ‘threshold’,  i.e. the individual baseline.  Accordingly, the  nine-month life events would be the only unequivocal index of environmental stressors.  Interpreted in this fashion, stressful  life events per se have no impact on 18 month-outcome. This interpretation points to the need for an analytic strategy based on within-subject comparisons.  Much of the  literature has used a repeated measures design, where the number of events prior to a psychotic episode is compared with events from an earlier period.  Such a method controls for individual  differences with respect to stressor-distress thresholds, as noted above.  182  Discussion  To summarize, two conclusions are possible from these short— term results.  First, the relationship between life events prior  to intake and outcome likely reflects an intra—individual ability to withstand stress.  Second, the months after a first psychotic  episode are a turbulent time.  The lack of a correlation with  outcome likely indicates that events in those first nine months do not affect the recovery process. model,  To assess a stress-process  it may be better to await later stages in the course of  illness. This study is not well aligned for direct comparison with other literature:  No other studies of schizophrenic patients  have examined the effect of life events on the course of illness in an aggregated fashion.  One way of retaining the  ‘aggregated’  feature would be to assess life events as they relate to a cumulative measure of relapse,  e.g.,  the number of subsequent  hospitalizations. Life events and onset.  When the methods used to assess the  effects of life events on onset or relapse (as used in the literature) are applied to these data, the results more or less support the “triggering” hypothesis for stressful life events in the onset of both schizophrenia and affective psychosis.  To test  this hypothesis, the number of life events occurring within six weeks of the onset of the prodromal period were contrasted with the number of events in the preceding six weeks  (i.e.,  7—13 weeks  prior to onset). Since the intake life events interview queried the year prior to the first treatment contact,  only those patients whose onset  183  Discussion  occurred within that time period could be the subject of this analysis.  The result of this constraint was a dramatic reduction  in sample size  (26 schizophrenic and 30 affective patients).  After a further loss in sample size due to missing data, the final sample sizes were 18 schizophrenic and 21 affective patients. The mean number of life events in the two six-week periods prior to the onset of the prodromal period are displayed in Figure 5.  Inspection of the figure suggests differences in both  diagnosis and over time. The distributions of the raw data shown in Figure 5, however, are skewed.  Therefore, to assess diagnosis by time differences  in the number of life events, the raw data were transformed using the log(n+1) convention.  Subsequently, a repeated measures  analysis of variance (ANOVA), with one between-subjects factor (diagnosis) and one within-subjects factor  (time), was conducted  with the log of the number of events as the dependent measure. The results indicate a main effect for diagnosis df=1,37, p<.0O5) and a main effect for time p<.OO5), but no interaction effect  (F=10.76,  (F=10.09, df=1,37,  (F=1.13, df=1,37, n.s.).  Subsequent paired i-tests indicate that affective patients had significantly more  (log) events in the six weeks leading up to  the onset of illness, compared to the previous six—week period (=2.94, df=20, p<.005).  A similar test for the schizophrenic  group indicates a nonsignificant trend towards a surplus of events in the period immediately prior to onset p=.O7).  For the schizophrenia patients,  (=1.56, df=17,  the increase in the  -o  N  C,.) C)  -o  CD  CD  3  H  0  -‘  U) 0  7c-  CD CD  0 )  0  -‘  0  (I)  CD CD  CD C-)  >  C,)  E  0  0 (.11 -  01  1\) 01  Number of Life Events 0) 01  (A)  U)  (D  m  (D  I  (D  -  (QUi  I-’.  cn  185  Discussion  (log) number of events in the weeks prior to onset, compared to the previous period, represents an effect size of =.43; the effect size for affective patients is =.65.  Thus,  it may be  that the failure to achieve significant results among the schizophrenic group is due to insufficient statistical power.  If  that Is the case, these results suggest that the life events have a triggering role for stressors in the onset of schizophrenia—-at least for those patients whose first treatment contact occurred within a year of the onset of the prodromal period.  A similar conclusion may be drawn for the affective  patients.  Further, these results indicate that affective  psychosis patients appear to be able to tolerate a higher baseline level of stressors, compared to schizophrenic patients. Finally,  there is some indication that schizophrenic patients do  not require fewer  (additional) stressors to precipitate psychosis  than do affective patients,  since the interaction term in the  repeated measures ANOVA is not significant.  Lateral Ventricle Size The literature suggests that some unknown proportion (range 7_949)  of people with schizophrenia have enlarged ventricles.  The best estimate for magnitude of enlargement is an effect size of  .60,  corresponding to 15—3O6 of original size.  Early MAP results  (Smith,  1986) show that neither  schizophrenic nor affective psychosis patients in this sample have larger ventricles,  compared both to a medical comparison  group and a matched normal comparison group.  Further, the  186  Discussion  distribution of VBR in both patient samples is not bimodal, indicating an absence of a distinct subset with enlarged ventricles.  other MAP results have appeared recently  Finally,  indicating temporal stability in ventricle size:  15 of the  first—episode schizophrenic patients were re-scanned after 1-3 years, and no further enlargement was detected (Sponheim et al., 1991).  In this light, these characteristics do not bode well for  VBR as a predictor of outcome. As it turns out, VBR does not predict outcome at five years. No relationship is apparent whether statistical tests used VBR in its continuous form, smallest ventricles  or compared patients with the largest and (the strategy used by a number of studies in  the area). The absence of any association between larger ventricles and poorer long—term outcome is surprising, since VBR does predict short—term outcome in this sample.  Katsanis et al.  (1991)  described the effect of large vs. small ventricles in outcome at 9 and 18 months, controlling for baseline functioning.  They  concluded that schizophrenics with larger ventricles had poorer short—term outcome*. not five-year outcome,  If ventricular size predicts 18—month, but then the biological vulnerability factor  indexed by VBR may have its principal effect in the early stages of illness.  