Clinical and non-clinical predictors of vocational recovery for Australians with psychotic disorders. (Clinical and Non-clinical Predictors).

The overall burden of disease associated with psychotic disorders is high in spite of the relatively low prevalence of these disorders in the community. In Australia, the prevalence of psychotic disorders was recently estimated at 4.7 per 1000 adult urban residents (Jablensky et al., 1999b). The burden of disease associated with two of the psychotic disorders, schizophrenia and bipolar affective disorder, is particularly high among young Australians. In young Australian males aged 15-24 years, these two disorders accounted for 10.8% of the total disease burden in 1996, representing the third leading cause of disease and injury behind road traffic accidents (13.2%), and alcohol dependence and harmful use (11.3%). Similarly, for Australian females aged 15-24 years, the combined disorders of schizophrenia and bipolar affective disorder were the second leading cause of disease burden (11.7%) behind depression, which accounted for 14.3% of the estimated total burden of disease (Mathers, Vos & Stevenson, 1999).

The high disease burden associated with psychotic disorders may be further compounded by the absence of sufficient assistance and opportunity for vocational recovery. In a recent Australian study (Waghorn, Chant & Whiteford, in press) an unemployment rate of 83.7% was found among persons with schizophrenia, a rate 10.6 times higher than for all Australians in 1997. Similar relatively high unemployment has been found among people with serious mental illness in the USA at 75-90% (Hughes, 1999), and in the UK, at 61-73% (Crowther, Marshall, Bond, & Huxley, 2001). However, unemployment is not a necessary consequence of having a serious mental illness. In a 32 year longitudinal study of people with severe mental illness, Harding, Brooks, Ashikaga, Strauss & Breier (1987) found that for one half to two thirds of participants with a retrospective DSM III diagnosis of schizophrenia, "... long-term outcome was neither downward nor marginal but an evolution into various degrees of productivity, social involvement, wellness, and competent functioning" (p. 730). Although the more detailed prognostic features of psychotic disorders are of ongoing research interest and debate, the heterogeneous pattern of the long-term disease process described by Harding et al. (1987) is widely supported (Carpenter & Strauss, 1991; Harrison, Croudace, Mason, Glazebrook & Medley, 1996; Kruger, 2000; Marengo, 1994; Mason et al., 1995).

The complex heterogeneity of process and outcomes for persons with psychotic disorders supports a continued search for clinical and demographic predictors of vocational recovery. Numerous investigations and several reviews have explored this topic (Anthony & Jansen, 1984; Crowther et al., 2001; Tsang, Lam, Ng & Leung, 2000) with results indicating that few variables confidently predict employment outcomes at the individual level. Tsang, Lam, Ng et al. (2000), reviewed controlled studies between 1985 and 1997, and found that diagnostic category and psychiatric symptoms were inconsistent predictors. Mixed results were obtained for age, sex, ethnicity, marital status, residential area, cognitive functioning, substance abuse, previous hospitalization and pre-morbid occupational performance. Pre-morbid functioning, work history and social skills were the most consistent and powerful predictors. Of the predictors reported by two or more controlled studies, only education consistently had no predictive value. Age at first hospitalization, family psychiatric history, severity of illness, and availability of employment assistance, emerged as having predictive utility in single studies.

Few studies have investigated longitudinal course of illness as a predictor of employment outcomes. Hoffman and Kupper (1996) and Kupper and Hoffman (2000) used repeated formal assessments to derive course of illness type. Although costly, these assessments were able to predict progress in vocational rehabilitation. Harrison et al. (1996) found that a less technical dichotomous classification of two-year course type (recovered versus chronic or relapsing) improved prediction of employment for persons with schizophrenia. Together these findings support the concept that differentiating course patterns of psychotic disorders may predict vocational recovery.

The generalizability of predictors across labor markets has not yet been demonstrated. Much of the research undertaken to date has taken place within the USA labor market, which is fundamentally different in many aspects from other established labor markets (Korpel, 1996; Tsang, Lam, Darasi, Ng & Chan, 2000; Waghorn & King, 1999). A clearer understanding is needed of the predictors of vocational recovery between and within labor markets to help achieve the progress goals of Anthony (1994), who sought to improve program design, encourage tailoring of programs to individual needs, and advance program evaluation by taking into account differences in user-characteristics.

In this study, psychotic disorders included schizophrenia, schizoaffective and schizophreniform disorders, bipolar disorder, depressive psychosis, mania, delusional disorders and other nonaffective psychoses. Psychotic disorders are disorders which "... have their origins in abnormal brain function and are characterized by fundamental distortions of thinking, perception, and emotional response. Clear consciousness and general intellectual capacity are usually maintained, although specific cognitive deficits may evolve in the course of time. Psychotic disorders affect the most basic functions that give a person a feeling of individuality, uniqueness, and self-direction" (Jablensky et al., 1999c, p. 2).

This investigation was prompted by the results of an examination of the predictors of vocational recovery for Australians with schizophrenia (Waghorn et al., in press). In the current study, predictor variables were selected to enable comparisons with the preceding investigation. We were interested in whether the significant predictors of vocational recovery for Australians with schizophrenia, specifically self-reported course of illness, premorbid work adjustment, age, education and skills, and diagnostic category, would predict vocational recovery for Australians with psychotic disorders. In view of the mixed evidence from North America concerning prediction by education and skills, and age, we were also interested in whether these variables would have more predictive value in the Australian labor market.

Method

The data were provided by the Data and Publications Committee of the Low Prevalence (Psychotic) Disorders Study Group. The data were collected as part of a large collaborative clinical and epidemiological investigation, one of three components of the National Survey of Mental Health and Wellbeing carried out in Australia in 1997-98 (Jablensky et al., 1999b). Data collection oCCurred over several stages in geographically representative areas of Australia. Initially, 5,710 persons who had been in contact with mental health and other related community services were screened for psychosis. From the 3,800 persons screening positive, a random sample of 980 consenting individuals completed in-depth interviews. Trained clinical staff conducted a structured clinical interview designed specifically for this study, the Diagnostic Interview for Psychosis (DIP, Jablensky et al., 1999a). The DIP incorporates elements of internationally established instruments OPCRIT (Operational Criteria for Psychoses, McGuffin et al., 1991), and the World Health Organization SCAN (Schedule for Clinical Assessment in Neuropsychiatry, Wing et al., 1990). Interview questions covered demographic variables, symptom profiles, socioeconomic descriptors, and disability items including role-related activities, service usage and unmet needs.