If the median split approach is used to assess the relationship between ventricle size and five-year outcome, no difference is seen between schizophrenic patients with larger vs. smaller ventricles. *  187  Discussion  In the schizophrenia literature, there are few established results that unequivocally link brain structure with function. This is particularly true for the gross anatomy measured in computed tomography:  this technology is not capable of providing  answers to specific questions about etiology,  and hence is  largely unable to point to mechanisms by which the rather gross concept of “biological vulnerability” used here might have an effect on short—term, but not longer—term,  outcome.  Eve-Tracking Performance Smooth pursuit eye movement performance failed to show an association with five—year outcome.  RMS error scores were not  related to outcome when used in their continuous form or dichotomized at the “natural break”  in the distribution.  Nonetheless, eye—tracking does predict short-term outcome (lacono,  1988).  Taken together,  this suggests that the liability  associated with poor eye-tracking is more pronounced immediately after onset. Using the empirically-derived cut—off described above, the prevalence of smooth pursuit dysfunction in this sample is much less than previous reviewers suggested patients;  Holzman,  1992).  (5O—856 of schizophrenic  In this sample, depending on the data  used to determine DSM—III diagnosis, some 15—20% of the schizophrenics show smooth pursuit dysfunction, compared to 8% of the combined affective psychosis cases  (11-12% of patients with  188  Discussion  Major Depression*;  5—6% of those with Bipolar disorder).  Since  the use of these empirical techniques on several other independent samples has derived similar prevalence estimates (Clementz, Grove,  lacono,  & Sweeney,  1992),  the true prevalence  rate of SPEM dysfunction is likely much lower than that suggested previously. Smooth pursuit eye—movement dysfunction was developed as a genetic marker, not as a predictor of course and outcome.  The  dramatic reduction in prevalence in this and other recent samples (Clementz et al.,  1992) will likely force a reassessment of the  role of SPEM dysfunction in theories of etiology.  With previous  prevalence estimates of 50-85% in schizophrenic probands, and 45% in their clinically nonschizophrenic parents and siblings, there was a strong prima facie case for a monogenetic contribution to etiology of all schizophrenic cases 1986).  (Mathysse, Holzman,  & Lange,  With quantitatively-derived prevalence estimates in the  order of 15—20%,  SPEM dysfunction will retain its status as a  genetic marker--but as a marker for a specific, of schizophrenia.  Nonetheless,  familial subtype  it will likely continue to be  important to the etiology of this subtype.  Biological Risk in General While lateral ventricle size and oculomotor dysfunction are likely independent, they are not the only two biological  Among the probands whose DSi1-III diagnosis is Major Depression (with psychotic features), smooth pursuit dysfunction is largely confined to the subgroup whose concurrent RDC diagnosis is Schizoaffective. *  189  Discussion  anomalies in schizophrenia.  The use of either SPEM dysfunction  or VBR makes no assumptions about other possible risk factors that may also influence the course of illness. That both biological factors examined here indicate a liability for short—term, intriguing.  but not longer-term outcome is  Since both VBR and SPEM performance are stable over  time, this may suggest the development of adaptation or compensatory processes.  It may reflect psychosocial adaptation,  biomedical intervention,  or more fundamental reorganization of  neurophysiology.  In any case, the putative adaptation would take  time to develop, reducing the impact of any biological liability only after the first year or two of illness.  The biological  disadvantage may again become more prominent if adaptational factors diminish in later stages of illness.  Models of Course and Outcome  Although conceptually compelling, merging the psychosocial and biological domains in schizophrenia may have better been postponed until the component parts were more firmly in place. The emphasis and rationale of this dissertation was primarily to integrate the two domains, yet the role of life events—-although well established in precipitating psychotic episodes--has not yet been well established as a predictor of outcome.  190  Discussion  Of the four predictive hypotheses, only that regarding the relationship between life events and outcome is crucial to the stress—process and diathesis—stress models.  Since that  relationship is not present in either the schizophrenic or affective psychosis groups, assessing either model is not possible.  Significance of The Project  The original aim of this project, as proposed in June 1991, was to assess four sets of variables as they pertain to the five— year outcome of schizophrenia.  To the degree that the four  hypotheses regarding the prediction of outcome were supported, a further aim was to build an integrated, biopsychosocial model. Despite the general lack of support for the hypotheses, and subsequent failure of such a model to materialize, valuable results are present. Of the descriptive results, the parallel trajectory seen for the two diagnostic groups strikes at a myth about schizophrenia: Clinical lore includes a profound pessimism about a decline in schizophrenia that is unmatched by other disorders. results show that,  These  in comparison with psychotic affective  patients, schizophrenic patients do not decline more; start off at a disadvantage,  they simply  and decline or improve at the same  rate in the early stages of illness.  191  Discussion  The parallel trajectories may also serve to highlight a fundamental similarity between the two diagnostic groups: shared psychotic status.  their  The early presence of psychosis may  signify a profound biological  ‘shock’  that requires lengthy  realignment at the most fundamental levels:  Post—psychosis  neurochemistry may take time to re—establish homeostasis. Similarly, the psychosocial adaptation may be of comparable complexity, e.g.  learning to draw on social support or developing  new coping styles.  Finally,  optimizing pharmacological  intervention often takes experience over many months or several episodes. Of the hypotheses regarding the prediction of five—year outcome, one, pertaining to social relationships, received partial support.  The results indicate that friends and family  present at the onset of illness continue to be important five years later.  While failing to achieve statistical significance,  it is nonetheless a remarkable finding, given the lengthy passage of time.  