Data were retained from those who met the DSM III R (American Psychiatric Association, 1987) diagnostic criteria for psychosis, and who were either not inpatients or were short-term inpatients with stays of eight weeks or less at the time of interview. Those who had earlier screened positive for psychosis who did not meet DSM III R criteria following the in-depth interview were excluded from the analysis (n=15). Short-term inpatients were retained because 47.7% of the sample had at least one hospital admission in the previous year, and we did not want to arbitrarily exclude people undergoing brief rehospitalization. Persons who reported their main occupation as housework, studying, or retired (n=78), and those whose occupational status was unknown or could not be assessed (n=3), were classified as non labor force participants, approximating the Australian Bureau of Statistics (ABS) method for calculating official unemployment rates (Australian Bureau of Statistics, 1999). The final analyses (Tables 1-3) were based on 782 persons with a DSM III R diagnosed psychotic disorder who could be classified as either unemployed or employed.

Dichotomous dependent variables were formed to investigate four aspects of vocational recovery. `Current employment' represented the responses of having either a full-time job or a part-time job, whereas `unemployed' reflected having no job at present. `Durable employment' represented continuing employment for three months or more in the previous 12 months, which was compared to less than three months or no employment. This distinction is appropriate because in 1997 the Federal Government's own vocational rehabilitation provider, CRS Australia (previously known as the Commonwealth Rehabilitation Service), defined durable employment as three months or more of continuous employment. `Work performance' was assessed by retrospective self-report by those (n=247) employed in the previous year who answered the questions: In the last 12 months, have you been less efficient in your work than you would have liked to be? Have others (e.g., supervisors, workmates, members of household) criticized your performance? Those reporting no dysfunction in occupational performance were compared to those with obvious or severe dysfunction. `Absenteeism' was assessed among those employed in the previous year (n=254), measured by self-report of the number of weeks absent from work in the past 12 months. Those with at most four weeks absence were compared to those who had five or more due to mental or physical health reasons. In the absence of any supporting literature, a four-week criterion was chosen to represent double the annual allocation of two weeks sick leave with pay, the typical Australian industry benefit.

The independent variable `course of illness' was based largely on self-report. Participants were asked several general questions about how their disorder had evolved over time, whether they recovered their normal self in between episodes, whether they were bothered by constant symptoms, and whether their ability to cope with everyday life was diminished in between major episodes. They were then asked to choose one of five graphic representations (Jablensky et al., 1999a) best fitting their illness pattern since onset of the disorder. Each diagram simultaneously reflected the number of psychotic episodes (one or multiple), the extent of return to wellness between episodes, and a recovery gradient. These questions followed the majority of other clinical questions, and raters were asked to use all available information in making the final rating. In practice, raters could modify the initial self-report response if they felt another description was more accurate given other clinical information collected. Further details of this measure are available from the corresponding author on request.

Multiple logistic regression analyses were used to estimate the effects of the predictors on the four dichotomous outcomes (Ostir & Uchida, 2000). All analyses were conducted using SAS version 8, and followed recommendations for reporting multiple binary logistic regression in medical research (Khan, Chien & Dwarakanath, 1999). Each of four regression models consisted of twenty-two predictors including diagnostic category. Nineteen were selected due to similarity with predictors recently reviewed (Tsang, Lam, Ng et al., 2000). An additional predictor, lack of insight, was added to check whether interviewer ratings of insight into the causes of illness symptoms helped predict vocational recovery. In a busy vocational setting where demand often exceeds resources, persons perceived as lacking insight may have more difficulty eliciting assistance from their vocational rehabilitation professional. Current treatment status was added to the regression models to check whether short stay inpatients were less likely than non-inpatients to show vocational recovery. In the interests of parsimony we restricted the investigation of interactions to those of the second order among variables with a significant main effect. Although a few statistically significant interactions were detected, none of these were of sufficient magnitude to alter our conclusions from a main effects only model. Consequently, interaction terms were not included in any of the four regression models.

Results

Unemployment rates

The overall unemployment rate for people with psychotic disorders in this sample is 77.8%, a rate 9.8 times that for all Australians in 1997. Females aged 17-65 years with psychoses had an unemployment rate of 76.1%, whereas 78.7% of males with psychoses aged 17-65 years, were unemployed. Young people aged 17-24 years with psychoses had an unemployment rate of 80.4%, 5.7 times the rate for all Australian youth in 1997 (Australian Bureau of Statistics, 1997, 1998a, 1998b).

Odds ratios

The relative contribution of each predictor is expressed in terms of adjusted odds ratios and 95% confidence intervals in Tables 1 and 2. In these analyses the odds are a ratio of the probability of an event (e.g., employment, or absenteeism of four weeks or less) to the probability of the complementary event. Each odds ratio is given with respect to a reference level that has a denominator of 1.0. Therefore, small or large odds ratios most deviant from unity represent greater effect sizes. Confidence intervals spanning unity indicate that the corresponding odds ratio is not statistically significant at the 95% confidence level.

Predictors of current employment

Course of illness at five levels, see Appendix A and Table 1, predicted current employment. Significant odds ratios were found for all four levels of course of illness with respect to the reference level, a single episode with good or unknown recovery. Those reporting chronic illness with clear deterioration had 4.9 times greater odds of being unemployed than employed, compared to those reporting a single episode with good or unknown recovery. The other significant clinical predictor, family history of psychiatric disorder other than schizophrenia, was associated with diagnostic category through a hereditary-like association with affective psychoses (chi square p < 0.001). Whereas a family history of schizophrenia had a similar hereditary-like association with schizophrenia (chi square p = 0.03), and did not predict current employment.

Of the non-clinical variables, marital status, and education and skills, predicted current employment. Those without an Australian secondary school Year 10 Certificate or higher, had 3.8 times greater odds of unemployment than employment compared to those with vocational qualifications. Those with a secondary school certificate had a reduced risk of 2.1 times greater odds of unemployment than those with vocational qualifications.

Predictors of durable employment in the previous year

Four clinical and five non-clinical variables predicted durable employment, see Table 1. Those reporting a chronic illness with clear deterioration had 2.9 times reduced odds of achieving durable employment in the previous year than those reporting a single episode with good or unknown recovery. Although those with a family history of a psychiatric disorder other than schizophrenia had 1.6 times greater odds of durable employment, the predictive effect appeared confounded by the association between family history and diagnostic category. Diagnostic category predicted durable employment, although only one odds ratio reached significance. Compared to people with schizophrenia, those with other non-affective psychotic disorders had 2.7 times greater odds of durable employment. Having no prior lifetime diagnosis of cannabis dependence increased the relative odds of durable employment. A prior cannabis use disorder was associated with being male (chi square p < .001), of younger age (chi square p < .001), a diagnosis of schizophrenia (chi square p = .03), with earlier illness onset (chi square p < .001) and more inpatient admissions in the previous year (chi square p < .001).

Of the non-clinical variables, education and skills, age, marital status, premorbid work adjustment, and successful use of the Commonwealth Employment Service (CES) in the previous year, predicted durable employment. Each circumstance of being aged less than 45 years, partnered, with good premorbid work adjustment, and having needs met by the CES, increased the relative odds of durable employment in the past 12 months compared to those with differing circumstances.

Predictors of work performance and absenteeism

Although mid-range age of illness onset was associated with better work performance, this relationship appeared non-linear and was further complicated by a correlation with diagnostic category (chi square p = 0.007). Both the shape and direction of the relationship between age of onset and work performance depended on the DSM III R diagnosis. Age, and successful use of the CES in the previous year, also predicted previous work performance, with increasing age positively associated with work performance. This result contrasted with the difficulties older workers appeared to experience in achieving durable employment, and highlighted the potential value of older workers to employers in terms of greater sustained work performance. Those who had used the CES in the previous year and had their needs meet, had 3.5 times greater odds of poor work performance compared to those who did not receive a service from the CES, or whose needs were not met by that service. This negative relationship contrasted with the positive association between having needs met by the CES and achieving durable employment. There was a strong association between the number of hospital admissions and absenteeism in the same year. Those with no hospital admissions had 7.2 times greater odds of having used four weeks absence or less in the previous year, compared to those with two or more hospital admissions. No other variables in the model predicted absenteeism.

Discussion

The results supported our hypothesis that course of illness, pre-morbid work adjustment, age, education and skills, and diagnostic category would predict aspects of vocational recovery for Australians with psychotic disorders. In addition, lifetime diagnosis of cannabis dependency, the number of inpatient admissions in the previous 12 months, marital status, and successful use of vocational assistance, contributed to differential prediction of vocational recovery.

Psychiatric diagnosis as a predictor of durable employment

The ability of psychiatric diagnosis to predict employment outcomes remains a controversial topic in vocational rehabilitation research (Tsang, Lam, Ng et al., 2000). With a standardized diagnostic protocol, we found a specific rather than a general predictive effect for diagnostic category within the psychotic disorders, indicating that persons with schizophrenia may need more intensive assistance to retain employment than those with other non-affective psychoses.

The predictive strength of self-reported course of illness

The variable `course of illness' measured at five levels largely from self-report, see Appendix A, predicted both current and durable employment but did not predict either work performance or absenteeism. This finding indicates that the intensity of vocational assistance to both obtain and keep employment may need to increase with the severity of the longitudinal course pattern of psychotic illness. The result also provides hope for people with more severe courses of psychotic illness, because seventeen individuals who reported a chronic and deteriorating course also reported current employment.

Other clinical predictors

Although the frequency of co-morbid substance use disorders associated with schizophrenia and other psychotic disorders are considered high at 29-50% (Jablensky et al., 1999b; Kirchner et al., 1998), the impact on vocational recovery remains unclear (Tsang, Lam, Ng et al., 2000). In this study, two variables explored substance use, lifetime diagnosis of alcohol dependency, and lifetime diagnosis of cannabis dependency. A prior cannabis use disorder was negatively associated with durable employment. While encouraging to know that past lifetime problems with cannabis are unlikely to be detrimental on three of four aspects of vocational recovery, the association between substance use and durable employment remains a promising area for further investigation.

The prior evidence for the number of hospitalizations as a predictor of employment outcome has been inconclusive (Mowbray, Bybee, Harris & McCrohan, 1995; Rogers, Anthony, Toole & Brown, 1991; Rogers, Anthony, Cohen & Davies, 1997; Tsang, Lam, Ng et al., 2000). However, this study found that the number of inpatient admissions in the previous year predicted absenteeism, but no other vocational variable. People with two or more hospital admissions in the previous year may need more assistance managing absences from work than others with less frequent inpatient admissions.

The importance of the finding that age of onset predicted work performance, but was not associated with other aspects of vocational outcome, remains unclear due to an association with diagnostic category. For instance, those with early onset at ages 5-16 years and no reported dysfunction in work performance were more likely to have a diagnosis of an affective psychosis (64.3%), than schizophrenia (21.4%) or other psychoses (14.3%). Hence, more data is needed to understand the effect of age of onset in predicting work performance. In the meantime, this predictor is not sufficiently understood to support applications in vocational rehabilitation.

Predictors which appear labor-market dependent

Education and skills predicted both current and durable employment, but did not predict either work performance or absenteeism. A result contrasting with the review identifying education level as a non-predictor of employment outcomes (Tsang, Lam, Ng et al., 2000), but consistent with a more recent longitudinal study by Mueser, Salyers and Mueser (2001) where education and mother's education level predicted employment at follow-up one and two years later. The predictive strength of education and skills in this study suggests that Australians without sufficient secondary education or specific vocational skills, may benefit from the tailored provision of supported education, or supported vocational training programs.