In particular, that specific components identified as  predictive of five-year outcome are consistent with those pertaining to 18-month outcome provides a partial replication in an area that has seen little confirmation of results. That this association overlaps with other factors in predicting five-year outcome is no less important than the unique and independent association of social relationships with short term outcome.  Social relationships are amenable to intervention  throughout the course of illness, whereas sex, diagnosis, and baseline functioning are not.  192  Discussion  The importance of the phases of illness is apparent in the descriptive results.  That there is first a decline in short-term  adaptive functioning, and then a rise over and above baseline functioning, provides a longitudinal view of the early course of illness.  If nothing else,  it provides support for the persistent  efforts of afflicted persons and their families.  It shows that  the demands of dealing with the challenge of a complex disorder will likely be greater in the earlier phases of the disorder. Moreover,  the tremendous demands in the early phases will be  followed by better functioning in the later phases. In the biological realm, no previous study has examined the prognostic significance of VBR or SPEM at multiple follow-up points.  That two putative indicators of biological liability,  each stable over time, both predict short— but not longer-term outcome, together point to possible compensatory processes as described above.  Limitations  While the MAP Project in general and this thesis in particular have a number of strengths, there are also some notable limitations.  They include aspects pertaining to the research  design itself, the nature of some of the measures, and the source of data.  193  Discussion  The research design,  though prospective and longitudinal,  is  nonetheless based on a “snapshot” approach to course and outcome. The dependent measure in this thesis,  the DSM-III Axis V rating,  and many measures of outcome in 9- and 18-month follow-up interviews, reflect a cross-sectional view,  and as such cannot  capture the rich and complex experience of illness that would be summarized in a “moving picture” view of course and outcome. Even as a snapshot measure,  the Axis V--like all global  ratings--combines the multidimensional concept of outcome into a single dimension. Stoffelmayr et al.,  Previous studies  (Strauss & Carpenter.  1983) have demonstrated that dimensions of  outcome are best considered as “open-linked systems”. example,  1977;  For  it is possible for a person to have good functioning at  work but poor social functioning,  or the reverse.  combining two dimensions into a single rating,  Axis V, by  loses complexity  inherent in the lives of individual patients. As a global rating, Axis V rates primarily the quantitative aspects of social and occupational functioning.  For example,  similar scores would be derived for a patient who had worked full—time as a civil engineer prior to onset, yet who is only able to manage full—time work as a car wash attendant following the onset of schizophrenia.  Occupational functioning is itself a  multidimensional concept, where other aspects  (not rated in Axis  V) might include the complexity of the job and the quality of performance in fulfilling the task. Use of the Axis V rating in this particular data set indicates one final limitation,  in that five-year social functioning data  194  Discussion  were not available for approximately one-third of the cases.  The  result is that the dependent measure in the study relies more heavily on occupational functioning than it otherwise would. Because the five-year data reflected only current functioning, we could not control for treatment and medication.  The  limitations of the abbreviated/proxy data at five years precluded even a retrospective recounting of medication and treatment in the past year  (data that were collected in the full five-year  interview). The final issue pertaining to research design is the amount of time that passed between assessment of predictor variables and five-year outcome.  For example,  any particular life event will  likely have its greatest effect within weeks of its occurrence. The results in this study likely indicate that the effect of life events leading up to 18—month follow-up is negligible at five years.  Similarly, the effect of social relationships was  stronger at 18 months compared to five years, when the effect has diminished to a nonsignificant trend. an intermediate point, more reasonable effort.  e.g.,  Measurement of outcome at  three or four years, may have been a  Alternatively, a reassessment of social  relationships at an intermediate point would have been instructive. The nature of the life events measure used in this study, the Social Readjustment Rating Scale  (Holmes & Rahe,  1967),  is  another limitation,  inasmuch as it was designed for nonclinical  populations.  there may be events of particular  Thus,  significance for psychiatric populations that are not measured  195  Discussion  with sufficient sensitivity; alternative instruments such as the Psychiatric Epidemiology Rating Interview (PERI) Life Events Scale (B.S. Dohrenwend, Krasnoff, Askenasy,  & Dohrenwend,  1978)  are available. The self—report nature of life events measures is another limitation,  in that they are susceptible to individual  differences in the interpretation of what constitutes a minor life event.  Minor events such as a mild reprimand from a  supervisor or a leaking waterpipe at home may or may not be interpreted as a life event.  At this end of the continuum, the  occurrence of an event is confounded with its meaning.  In our  study, the use of trained interviewers and a structured format minimized this danger.  Future Research  There are a number of directions that one might take upon conclusion of this project, some of which are possible within this data set.  In the psychosocial domain,  it would be  instructive to examine the possible differential effects of social relationships and life events on various subtypes in schizophrenia.  Cohen and Sokolovsky (1978),  for example,  found  that social relationships had no effect on the prognosis of severely impaired patients.  Discussion  196  If results from investigations such as these are to have clinical applications,  a more fine-grained analysis of social  relationships data will be important.  Here, we might consider  the structural characteristics of networks,  e.g. network Given the  multiplexity or density as they relate to outcome.  magnitude of the prospective correlations with five-year outcome, this work would need to be based on 18-month outcome from the MAP Project, or on data from a new sample. The dual nature of the role of family members in the social network is intriguing.  If the nonsignificant correlations with  five—year outcome are seen as supporting the significant correlations with 18-month outcome, there one can interpret the findings as indicating only that greater numbers of kin show a tendency to exert a small but significant negative influence on outcome.  Reversing the wording is probably more constructive for  afflicted persons and their family:  smaller numbers of kin are  beneficial, compared with larger numbers of kin.  No evidence has  been presented that can comment on the effect of an increased or decreased presence of any particular family member. Superficially, this result bears some similarity with the ‘expressed emotion’  (EE) phenomenon  (Leff & Vaughn,  1985),  an  avenue of investigation that must be explored in a future project.  The similarity, however,  components of style’  ‘critical comments’,  is only superficial:  the EE  ‘overinvolved’, and  ‘affective  characterize the nature of a transaction, and do not  comment on the role of greater or fewer family members making those comments.  A direct comparison of EE and the social  197  Discussion  relationships as measured here is necessary before claiming any overlap in the two concepts.  Indeed,  before measuring EE, a  preliminary avenue for future research would be to assess the network—related characteristics of any family members who do have a negative influence on outcome.  Alternatively,  it may be that  no individual family member has a negative effect; rather,  it may  simply be that--for schizophrenic probands--family can sometimes be too much of a good thing. If both social relationships and stressful life events had predicted outcome, a consideration of their interaction would have constituted a stress-process model of the course of illness. Such a notion continues to be worthy of pursuit.  A more  reasonable assessment of the role of life events, however, would be on short—term outcome.  For example,  life events could be  assessed every six months over a two—year period.  In this way,  the interaction between life events and social relationships may be assessed.  In such a fashion,  resources such as coping styles,  one could also consider personal locus of control, or mastery as  they affect the relationship between stressors and distress over t i me. In the biological realm, a different strategy may be more productive.  While brain morphology indicators have told us much  about the likely primacy of brain dysfunction, and while smooth pursuit dysfunction may be instructive in understanding genetic etiology, neither phenomenon was developed to predict course and outcome.  Even if both predict short-term but not long-term  outcome, the results are difficult to interpret.  It may be more  Discussion  198  fruitful to use investigative tools that have been designed to relate structure to function, or fast MRI.  e.g. neuropsychological assessment  199 REFERENCES  Al Khani, M., Bebbington, P., Watson, J., & House, F. (1986). Life events and schizophrenia: A Saudi Arabian study. British Journal of Psychiatry, 148, 12-22. Alloway, R., & Bebbington, P. (1987). The buffer theory of social support: A review of the literature. Psychological Medicine, II, 91—108. American Psychiatric Association (APA). statistical manual of mental disorders, DC: Author.  (1980). Diagnostic and (3rd ed.). Washington,  Andreasen, N., Olsen, S., Dennert, J., & Smith, M. (19 82). Ventricular enlargement in sch izophrenia: Relationship to positive and negative sympto ms. American Journal of Psy chiatry, U., 297—302. Andreasen, N., Nasrallah, H., Dunn V., Olsen, S., Grove, W.., Erhardt, J., Coffman, A., & Crossett., J. (1986). Structural abnormalities in the fronta l system in schizophrenia: A magnetic resonance imaging study. Archives of General Psychia try, 43, 136—44. Andreasen, N., Erhardt, J., Swayze, V., (1990). Magnetic resonance imaging in schizop hrenia: The path ophysiologic significance of structural abnormalities. Archives of General Psychiatry, ill Angermeyer, M., Kuehn, L., & Goldstein, J. (1989). Gender differences in schizophrenia: Rehospitalization and com munity survival. Psychological Medicine, , 365— 82. Bartfai, A., Levander, S., Nyb eck, H., Bergrren, B-M., & Schalling, D. (1986). Smooth pursuit eye tracking, neuropsychological test perform ance, and computed tomograph y in schizophrenia. Psychiatry Research, j, 49—62. Beiser, M., Bean, G., Erickson, D., Zhang, J., lacono, W., & Rector, N. (1993). Biological and psychosocial pre dictors of iob performance following a first episode of psychosis. Manuscript submitted for publication. Beiser, M., Erickson, D., Fle ming, J., & lacono, W. (19 93). Establishing onset in psychosis. American Journal of Psychia try, in press. Beiser, M., & lacono, W. (1990) . An update on the epidemiolo gy of schizophrenia. Canadian Journal of Psychiatr y, 35, 657-668.  200  Predicting more of the people more Bern, D., & Funder, D. (1978). Assessing the personality of situations. of the time: Psychological Review, 85, 485—501. Steroids Benston, J., Reza, M., Winter, J., & Wilson, G. (1978). and apparent cerebral atrophy on computer tomography scans. Journal of Computer Assisted Tomography, 2, 16—23. Besson, J., Corrigan, F., Cherryman, G., & Smith, N. (1987). Nuclear magnetic resonance brain imaging in chronic British Journal of Psychiatry, jjQ, 161-163. schizophrenia. Bird, J. (1982). Computerized tomography, atrophy and dementia: A review. Progress in Neurobiology, li 91—115. Predicting the outcome in schizophrenia. Bland, R. (1982). Canadian Journal of Psychiatry, 27, 52-62. Sex differences in a Blishen, B., & Carroll, W. (1978). Canadian Review socioeconomic index for occupations in Canada. of Sociology and Anthropology, , 352-371. Blishen, B., & McRoberts, H. (1976). A revised socioeconomic index for occupations in Canada. Canadian Review of Sociology and Anthropology, II, 71-79. Bleuler, N. (1978). The schizophrenic disorders: Long—term New Haven: patient and family studies (S. Clemens, Trans.). Yale University Press. Premorbid functioning Bromet, B., Harrow, N., & Kasl, S. (1974). Archives of and outcome in schizophrenics and nonschizophrenics. General Psychiatry, Q, 203—207. Brown, G., & Birley, J. onset of schizophrenia. , 203—214.  Crises and life changes and the (1968). Journal of Health and Social Behaviour,  Brown, G., & Harris, T. (1978). London: Tavistock. depression.  The social origins of  Life Brown, G., Sklair, F., Harris, T., & Birley, J. (1973). Some methodological Part 1. events and psychiatric disorders: issues. Psychological Medicine, 3, 74—87. Prognosis in Burstein, A., Adams, R., & Chapman, L. (1974). Journal of Nervous and schizophrenia: A 5—year follow—up. 137—140. Mental Disease, Classification and outcome in Cancro, R., & Sugerman, A. (1968). process/reactive schizophrenia. Comprehensive Psychiatry, , 227—32.  