The relationship between age and employment outcomes remains unclear due to the mixed results of prior studies (e.g., Rimmerman, Botuck, & Levy, 1995; Corrigan, Reedy, Thadani, & Ganet, 1995; Rogers et al., 1991). However, in this study, age was not associated with current employment, showed a negative association with durable employment, and was positively linked to work performance. Although older Australian workers with psychosis appeared to have more difficulty keeping a job, this is unlikely to be due to declining work performance because younger workers had greater odds of reporting work performance difficulties. In the Australian labor market, older workers aged 45-65 years, may need more assistance with retaining employment for other than work performance reasons, whereas younger workers aged 17-44 years may need specific help to maintain work performance.

Use of the CES in the past year represented access to an Australian Federal Government gateway to several forms and levels of employment assistance. However, the methods of service access, along with the type and range of employment services provided, changed when the bulk of CES functions were contracted to independent non-Government Jobnetwork providers. Although the service is no longer available, having vocational needs met by the CES was associated with achieving durable employment during the same year.

Pre-vocational psychiatric rehabilitation was unrelated to any aspect of vocational recovery. The inability of prevocational psychiatric rehabilitation to achieve employment outcomes has been a consistent finding in international vocational rehabilitation research (Bond, 1998; Bond, Dietzen, Mcgrew & Miller, 1995; Crowther et al., 2001). In Australia, the lack of any relationship between pre-vocational psychiatric rehabilitation and vocational recovery highlights the possible need for greater coordination between mental health services and the vocational sector.

Other non-clinical predictors

Marital status was classified dichotomously according to the presence or absence of a partner on the assumption that partners may facilitate obtaining and keeping employment. Marital status predicted both current and previous durable employment but not work performance or absenteeism in the previous year. The ability to remain partnered may indicate social competence, which is considered a strong predictor of vocational recovery (Tsang, Lam, Ng et al., 2000). However, the relationship is complicated by high correlations between psychopathology and social skills (Hoffman & Kupper, 1997).

Social competence was partly represented in this study by interviewer rated premorbid social adjustment, based on a series of questions about social activities prior to having psychiatric problems for the first time. In addition, premorbid functioning was represented in the predictive models by both premorbid social adjustment and premorbid work adjustment. Premorbid work adjustment rather than premorbid social adjustment predicted vocational recovery, suggesting a possible third dimension of premorbid functioning, adding to the social and academic dimensions identified by Allen et al. (2001). Poor premorbid work adjustment may indicate persons more likely to need intensive assistance to retain employment.

Non-predictors of vocational recovery

The non-predictors of all four aspects of vocational recovery are listed in Appendix A along with the significant predictors discussed. Inter-correlations among predictors are indicated by frequency differences of up to 13% across current employment status, that were not significant when adjusted for all other predictors in the logistic regression models.

Limitations of this study

As this study was limited to the available predictors in the data set it was not possible to investigate three predictors identified from the literature, namely work history, social skills, and the nature of vocational services received. Because approximately 80% of the sample were contacted through the public mental health system, it is likely that the sample represents those with more serious courses of illness than might be found among current vocational rehabilitation participants. Nevertheless, this limitation also strengthened the study, through reducing confounding by any pre-existing selection criteria for vocational assistance. Until further work has established the psychometric properties of the construct `course of illness', cautious monitoring of this variable in vocational settings is warranted. Another limitation of this study is indicated by the direction of prediction, which was retrospective with respect to durable employment, work performance and absenteeism, whereas the intended application of these results is prospective.

Conclusions

The major implication for vocational rehabilitation professionals is that in the absence of better information, valid predictors suggest subgroups of parsons with psychotic disorders in need of assistance that is more intensive, in distinct stages of vocational rehabilitation. However, even valid predictors may contribute little to vocational programs unless combined with other sources of individual-level information such as work interests, work history and career development (Ciardiello & Bingham, 1982; Griffiths, 1975), social skills assessments (Tsang & Pearson, 1996), work capacity evaluations (Zarate, Liberman, Mintz, & Massel, 1998), and on site assessments of work performance (Bryson, Bell, Greig, & Kaplan, 1999; Bryson, Bell, Lysaker, & Zito, 1997). As a guide, Table 3 summarizes the clearer findings of this study by indicating those persons more likely to need intensive assistance across four aspects of vocational recovery. Family history of psychiatric disorder other than schizophrenia, and age of onset were not included in Table 3 because the independent effects of these predictors are not sufficiently clear. Likewise, use of the CES in the previous year was omitted from Table 3 due to the discontinuation of this form of vocational assistance in Australia.

If validated by further research, self-reported course of illness may represent a low cost and practical predictor of vocational recovery with potential utility in setting initial intensity levels for vocational assistance. In addition, a shift towards a greater valuing of self-reported information by rehabilitation professionals may strengthen the rehabilitation alliance, which is likely to further enhance rehabilitation outcomes (Calsyn, Morse & Allen, 1999; Gehrs & Goering, 1994). For researchers, these results indicate several promising predictors for further investigation both within and between labor markets. Agreement among researchers is needed on the relative strength and precedence of individual predictors, along with an understanding of the mechanisms underlying each predictor, and how predictors work together as risk factors (Kraemer, Stice, Kazdin, Offord & Kupfer, 2001) for more intensive assistance needs in progressive phases of vocational rehabilitation.

Appendix A
Current employment status of Australians with psychotic disorders by
all predictors included in the regression models.