201 Caplan, G. New York:  Support systems and community mental health. (1974). Behavioural Publications.  Carpenter, W., Heinrichs, D., & Wagman, A. (1988). Deficit and non-deficit forms of schizophrenia: The concept. American Journal of Psychiatry, 145, 578—583. Chung, R., Langeluddecke, P., & Tennant, C. (1986). Threatening life events in the onset of schizophrenia and schizophreniform psychosis. British Journal of Psychiatry, jj, 680-85. Ciompi, L. (1980). Three lectures on schizophrenia. Journal of Psychiatry, JJj, 413—420.  British  Ciompi, L. (1988). Learning from outcome studies: Toward a comprehensive biological-psychological understanding of schizophrenia. Schizophrenia Research, 1, 373—384. Clementz, B., Grove, W., lacono W., & Sweeney, J. (1992). Smooth-pursuit eye movment dysfunction and liability for schizophrenia: Implications for genetic modelling. Journal of Abnormal Psychology, 101, 117-29. Cobb, S. (1976). Social support as a moderation of life stress. Psychosomatic Medicine, , 300—314. Cohen, C., & Sokolovsky, J. (1978). Schizophrenia and social networks: Ex-patients living in the inner city. Schizophrenia Bulletin, 4., 546—60. Cohen, 3. sciences.  Statistical power analysis for the behavioural (1977). New York: Academic Press.  Cohen, S., & Syme, S. (Eds.) (1985). Orlando, Florida: Academic Press.  Social support and health.  Coyne, 3., & DeLongis, A. (1986). Going beyond social support: Journal of The role of social relationships in adaptation. Consulting and Clinical Psychology, j, 454-46. Molecular pathology of schizophrenia: Crow, T. (1980). British Journal of Medicine, than one disease process. 68.  More  ZQ,  66-  The way ahead. In P. Aetiology of psychosis: Crow, T. (1988). The maior Bebbington & P. McGuffin (Eds.), Schizophrenia: Oxford: Heinemann Publishers. issues, pp. 127-34. Daniel, D., & Weinberger, D. (1991). Ex Multi Uno: A case for In C. Tamminga & neurobiological homogeneity in schizophrenia. Advances in Neuropsychiatry and S. Schulz (Eds.), Volume 1. Schizophrenia Research, pp. 227Pychopharmacology: Raven. 35. New York:  202  Social support as Davidson, T., Bowden, L., & Tholen, T. (1979). Archives of Physical Medicine a moderator in burn rehabilitation. and Rehabilitation, Q, 556. Day, R., Nielsen, J., Korten, A., Ernberg, G., Dube, K., Gebhart, J., Jablensky, A., Leon, C., Marsella, A., Olatawura, M., Sartorius, N., Stromgren, E., Takahashi, R., Wig, N., & Wynne, L. Stressful life events preceding the onset of (1987). A cross-national study from the World Health schizophrenia: Organization. Culture. Medicine and Psychiatry, ji, 123-205. DeLisi, L., Goldin, L., Hamovit, J., Maxwell, E., Kurtz, D., A family study of the association of Gershon, E. (1986). Archives of increased ventricular size with schizophrenia. General Psychiatry, il 148-53.  &  DeLisi, L., Schwartz, C., Targum, S., Cannon-Spoor, E., Ventricular brain Weinberger, D., & Wyatt, R. (1983). enlargement and outcome of acute schizophreniform disorder. Psychiatry Research, , 169-71. Dewan, M., Haldipur, C., Lane, E., Ispahani, A., Boucher, M., & Comprehensive MaJor, L. (1988). Bipolar affective disorder I. quantitative computed tomography. Acta Psychiatrica Scandinavica, 77, 670—676. Docherty, J., Van Kammen, D., Sins, S., & Marder, S. (1978). American Journal of Stages of onset of schizophrenic psychosis. 420—426. Psychiatry, Dohrenwend, B.P., & Egri, G. (1981). events and episodes of schizophrenia. 12—23.  Recent stressful life Schizophrenia Bulletin,  7,  Dohrenwend, B.S., Knasnoff, L., Askenasy, A., & Dohrenwend, B.P. Exemplification of a method for scaling life events: (1978). Journal of Health and Social The PERI life events scale. Behaviour, , 205—29. Cerebral ventricular Dolan, R., Calloway, S., & Mann, A. (1985). size in depressed patients. Psychological Medicine, 15, 873—78. The structure of social relationships: Duncan—Jones, P. (1981a). Social Psychiatry, , Analysis of a survey instrument, Part 1. 55—61. The structure of social relationships: Duncan—Jones, P. (1981b). Social Psychiatry, 16, Analysis of a survey instrument, Part 2. 143—149. Stress and vulnerability. In Endler, N., & Edwards, J. (1988). C. Last & M. Hersen (Eds.), Handbook of anxiety disorders, pp. Pergamon. 278-92. New York:  203  Erickson, D.., Beiser, H., lacono, W., Fleming, J., & Lin, T. The role of social relationships in the course of first— (1989). American Journal episode schizophrenia and affective psychosis. of Psychiatry, )j, 1456—1461. Farmer, A., Jackson, R., McGuffin, P., & Story, P. (1987). Cerebral ventricular enlargement in chronic schizophrenia: British Journal of Psychiatry, Consistencies and contradictions. Q., 324—30. Finlay-Jones, R. (1988). Life events and psychiatric illness. In A. Henderson (Ed.), Handbook of Social Psychiatry, pp. 27-40. Elsevier. New York: Goldstein, J. (1988). Gender differences in the course of schizophrenia. American Journal of Psychiatry, J.j., 684-89. Goldstein, J., Santangelo, S., Simpson, J., & Tsuang, M. (1990). The role of gender in identifying subtypes of schizophrenia. Schizophrenia Bulletin, 16, 263—75. The Gottesman, I., & Shields, J. (1982). Schizophrenia: Cambridge: Cambridge tiniveristy Press. epigenetic puzzle. Greenblatt, M., Becerra, R., networks and mental health : Psychiatry, U2, 977—84. Gureje, 0., in Nigeria: Psychiatry,  Social & Serafetinides, E. (1982). An overview. American Journal of  Life events and schizophrenia & Adewumni, A. (1988). British Journal of A controlled investigation. 153, 367—375.  Hammer, M. (1981). Social supports, social networks, schizophrenia. Schizophrenia Bulletin, j, 45-57.  and  Harding, C., Brooks, G., Ashikaga, T., Strauss, J., & Brier, A. (1987). The Vermont longitudinal study of persons with severe mental illness: I. Long—term outcome of subjects who retrospectively met DSM—III criteria for schizophrenia. American Journal of Psychiatry, jjj, 727—735. Haug, J. disease.  Pneumoencephalographic studies in mental (1962). Acta Psychiatrica Scandinavica, IQ (Suppi. 165), 9-106.  Henderson, A. (1980). A development in social psychiatry: The systematic study of social bonds. Journal of Nervous and Mental Disorder, 63—69. Henderson, A. (1984). Interpreting the evidence on social support. Social Psychiatry, , 49—52. Henderson, A. (1988). An introduction to Social Psychiatry. York: Oxford linversity Press.  New  204  Social support: The Henderson, A., & Brown, G. (1988). In A. Henderson (Ed.), Handbook of hypothesis and the evidence. New York: Elsevier. Social Psychiatry, pp. 73-85. Henderson, A., Duncan-Jones, P., Byrne, D., & Scott, R. (1980). The Interview Schedule for Measuring social relationships: Social Interaction. Psychological Medicine, 10, 723—734. Hirsch, S., Bowen, 3., Emami, J., Cramer, P., Jolly, A., & Haw, C. (1993). The effect of life events and medication in the Schizophrenia aetiology of schizophrenic relapse [Abstract]. Research, 9, 266. The Social Readjustment Rating Holmes, T., & Rahe, R. (1967). 213-218. Journal of Psychosomatic Research, Scale. Behavioural markers of schizophrenia Holzman, P. (1992). Journal of Psychiatric research useful for genetic studies. Research, , 427—45. Cerebrospinal fluid Houston, 3., Maas 3., & Bowden C. (1986). HVA, central brain atrophy, and clinical state in schizophrenia. Psychiatry Research, j, 207-14. Negative symptoms Husted, J., Beiser, M., & lacono, W. (1992). and the early course of schizophrenia. Psychiatry Research, jj, 215—22. lacono, W. (1988). Eye movements abnormalities in schizophrenia and affective disorders. In C. Johnston & F. Pirozzolo (Eds.), Hilisdale, The neuropsychology of eye movements, pp. 115—145. NJ: Erlbaum. Two year retest stability of lacono, W., & Lykken, D. (1981). eye-tracking performance and a comparison of electro-oculographic and infrared recording techniques. Psychophysiology, j., 49-55. lacono, W., Moreau, M., Beiser, M., Fleming, 3., & Lin, T. Smooth pursuit eye tracking in first epeisode psychotic (1992). Journal of Abnormal Psychology, patients and their relatives. 104—16. 121, lacono, W., Smith, N., Moreau, M., Beiser, M., Fleming, J., Lin, Ventricular and sulcal size at onset of T., & Flak, B. (1988). 820—24. Journal of Psychiatry, psychosis. American Illowsky, B., Juliano, D., Bigelow, L., & Weinberger, D. (1988). Stability of CT scan findings in schizophrenia: Results of an 8of year follow-up study. Journal Neurology, Neurosurgery, and 209—213. Psychiatry, j, Recent life events and acute Jacobs, A., & Meyers, 3., (1976). Journal of Nervous schizophrenic psychosis: A controlled study. and Mental Disease, 162, 75-87.  205  Computed tomography in the Jacoby, R., & Levy, R. (1980). British Journal of Psychiatry, 3. Affective disorders. elderly: iJ., 270—75. Jeste, D., Kleinman, J.., Potkin, S., Luchins, D., & Weinberger, Subtyping the schizophrenic syndrome. D. (1982), Ex Uno Multi: 17, 199—222. Biological Psychiatry, Johnston, C., & Pirozzolo, F. (Eds.) (1988). Hilisdale, NJ: Eribaurn. of eve movements.  The neuropsychology  Johnstone, E., Crow, T., Frith, C., Husband, J., & Kreel, L. Cerebral ventricle size and cognitive impairment in (1976). Lancet, ii, 924—926. schizophrenia. Kaiser, M. (1974). Psychometrika, 39,  An index of factorial simplicity. 31—36.  The relationship Katsanis, J., lacono, W., & Beiser, M. (1991). of lateral ventricle size to psychophysiological measures and Psychiatry Research, 37, 115-29. short-tern outcome. Kemali, D., Mai, M., Galderisi, S., Ariano, M., Cesarelli, M., Milici, N., Salvati, A., Valente, A., & Volpe, M. (1985). Clinical and neuropsychological correlates of cerebral Journal of Psychiatric ventricular enlargement in schizophrenia. Research, ), 587-96. Kemali, D., Maj, M., Galderisi, S., Salvati, A., Starace, F., Clinical, biological, and Valente, A., & Pirozzi, R. (1987). neuropsychological features associated with cerebral ventricular Psychiatry enlargement in DSM-III schizophrenic disorder. Research, II, 137—49. Kufferle, B., Grunberger, J., Linzmayer, L., & Saletu, B. (1988). Distinction between positive and negative schizophrenia in the Schizophrenia light of psychophysiological measurements. Research, 1, 179-180. Lazarus, R., & Folkman, Springer. New York:  S.  (1984).  Stress,  appraisal and coping.  Leff, J.., Hirsch, S., Gaind, R., Rhode, P., & Stevens, B. (1973). British Life events and maintenance therapy in schizophrenia. Journal of Psychiatry, 123, 659—60. Leff, J., Kuipers, E., Berkowitz, R., Vaughn, C., & Sturgeon, D. Life events, relatives’ expressed emotion and (1983). Psychological maintenance neuroleptics in schizophrenic relapse.  Medicine, , 799—806. Leff, J., New York:  & Vaughn, C. Guilford.  (1985).  Expressed emotion in families.  206 Social supports——The consequences of Lieberman, M. (1985). Journal of Consulting and Clinical Psychology, psychologizing. j, 461—65. The symptoms of chronic schizophrenia: A re Liddle, P. (1987). British Journal examination of the postitive—negative dichotomy. of Psychiatry, jL, 145-51. Lorei, T., & Guel, L. (1973). Demographic characteristics as predictors of post—hospital employment and readmission. Journal of Consulting and Clinical Psychology, jQ, 426-430. Lateral Luchins, D., Lewine, R., & Meltzer, H. (1984). ventricular size, psychopathology, and medication response in the psychoses. Biological Psychiatry, , 29-44. A comparison of CT findings Luchins, D., & Meltzer, H. (1986). in acute and chronic ward schizophrenics. Psychiatry Research, 12.., 7—14. MacMillan, J., Crow, T., Johnson, A., & Johnstone, E. (1986). III. The Northwick Park study of first—episode schizophrenia: Short—term outcome in trial entrants and trial—eligible patients. British Journal of Psychiatry, jj, 128-33. Mathew, R., Parain, C., Prakash, R., Kulkarni, M., Logan, T., & Wilson, W. (1985). A study of the septum pellucidum and corpus Acta Psychiatrica callosum in schizophrenia with MR imaging. Scandinavica, , 414-21. The genetic Matthyse, S., Holzman, P., & Lange, K. (1986). Application of Mendelian latent transmission of schizophrenia: structure analysis to eye-tracking dysfunction in schizophrenia Journal of Psychiatric Research, , 57— and affective disorder. 65. Predictors of shorter-, medium-, and McGlashan, T. (1986). American Journal of longer-term outcome in schizophrenia. Psychiatry, Ui, 50-55. McGlashan, T. (1988). A selective review of recent North American long—term follow—up studies of schizophrenia. Schizophrenia Bulletin, j, 515—542. Medalie, J., 10,000 men.  Angina pectoris among & Goldbourt, U. (1976). American Journal of Medicine, EQ., 910—921.  Schizotaxia, schizoptypy, schizophrenia. Meehl, P. (1962). American Psychologist, 17, 827—38. A promising direction for Social networks: Mueller, D. (1980). research on the relationship of the social environment to 147—161. psychiatric disorder. Social Science and Medicine,  207  Murphy, H., & Raman, A. (1971). The chronicity of schizophrenia in indigenous tropical peoples: Results of a 12 year follow—up survey in Mauritius. British Journal of Psychiatry, JJ&, 489497. Nasrallah, H., Kuperman, S., Hamra, B., & McKalley-Whitters, M. Clinical differences between schizophrenic patients with (1983). and without large cerebral ventricles. Journal of Clinical Psychiatry, 44, 407—409. Nasrallah, H., McCalley—Whitters. M., & Jacoby, C. (1982). Cortical atrophy in schizophrenia and mania: A comparative study. Journal of Clinical Psychiatry, j, 439-441. Neale, Wiley.  J.,  & Oltmanns,  T.  (1980).  Schizophrenia.  