                                                    Labor force
                                                    participants
Predictors                        employed           unemployed
  Clinical                           n        %          n          %

Current treatment status
  not an inpatient                  167      23.5       544        76.5
  inpatients up to 8 weeks            7       9.9        64        90.1

Course of illness
  single episode good recovery       26      39.4        40        60.6
  multiple episodes with
    good recovery                    53      31.0       118        69.0
  multiple episodes with
    partial recovery                 53      23.1       176        76.9
  chronic illness with
    little deterioration             25      15.4       137        84.6
  chronic illness with
    clear deterioration              17      11.0       137        89.0

DSM III R Diagnostic group
  affective psychoses                87      27.3       232        72.7
  other psychoses                    25      30.1        58        69.9
  schizophrenia                      62      16.3       318        83.7

Positive thought disorder (a)
  yes                                 7      15.9        37        84.1
  no                                167      22.6       571        77.4

Negative thought disorder (b)
yes                                  13      21.7        47        78.3
no                                  161      22.3       561        77.7

Age at onset
  5-16 years                         17      17.5        80        82.5
  17-19 years                        41      25.5       120        74.5
  20-22 years                        30      19.1       127        80.9
  23-25 years                        26      22.0        92        78.0
  26-61 years                        60      24.1       189        75.9

Impairment due to
medication (c)
  moderate to severe                 37      17.1       179        82.9
  mildly                             66      24.3       205        75.7
  none, not on medication
    or unknown                       71      24.1       224        75.9

Lifetime diagnosis of alcohol
abuse/dependence (d)
  yes                                42      16.7       210        83.3
  no                                132      24.9       398        75.1

Lifetime diagnosis of cannabis
abuse/dependence (e)
  yes                                38      17.8       175        82.2
  no                                136      23.9       433        76.1

Family history of psychiatric
disorder (other than
schizophrenia)
  definite family history            96      30.0       224        70.0
  no family history                  78      16.9       384        83.1

Family history of schizophrenia
  definite family history            22      15.9       116        84.1
  no family history                 152      23.6       492        76.4

Number of inpatient admissions
in the past year
  two or more admissions             29      15.6       157        84.4
  one admission                      43      23.0       144        77.0
  nil admissions                    102      24.9       307        75.1

Lack of insight (f)
  yes                                67      16.8       333        83.2
  insight present                   107      28.8       275        71.2

Participation in psychiatric
rehabilitation in the past
year
  participated                       33      20.9       125        79.1
  no participation or
    unknown                         141      22.6       483        77.4

Non-clinical

Age
  17-24 years                        18      19.6        74        80.4
  25-34 years                        61      28.1       156        71.9
  35-44 years                        49      22.9       165        77.1
  45-65 years                        46      17.6       213        82.2
  all ages                          174      22.2       608        77.8

Sex
  M                                 105      21.3       388        78.7
  F                                  69      23.9       220        76.1

Language other than English
spoken at home
  Yes                                17      16.0        89        84.0
  No                                157      23.2       519        76.8

Education and skills (g)
  left school no                     43      11.6       329        88.4
  qualifications
  secondary school
    qualification                    34      22.1       120        77.9
  vocational qualification           62      34.3       119        65.7
  Bachelors degree or higher         35      46.7        40        53.3

Marital status
  partner                            42      35.6        76        64.4
  no partner                        132      19.9       532        80.1

Premorbid work adjustment
  poor work adjustment               38      15.4       209        84.6
  good work adjustment              136      25.4       399        74.6

Premorbid social adjustment
  poor social adjustment             45      16.1       235        83.9
  good social adjustment            129      25.7       373        74.3

Use of CES in past year (h)
  service met needs                  34      26.8        93        73.2
  needs not met, did not
  receive service or
  unknown                           140      21.4       515        78.6

Predictors                        Total
  Clinical

Current treatment status
  not an inpatient                 711
  inpatients up to 8 weeks          71

Course of illness
  single episode good recovery      66
  multiple episodes with
    good recovery                  171
  multiple episodes with
    partial recovery               229
  chronic illness with
    little deterioration           162
  chronic illness with
    clear deterioration            154

DSM III R Diagnostic group
  affective psychoses              319
  other psychoses                   83
  schizophrenia                    380

Positive thought disorder (a)
  yes                               44
  no                               738

Negative thought disorder (b)
yes                                 60
no                                 722

Age at onset
  5-16 years                        97
  17-19 years                      161
  20-22 years                      157
  23-25 years                      118
  26-61 years                      249

Impairment due to
medication (c)
  moderate to severe               216
  mildly                           271
  none, not on medication
    or unknown                     295

Lifetime diagnosis of alcohol
abuse/dependence (d)
  yes                              252
  no                               530

Lifetime diagnosis of cannabis
abuse/dependence (e)
  yes                              213
  no                               569

Family history of psychiatric
disorder (other than
schizophrenia)
  definite family history          320
  no family history                462

Family history of schizophrenia
  definite family history          320
  no family history                644

Number of inpatient admissions
in the past year
  two or more admissions           186
  one admission                    187
  nil admissions                   409

Lack of insight (f)
  yes                              400
  insight present                  382

Participation in psychiatric
rehabilitation in the past
year
  participated                     158
  no participation or
    unknown                        624

Non-clinical

Age
  17-24 years                       92
  25-34 years                      217
  35-44 years                      214
  45-65 years                      259
  all ages                         782

Sex
  M                                493
  F                                289

Language other than English
spoken at home
  Yes                              106
  No                               676

Education and skills (g)
  left school no                   372
  qualifications
  secondary school
    qualification                  154
  vocational qualification         181
  Bachelors degree or higher        75

Marital status
  partner                          118
  no partner                       664

Premorbid work adjustment
  poor work adjustment             247
  good work adjustment             535

Premorbid social adjustment
  poor social adjustment           280
  good social adjustment           502

Use of CES in past year (h)
  service met needs                127
  needs not met, did not
  receive service or
  unknown                          655

Notes

(a) Interviewer rated speech problems characterized by a lack of logical
connections, unexpected shifts from topic to topic, answers past the
point, bizarre use of words and phrases; or words with no generally
accepted meanings.

(b) Interviewer rated speech problems characterized by blocking, poverty
of content, and restricted quantity of speech.

(c) Self-reported impairment of daily life attributed to side effects of
medication.

(d) Self-reported recurrent or continued use of alcohol for at least one
month despite knowledge of previous problems
during lifetime caused or exacerbated by alcohol.

(e) Self-reported recurrent or continued use of cannabis for at least
one month despite knowledge of previous problems during lifetime caused
or exacerbated by cannabis.

(f) Interviewer ratings of participants' causal explanations for
psychotic symptoms and participants' general understanding
of their psychotic illness.

(g) Vocational qualifications included trade certificates, undergraduate
diplomas, other vocational certificates and nursing or
teaching qualifications not requiring a university degree.

(h) In 1997 the Commonwealth Employment Service (CES) provided access to
job vacancies and sponsored participation in a range of labor market
programs, sometimes referring people to other training or rehabilitation
services. Unfortunately, no data were available on the actual CES
services received.