New York:  Nuckolls, K., Cassel, J., & Kaplan, B. (1972). Psychosocial assets, life crises and the prognosis of pregnancy. American Journal of Epidemiology, 95, 431—441. Nuechterlein, K. (1987). Vulnerability models for schizophrenia: State of the art. In H. Hafner, W. Gattaz, & W. Janzarik (Eds.), Search for the causes of schizophrenia, pp. 297-316. New York: Springer. Nuechterlein, K., & Dawson, M. (1984). A heuristic vulnerability! stress model of schizophrenic episodes. Schizophrenia Bulletin, Q., 300—312. Olsen, S., Nasrallah, H., Coffman J., & Swartzkopf, S. (1990). CT and MRI abnormalities in schizophrenia: Relationship with negative symptoms. In J. Greden & R. Tandon (Eds.), Negative Pathophysiology and clinical schizophrenic symptoms: Washington, DC: American Psychiatric implications, pp. 147-60. Association Press. Owen, M., Lewis, S., & Murray, R. (1989). Family history and A case cerebral ventricular enlargement in schizophrenia: British Journal of Psychiatry, jj, 629—34. control study. Owens, D., Johnstone, E., Crow, T., Frith, C., Jagoe, J., & Kreel, L. (1985). Lateral ventricle size in schizophrenia: Relationship to the disease process and its clinical manifestations. Psychological Medicine, , 27—41. Pandurangi, A., Bilder, R., Rieder, R., Mukherjhee, S., & Hamer, R. (1988). Schizophrenic symptoms and deterioration in relation Journal of Nervous and Mental to computed tomographic findings. Disease, 200—206. Pattison, E., DeFrancisco, D., Wood, P., Frazier, H., & Crowder, J. (1975). A psychosocial kinship model for family therapy. 1246—1251. American Journal of Psychiatry,  208  Pearison, G., Garbacz, D., Breakey, W., Ahn, H., & DePaulo, .J. Lateral ventricular enlargment associated with (1984). persistent unemployment and negative symptoms in both schizophrenia and bipolar disorder. Psychiatry Research, , 19. Pearlson, G., Garbacz, D., Moberg, P., Ahn, H., & DePaulo, J. Symptomatic, familial, perinatal, and social correlates (1985). of computerized axial tomography (CAT) changes in schizophrenics and bipolars. Journal of Nervous and Mental Disease, 131, 42-50. Ventricular volume Penn, R., Belanger, M., & Yasnoff, W. (1978). Annals of Neurology, , 216-223. in man computed from CAT scans. Pfefferbaum, A., Zipursky, R., 1dm, K., Zatz, L., Stahl, S., & Computed tomographic evidence for Jernigan, T. (1988). generalized sulcal and ventricular enlargement in schizophrenia. Archives of General Psychiatry, j., 633—640. Quantity or Porrit, D. (1979). Social support in crisis: quality? Social Science and Medicine, (6A), 715—721. .  Prudo, R., & Blum, H. (1987). Five year outcome and prognosis in schizophrenia. British Journal of Psychiatry, jQ., 345—354. Raz, S., & Raz, N. (1990). Structural brain abnormalities in the A quantitative review of the evidence from major psychoses: computerized imaging. Psychological Bulletin, j.Q, 93-108. Reddy, R., Mukherjee, S., Schnur, D., Chin, J., & Degreef, G. History of obstetrical complications, family history, (1990). and CT scan findings in schizophrenic patients. Schizophrenia Research, j, 311-14. Reveley, A., Reveley, M., Clifford, C., & Murray, R. (1982). Cerebral ventricular size in twins discordant for schizophrenia. Lancet, , 540—541. Rieder, R., Mann, L., Weinberger, D., van Kammen, D.,, & Post, R. Computed tomographic scans in patients with (1983). schizophrenia, schizoaffective, and bipolar affective disorder. Archives of General Psychiatry, jQ, 735—739. The Robins, L., Helzer, J., Croughan, J., & Ratcliff, K. (1981). Its history, characteristics NIMH Diagnostic Interview Schedule: and validity. Archives of General Psychiatry, , 381-389. Roy—Byrne, P., Post, R., Keilner, C., Joff, T., & Vhde, T. Ventricular—brain ratio and life course of patients with (1988). affective disorder. Psychiatry Research, , 277-284.  209  Sacchetti, E., Caizeroni, A., Vita, A., Terzi, A., Pollastro, F., & Cazzulo, C. (1992). The brain damage hypothesis of the seasonality of births in schizophrenia and major affective disorders: Evidence from computerized tomography. British Journal of Psychiatry, 160, 390—97. Schradle, S., & Dougher, M. (1985). Social support as a mediator of stress: Theoretical and empirical issues. Clinical Psychology Review, , 641—661. Schulz, S., Koller, M., Kishore, P., Hamer, R., Gehi, 3., & Friedel, R. (1983). Ventricular enlargement in teenage patients with schizophrenia spectrum disorder. American Journal of Psychiatry, 140, 1592—1595. Scott, M., Golden, C., Ruedrich, S., & Bishop, R. (1983), Ventricular enlargement in major depression. Psychiatry Research, , 91—93. Seeman, M. (1982). Gender differences in schizophrenia. Canadian Journal of Psychiatry, , 107-112. Seidman, L. (1983). Schizophrenia and brain dysfunction: An integration of recent neurodiagnostic findings. Psychological Bulletin, j, 195—238. Seidman, L., Sokolove, R., McElroy, C., Knapp, P.., & Sabin, T. Lateral ventricular size and social network (1987). differentiation in young, nonchronic schizophrenic patients. American Journal of Psychiatry, jj, 512—514. Shagass, C., Roemer, R., & Amadeo, M. (1976). Eye-tracking performance and engagement of attention. Archives of General Psychiatry, 33, 121-125. Shaw, R. Toronto:  Intermetropolitan migration in Canada. (1985). NC Press.  Shima, S., Shikano, T., Kitamura, T., Masuda, Y., Tsukumo, Depression and ventricular Kanba, S., & Asai, M. (1984). enlargement. Acta Psychiatrica Scandinavica, IQ., 275-77.  T.,  Siever, L., Coursey, R., Alterman, I., Buchsbaum, M., & Murphy, Vulnerability D. (1984). Impaired smooth pursuit eye movement: marker for schizotypal personality disorder in a normal volunteer population. American Journal of Psychiatry, jJi, 1560-1566. Siever, L., van Kammen, D., Linnoila, M., Alterman, I., Hare, T., Smooth pursuit eye movement disorder and & Murphy, D. (1986). its psychobiologic correlates in unmedicated schizophrenics. Biological Psychiatry, 2J, 1167—1174.  210  A computerized tomographic evaluation of brain Smith, G. (1986). Unpublished morphology in first episode psychotic patients. doctoral dissertation, University of British Columbia, Vancouver. Morphological and Smith, R., & Maser, J. (1983). neuropsychological abnormalities as predictors of clinical response to psychotropic drugs. Psychopharmacological Bulletin, 505—09. .i, Sokolovsky, 3., Cohen, C.., Berger, D., & Geiger, J. (1978). Personal networks of ex-mental patients in a Manhattan SRO hotel. 5-15. Human Organization, 3, Sponheim, S., lacono, W., & Beiser, M. (1991). ventricular size after the onset of psychosis. Research, 40, 21-29. Spring, issues.  Stress and schizophrenia: B. (1981). 24—33. Bulletin, Schizophrenia  Stability of Psychiatry  Some definitional  ,  Long—term prognosis and follow-up in Stephens, J. (1978). Bulletin, ,j, 25—47. schizophrenia. Schizophrenia Premorbid Stoffelmayr, B., Dillavou, D., & Hunter, J. (1983). Journal of Consulting and functioning and schizophrenia. Clinical Psychology, j, 338—352. The prediction of outcome Strauss, J., & Carpenter, W. (1977). III. Five year outcome and its predictors. in schizophrenia: Archives of General Psychiatry, 34, 159—163. Sturgeon, D., Turpin, G., Kuipers, L., Berkowitz, R., & Leff, .3. Psycophysiological responses of schizophrenic patients (1984). A follow-up study. to high and low expressed emotion relatives: British Journal of Psychiatry, 145, 62—69. Suddath, R., Christison, G., Torrey, E., & Weinberger, D. (1990). Anatomical abnormalities in the brains of monozygotic twins New England Journal of Medicine, discordant for schizophrenia. 