Table 1

Predictors of current and durable (a) employment for Australians with
psychotic disorders. Odds-ratios and 95% confidence intervals adjusted
for all predictors.

Predictors               Current employment      Durable employment
            Clinical
Course of illness
  single episode,
    good or unknown
    recovery (ref.
    level)             1.0                      1.0
  multiple episodes,
    good recovery      0.419 (0.200, 0.876) *   0.891 (0.442, 1.797)
  multiple episodes,
    partial recovery   0.339 (0.164, 0.700) *   0.514 (0.257, 1.026)
  chronic illness
    with little
    deterioration      0.235 (0.103, 0.532) *   0.383 (0.175, 0.836) *
  chronic illness
    with clear
    deterioration      0.202 (0.085, 0.482) *   0.348 (0.156, 0.779) *
Family history of
psychiatric disorder
(other than
schizophrenia)
  no family history
    (ref. level)       1.0                      1.0
  definite family
  history              2.363 (1.569, 3.560) *   1.649 (1.121, 2.427) *
Lifetime diagnosis
of cannabis abuse/
dependence
  yes (ref. level)     1.0                      1.0
  no                   1.187 (0.702, 2.005)     1.711 (1.056, 2.773) *
DSM III R Diagnostic
Group
  schizophrenia
    (ref. level)       1.0                      1.0
  affective
    psychoses (b)      1.166 (0.737, 1.845)     1.313 (0.856, 2.013)
  other psychoses
    (c)                1.749 (0.935, 3.273)     2.728 (1.524, 4.883) *

        Non-clinical
Education and skills
  left school no
    qualifications     0.263 (0.158, 0.437) *   0.370 (0.232, 0.590) *
  secondary school
    qualification      0.483 (0.277, 0.844) *   0.483 (0.281, 0.829) *
  vocational
    qualification
    (ref. level)       1.0                      1.0
  bachelors degree
    or higher          1.662 (0.885, 3.119)     1.854 (0.988, 3.480)
Age
  17-24 years          1.466 (0.637, 3.375)     2.620 (1.227, 5.594) *
  25-34 years          1.866 (1.044, 3.335)     3.544 (2.014, 6.238) *
  35-44 years          1.486 (0.877, 2.516)     2.194 (1.318, 3.652) *
  45-65 years (ref.
    level)             1.0                      1.0
Marital status
  partner (ref.
    level)             1.0                      1.0
  no partner           0.533 (0.321, 0.885) *   0.542 (0.331, 0.889) *
Premorbid work
adjustment
  poor work
    adjustment (ref.
    level)             1.0                      1.0
  good work
    adjustment         1.341 (0.842, 2.136)     2.160 (1.380, 3.383) *
Use of CES in past
year
  service met needs    1.164 (0.691, 1.962)     1.687 (1.052, 2.705) *
  needs not met or
    did not receive
    service or
    unknown (ref.
    level)             1.0                      1.0

Notes * Indicates both a significant main effect and a significant
confidence interval at the [alpha] = .05 level.

(a) Durable employment was defined as three months or more employment
in the previous year

(b) Affective psychoses included definite and probable diagnoses of
psychotic depression, mania, bipolar affective disorder,
bipolar with psychosis and schizoaffective disorder.

(c) Other psychoses included atypical psychosis, delusional disorder
and schizophreniform psychosis.

Table 2
Predictors of self-reported work performance a and absenteeism (b) in
the past year for Australians with psychotic disorders. Odds-ratios
and 95% confidence intervals adjusted for all predictors.

Predictors            No dysfunction in         At most four weeks
                      work performance in       absenteeism in the
                      the past year             past year

      Clinical
Number of inpatient
  admissions in the
  past year two or
  more admissions
  (ref. level)        1.0                       1.0
  one admission       2.108 (0.843, 5.270)      1.769 (0.752, 4.164)
  nil admissions      2.462 (1.011, 5.996)      7.248 (3.039, 17.289) *
Age at onset
  5-16 years          2.112 (0.584, 7.635)      1.064 (0.299, 3.788)
  17-19 years         3.869 (1.390, 10.770) *   1.782 (0.654, 4.856)
  20-22 years         3.320 (1.174, 9.388) *    1.790 (0.645, 4.973)
  23-25 years         1.138 (0.401, 3.227)      0.859 (0.312, 2.363)
  26-61 years
    (ref. level)      1.0                       1.0

      Non-clinical
Age
  17-24 years         0.126 (0.031, 0.514) *    0.335 (0.082, 1.372)
  25-34 years         0.166 (0.057, 0.480) *    0.385 (0.136, 1.086)
  35-44 years         0.195 (0.068, 0.560) *    0.463 (0.170, 1.258)
  45-65 years
    (ref. level)      1.0                       1.0
Use of CES in
  past year service
  met needs           0.284 (0.130, 0.620) *    0.782 (0.362, 1.686)
  needs not met or
    did not receive
    service or
    unknown (ref.
    level)            1.0                       1.0

Notes * Indicates both a significant main effect and a significant
confidence interval at the [alpha] = .05 level.

(a) Work performance was assessed for those who had held a job in the
previous 12 months.

(b) Absenteeism was defined as five or more weeks absence from work in
the previous 12 months for mental or physical health reasons.

Table 3
Estimated vocational assistance needs of Australians with psychotic
disorders.

Predictor                 Aspect of vocational recovery

                 Obtaining               Good work
                 employment              performance
Course of
illness          Those with more
                 severe courses of
                 illness may need more
                 assistance to prepare
                 for and commence
                 employment

DSM III R
diagnostic
group

Lifetime
diagnosis of
cannabis abuse

Number of                                Those with two or
inpatient                                more admissions may
admissions                               need more assistance
in the past                              with work
year                                     performance

Age                                      Workers aged 44 years
                                         or less may need more
                                         assistance with work
                                         performance

Education and    Those with lower
skills           education and skills
                 may need more
                 assistance to prepare
                 for and commence
                 employment

Marital status   Those without
                 partners may need
                 more assistance to
                 prepare for and
                 commence
                 employment

Premorbid
work
adjustment

Predictor                 Aspect of vocational recovery

                    Low                   Achieving durable employment
                 absenteeism
Course of
illness                                   Those with more severe
                                          courses of illness may need
                                          more intensive post-placement
                                          assistance

DSM III R                                 Those with schizophrenia may
diagnostic                                need more intensive
group                                     post-placement assistance

Lifetime                                  Those who have had a previous
diagnosis of                              problem with cannabis may
cannabis abuse                            need more intensive
                                          post-placement assistance

Number of        Those with two or
inpatient        more admissions may
admissions       need more assistance
in the past      to prevent absenteeism
year

Age                                       Workers aged 45 years or more
                                          may need more intensive
                                          post-placement assistance
                                          for reasons other than
                                          work- performance

Education and                             Those with lower education
skills                                    and skills may need more
                                          intensive post-placement
                                          assistance

Marital status                            Those without partners may
                                          need more intensive
                                          post-placement assistance

Premorbid                                 Those with poor premorbid
work                                      work adjustment may need more
adjustment                                intensive post-placement
                                          assistance

Acknowledgements

We thank Professor John McGrath for comments on early drafts of this paper and for supervising access to Confidentialised Unit Record File data. This investigation was conducted with ethics approval from The University of Western Australia Human Research Ethics Committee. This paper is based on data collected in the framework of the collaborative Low Prevalence (Psychotic) Disorders Study, an epidemiological and clinical investigation forming part of the National Mental Health and Wellbeing Survey, Australia 1997-98. A complete list of investigators, funding arrangements, detailed acknowledgements, and a full description of the survey is available elsewhere (Jablensky et al., 1999b).

References

Allen, D. N., Kelley, M. E., Miyatake, R. K., Gurkliss, J. A., & van Kammen, D. P. (2001). Confirmation of a two factor model of premorbid adjustment in males with schizophrenia. Schizophrenia Bulletin, 27(1), 39-46.

Anthony, W. A. (1994). Characteristics of people with psychiatric disabilities that are predictive of entry into the rehabilitation process and successful employment. Psychosocial Rehabilitation Journal, 17(3), 3-13.

Anthony, W. A., & Jansen, M. A. (1984). Predicting the vocational capacity of the chronically mentally ill. American Psychologist, 5,537-544.

American Psychiatric Association, (1987). Diagnostic and Statistical Manual of Mental Disorders, Third Edition, Revised. Washington DC: American Psychiatric Association.

Australian Bureau of Statistics, (1997). Austats: Time Series Spread Sheet no. 6291.0.40.001, July, (Electronic Data Base). Canberra: Commonwealth Government.

Australian Bureau of Statistics, (1998a). Australian Social Trends (4102.0). Canberra: Commonwealth Government.

Australian Bureau of Statistics, (1998b). Transition from Education to Work (6227.0). Canberra: Commonwealth Government.

Australian Bureau of Statistics, (1999). ABS Standards for Social, Labor and Demographic Variables. Canberra: Commonwealth Government.

Bond, G. (1998). Principles of the Individual Placement and Support model: empirical support. Psychiatric Rehabilitation Journal, 22(1), 11-23.

Bond, G. R., Dietzen, L. L., McGrew, J. H., & Miller, L. D. (1995). Accelerated entry into supported employment for persons with severe psychiatric disabilities. Rehabilitation Psychology. 40(2), 75-94.

Bryson, G., Bell, M. D., Greig, T., & Kaplan, E. (1999). The Work Behaviour Inventory: prediction of future work success of people with schizophrenia. Psychiatric Rehabilitation Journal, 23(2), 113-117.

Bryson, G., Bell, M. D., Lysaker, P., & Zito, W. (1997). The work behavior inventory: a scale for the assessment of work behaviour for people with mental illness. Psychiatric Rehabilitation Journal, 20(4), 47-55.

Calsyn, R. J., Morse, G. A., & Allen, G. (1999). Predicting the helping alliance with people with a psychiatric disability. Psychiatric Rehabilitation Journal, 22(3), 283-287.

Carpenter, W. T., & Strauss, J. S. (1991). The prediction of outcome in schizophrenia IV: eleven-year follow-up of the Washington IPSS Cohort. The Journal of Nervous and Mental Disease, 179(9), 517-525.

Ciardiello, J. A., & Bingham, W. C. (1982). The career maturity of schizophrenic clients. Rehabilitation Counseling Bulletin, September, 3-9.

Corrigan, P., Reedy, P., Thadani, D., & Ganet, M. (1995). Correlates of participation and completion in a job club for clients with psychiatric disability. Rehabilitation Counseling Bulletin, 39(1), 43-53.

Crowther, R. E., Marshall, M., Bond, G. R., & Huxley, P. (2001). Helping people with severe mental illness to obtain work: systematic review. British Medical Journal, 322, 204-208.

Gehrs, M., & Goering, P. (1994). The relationship between the working alliance and rehabilitation outcomes of schizophrenia. Psychosocial Rehabilitation Journal, 18(2), 43-54.

Griffiths, R. D. P. (1975). Vocational guidance conducted with psychiatric patients. Bulletin of the British Psychological Society, 28, 427-436.

Harding, C. M., Brooks, G. W., Ashikaga, T., Strauss, J. S., & Breier, A. (1987). The Vermont longitudinal study of persons with severe mental illness, II: long term outcome of subjects who retrospectively met DSM-III criteria for schizophrenia. American Journal of Psychiatry, 144, 727-735.

Harrison, G., Croudace, T., Mason, P., Glazebrook, C., & Medley, I. (1996). Predicting the long-term outcome of schizophrenia. Psychological Medicine, 26(4), 697-705.

Hoffman, H., & Kupper, Z. (1996). Patient dynamics in early stages of vocational rehabilitation: a pilot study. Comprehensive Psychiatry, 37(3), 216-221.

Hoffman, H., & Kupper, Z. (1997). Relationships between social competence, psychopathology, and work performance and their predictive value for vocational rehabilitation of schizophrenic outpatients. Schizophrenia Research, 23, 69-79.

Hughes, R. (1999). Psychosocial rehabilitation, new protocols, ethics and outcomes. International Journal of Mental Health, 28(1), 3-33.