789—794.  aaz.,  A selective Szymanski, S., Kane, 3., & Lieberman, J. (1991). Schizophrenia review of biological markers in schizophrenia. Bulletin, U, 99-111. CT scanning and Takahashi, R., manage, V., & Kato, N. (1982). In C. Perris, G. Struwe, & the investigation of schizophrenia. B. Jansson (Eds.), Biological Psychiatry, pp. 259-68. Targum, S., Rosen, L., DeLisi, L., Weinberger, D., & Citrim, C. (1983). Cerebral ventricular size in Major Depressive disorder: Biological Psychiatry, Association with delusional symptoms. 329—336.  ,  211  Hydrocephalus and atrophy. TerBrugge, K., & Rao, K. (1983). S. Lee & K. Rao (Eds.), Cranial computed tomography, 170-199. McGraw Hill. Toronto:  In  Dimensions of life events that influence Thoits, p. (1983). An evaluation and synthesis of the psychological distress: Trends in In H. Kaplan (Ed.), Pychosocial stress: literature. New York: Academic. theory and research, pp. 33-103. Social networks, support, and coping: Tolsdorf, C. (1976). exploratory study. Family Process, j, 407—417.  An  Direct, indirect, and moderating effects of Turner, R. (1983). social support on psychological distress and associated Trends in In H. Kaplan (Ed.), Pychosocial stress: conditions. Academic. New York: theory and research, pp. 105-55. Physical disability and Turner, R., & McLean, p. (1989). Rehabilitation Psychology, jj, 225—242. psychological distress. Computed Turner, S., Toone, B., & Brett—Jones, J. (1986). Preliminary tomographic changes in early schizophrenia: findings. Psychological Medicine, , 219—26. Prospective prediction of schizophrenic Vaillant, G. (1964). Archives of General Psychiatry, 11, 509—518. remission. Ventura, J., Nuechterlein, K., Lukoff, D., & Pederson-Hardesty, J. (1989). A prospective study of stressful life events and Journal of Abnormal Psychology, , 407schizophrenic relapse. 411. Vita, A., Saccetti, E., Valvassori, G., & Cazzullo, C. (1988). Brain morphology in schizophrenia: a 2- to 5-year CT scan followActa Psvchiatrica Scandinavica, Z., 618-621. up study. Is outcome for schizophrenia better in Waxler, N. (1977). nonindustrial countries? Journal of Nervous and Mental Disorder, 167, 144—158. Implications of normal brain development Weinberger, D. (1987). Archives of General for the pathogenesis of schizophrenia. Psychiatry, jj, 660-669. Weinberger, D., Cannon—Spoor, E., Potkin, S., & Wyatt, R. (1980). Poor premorbid adjustment and CT scan abnormalities in chronic American Journal of Psychiatry, i, 1410—1413. schizophrenia. Weinberger, D., DeLisi, L., Perman, G., Targum, S., & Wyatt, R. Computed tomography in schizophreniform disorder and (1982). Archives of General Psychiatry, , other psychiatric disorders. 778—783.  212  Weinberger, D., DeLisi, L., Neophytides, A., & Wyatt, R. Familial aspects of CT scan abnormalities in chronic schizophrenic patients. Psychiatry Research, , 65-71.  (1979).  Weiss, R. (1974). The provisions of social relationships. In Z. Rubin (Ed.), Englewood Cliffs, Doing unto others, pp. 17—26. Prentice-Hall. NJ: Westermeyer, J., & Pattison, E. (1981). Social networks and mental illness in a peasant society. Schizophrenia Bulletin, 125—134.  7,  Williams, A., Reveley, M., Kolakowksa, T., Ardern, M., & Mandelbrote, B. (1985). Schizophrenia with good and poor outcome II.: Cerebral ventricle size and its clinical significance. British Journal of Psychiatry, Jj.., 239-46. The measurement Wing, J., Cooper, J., & Sartorius, N. (1974). London: Cambridge and classification of psychiatric symptoms. Unversity Press. World Health Organization (1972). The international pilot study Geneva: Author. on schizophrenia. Mental disorders: Glossary World Health Organization (1978). and guide to their classification in accordance with the ninth revision of the International Classification of Diseases. Geneva: Author. Schizophrenia: World Health Organization (1979). Wiley. New York: international follow-up study.  An  Vulnerability: A new view of Zubin, J., & Spring, B. (1977). Journal of Abnormal Psychology, j, 103-126. schizophrenia. The metamorphosis of Zubin, J., & Steinhauer, S. (1983). Pychological schizophenia: From chronicity to vulnerablity. Medicine, , 551—71.  213 Arjvendix 1. Summary of DSM-III Axis V Levels  Adult examples  Child/Adol.  examples  SUPERIOR--Unusu ally effective func tioning in social relations, occupa tional functioning, & use of leisure t i me.  Single parent living in deteriorating neighbourhood takes excellent care of children & home, has warm relations with friend, & finds time to pursue hobby.  A 12—year—old girl gets superior grades in school, is extre mely popular among peers, & excels in many sports. She does all of this with ease & comfort.  VERY GOOD--Better than average func tioning in social relations, work functioning, & use of leisure time.  A 65—yr.—old retired widower does some volunteer work, often sees old friends, & pursues hobbies.  Adolescent boy gets excellent grades, works part—time, has several close friends & plays in jazz band.  GOOD--No more than slight impairment in either social or occupational functioning.  A woman with many friends functions well at a diffi cult job, but says strain is too much.  An 8-year-old boy does well at school, has several friends, but bullies younger children.  FAIR--Moderate im pairment in either social or occupa tional functioning, or some impairment in both.  A lawyer has trou ble carrying through assignments; has sev eral acquaintances but hardly any close friends.  A 10—year—old girl does poorly in school but has adeq uate peer and family relations.  POOR--Marked im pairment in either social relations or occupational funct ioning, or moderate impairment in both.  A man with one or two friends has trouble keeping a job for more than a few weeks.  A 14-year-old boy almost fails in school & has trouble getting along with his peers.  VERY POOR--Marked impairment in both social relations & occupational funct tioning.  A woman is unable to do any of her house work & has violent outbursts toward her neighbours.  A 6—year—old girl needs special help in all subjects & has virtually no peer relationships.  GROSSLY IMPAIRED-Gross impairment in virtually all areas of functioning.  An elderly man needs supervision to main tain minimal hygiene and is usually incoherent.  A 4-yr-old boy needs constant restraint to avoid self-injury & is almost totally lacking in social skills.  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items