Jablensky, A., McGrath, J., Herrman, H., Castle, D., Gureje, O., Morgan, V., & Korten, A. (1999a). Diagnostic Interview for Psychoses (Interview Schedule Version 1.0, National Survey of Mental Health and Wellbeing Report 4: People living with psychotic illness: An Australian Study 1997-98). Canberra: Commonwealth Department of Health and Aged Care.

Jablensky, A., McGrath, J., Herrman, H., Castle, D., Gureje, O., Morgan, V., & Korten, A. (1999b). National Survey of Mental Health and Wellbeing. Report 4. People Living with Psychotic Illness: An Australian Study 1997-98. Canberra: Commonwealth Department of Health and Aged Care.

Jablensky, A., McGrath, J., Herrman, H., Castle, D., Gureje, O., Morgan, V., & Korten, A. (1999c). National Survey of Mental Health and Wellbeing. Bulletin 1, People Living with Psychotic Illness: An Australian Study 1997-98. An Overview. Canberra: Commonwealth Department of Health and Aged Care.

Khan, K.S., Chien, P.F.W., & Dwarakanath, L.S. (1999). Logistic regression models in Obstetrics and Gynecology literature. Obstetrics & Gynecology, 93(6), 1014-1020.

Kirchner, J. E., Owen, R. R., Nordquist, C., & Fischer, E. P. (1998). Diagnosis and management of substance use disorders among inpatients with schizophrenia. Psychiatric Services, 49(1), 82-85.

Korpel, M. H. T. (1996). Supported employment behind the Netherland's dikes: Trends in vocational rehabilitation for persons with psychiatric disabilities. Journal of Vocational Rehabilitation, 6, 97-105.

Kraemer, H. C., Stice, E., Kazdin, A., Offord, D., & Kupfer, D. (2001). How do risk factors work together? Mediators, moderators, and independent overlapping and proxy risk factors. American Journal of Psychiatry, 158(6), 848-856.

Kruger, A. (2000). Schizophrenia: recovery and hope. Psychiatric Rehabilitation Journal, 24(1), 29-37.

Kupper, Z., & Hoffman, H. (2000). Course patterns of psychosocial functioning in schizophrenia patients attending a vocational rehabilitation program. Schizophrenia Bulletin, 26(3), 681-698.

Marengo, J. (1994). Classifying the courses of schizophrenia. Schizophrenia Bulletin, 20(3), 519-536.

Mason, P., Harrison, G., Glazebrook, C., Medley, I., Dalkin, T., & Croudace, T. (1995). Characteristics of outcome in schizophrenia at 13 years. British Journal of Psychiatry, 167,596-603.

Mathers, C., Vos, T., & Stevenson, C. (1999). The Burden of Disease and Injury in Australia. Canberra: Australian Institute of Health and Welfare.

McGuffin, P., Farmer, A. E., & Harvey, I. (1991). A polydiagnostic application of operational criteria in studies of psychotic illness. Development and reliability of the OPCRIT system. Archives of General Psychiatry, 48, 764-770.

Mowbray, C. T., Bybee, D., Harris, S. N., & McCrohan, N. (1995). Predictors of work status and future work orientation in people with a psychiatric disability. Psychiatric Rehabilitation Journal, 19(2), 17-28.

Mueser, K.T., Salyers, M.P. & Mueser, P.R. (2001). A prospective analysis of work in schizophrenia. Schizophrenia Bulletin, 27(2), 281-296.

Ostir, G. V., & Uchida, T. (2000). Logistic regression a non-technical review. American Journal of Physical Medicine and Rehabilitation, 79(6), 565-572.

Rimmerman, A., Botuck, S., & Levy, J. M. (1995). Job placement for individuals with psychiatric disabilities in supported employment. Psychiatric Rehabilitation Journal, 19(2), 37-43.

Rogers, E. S., Anthony, W. A., Toole, J., & Brown, M. A. (1991). Vocational outcomes following psychosocial rehabilitation. A longitudinal study of three programs. Journal of Vocational Rehabilitation, 1 (3), 21-29.

Rogers, E.S., Anthony, W.A., Cohen, M., & Davies, R.R. (1997). Prediction of vocational outcome based on clinical and demographic predictors among vocationally ready clients. Community Mental Health Journal, 33(2), 99-112.

Tsang, H., Lam, P., Darasi, B., Ng, B., & Chan, F. (2000). Predictors of post-hospital employment status for persons with psychiatric disability: perceptions of rehabilitation health professionals in Hong Kong. Psychiatric Rehabilitation Journal, 24(2), 169-173.

Tsang, H., Lam, P., Ng, B., & Leung, O. (2000). Predictors of employment outcome for people with psychiatric disabilities: a review of the literature since the mid `80s. Journal of Rehabilitation, 66(2), 19-31.

Tsang, W. H. H., & Pearson, V. (1996). A conceptual framework for work-related social skills in psychiatric rehabilitation. Journal of Rehabilitation, 62(3), 61-67.

Waghorn, G., Chant, D., & Whiteford, H. (in press). The strength of self-reported course of illness in predicting vocational recovery for Australians with schizophrenia. Journal of Vocational Rehabilitation.

Waghorn, G., & King, R. (1999). Australian trends in vocational rehabilitation for psychiatric disability. Journal of Vocational Rehabilitation, 13(3), 153-163.

Wing, J. K., Babor, T., Brugha, T., Burke, J., Cooper, J. E., Giel, R., Jablenski, A., Regier, D., & Sartorius, N. (1990). SCAN: schedules for clinical assessment in neuropsychiatry. Archives of General Psychiatry, 47(June), 589-593.

Zarate, R., Liberman, R. P., Mintz, J., & Massel, H. K. (1998). Validation of a Work Capacity Evaluation for individuals with psychiatric disorders. Journal of Rehabilitation, 64(4), 28-33.

Geoff Waghorn
The Park, Centre for Mental Health

David Chant
Harvey Whiteford

The University of Queensland

Geoff Waghorn, The Park, Centre for Mental Health, Wacol, Queensland, 4076, Australia. E-mail: [e-mail]