- Email: [email protected]
Socio-demographic and clinical predictors of occupational status in schizophrenic psychoses—follow-up within the Northern Finland 1966 Birth Cohort
Psychiatry Research 150 (2007) 217 – 225 www.elsevier.com/locate/psychres
Socio-demographic and clinical predictors of occupational status in schizophrenic psychoses—follow-up within the Northern Finland 1966 Birth Cohort Jouko Miettunen a,⁎, Erika Lauronen a , Juha Veijola a,b , Hannu Koponen b,c , Outi Saarento d , Anja Taanila e , Matti Isohanni a a
Department of Psychiatry, University of Oulu, P.O.Box 5000, 90014 Oulun Yliopisto, Finland b Academy of Finland, Helsinki, Finland c Department of Psychiatry, University of Kuopio, Kuopio, Finland d Department of Psychiatry, Oulu University Hospital, Oulu, Finland e Department of Public Health and General Practice, University of Oulu, Oulu, Finland Received 3 March 2006; received in revised form 16 June 2006; accepted 30 August 2006
Abstract We studied occupational status of persons with schizophrenic psychoses by age 34 in a longitudinal population-based cohort and predicted which demographic and illness-related factors could support the patients to maintain their occupational capacity. Subjects of the Northern Finland 1966 Birth Cohort with the diagnosis of DSM-III-R schizophrenic psychoses (n = 113) by the year 1997 were followed until the end of year 2000. Various illness and socio-demographic factors at the time of onset of illness were used as predictors. At the end of the follow-up time 50 (44%) of patients were not pensioned and 22 (20%) were also working at least half of the time during year 2000. After adjusting for gender, being unemployed at onset, educational level and proportion of time spent in psychiatric hospitals, those who were married or cohabiting at the time of onset of illness were less often on pension than those who were single (OR 6.51; 95% CI 1.83–23.12). Thus, nearly half of the patients with schizophrenic psychoses were not pensioned after an average 10 years follow-up. Based on our findings, those who were single at time of their onset of illness probably need most support to retain their contacts to work life. © 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Employment; Pension; Psychotic disorders; Schizophrenia; Functioning; Marital status
1. Introduction Schizophrenia is a severe, often life-long disease, which may cause severe functional decline and unemployment in the patients. Maintaining working ability is
⁎ Corresponding author. Tel.: +358 8 3156923; fax: +358 8 333167. E-mail address: [email protected] (J. Miettunen).
an important goal during the course of schizophrenia (Bell and Lysaker, 1997; Priebe et al., 1998). Many studies have reported low rates of employment in schizophrenia patient. A recent study by Marwaha and Johnson (2004) reviewed 22 studies with patients having schizophrenia. They reported wide variation in employment rates (between 4% and 90%). Most of the high rates were from the International Pilot Study of Schizophrenia (IPSS), where study samples excluded
0165-1781/$ - see front matter © 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.psychres.2006.08.011
218
J. Miettunen et al. / Psychiatry Research 150 (2007) 217–225
people with chronic schizophrenia with poorer outcome. In the same review, the employment rates were between 13% and 65% in first-episode psychosis studies. There have been some studies which have found predictors for occupational outcome (employment or disability pension) in schizophrenia subjects. Chronicity of illness (Tsang et al., 2000; Mueser et al., 2001) and male gender (Harrison et al., 1996; Agerbo et al., 2004) has been associated with poor occupational functioning in several studies. Also prior work experience, patient's and mother's higher educational level, and good social functioning have predicted work ability in people with schizophrenia (Mueser et al., 2001). In this article, we explore occupational status in schizophrenic psychoses within an early middle-aged, general population-based sample of the Northern Finland 1966 Birth Cohort. By using various demographic and illness-related variables, we also predict which of them could support the patients in maintaining their employment. We use register and hospital notes-based data, which minimizes the typical problem of drop-outs seen in most of the previous studies. Our special focus is on the variables concerning the time of the early phase of psychosis and first psychiatric hospitalization. 2. Methods 2.1. Data collection The Northern Finland 1966 Birth Cohort is an unselected, general population birth cohort ascertained during mid-pregnancy. It is based upon 12,058 live-born children in the provinces of Lapland and Oulu (Rantakallio, 1969). Permission to gather data was obtained from the Ministry of Social and Health Affairs and the study design is under review by the Ethical Committee of the Northern Ostrobothnia Hospital District. Altogether, 83 subjects have forbidden the use of their data and have been excluded. The current data include all subjects living in Finland at the age 16 years and still alive at the end of the year 2000 (n = 10,748). 2.2. Subjects of the study The nationwide Finnish Hospital Discharge Register (FHDR) covers all mental and general hospitals as well as beds in local health centers and private hospitals nationwide. All cohort members over 16 years appearing on the FHDR until the end of 1997 for any mental disorder were identified and their diagnoses were rechecked twice by professionals using DSM-III-R criteria (APA, 1987). The reliability for schizophrenia diagnoses
of this procedure was good (kappa = 0.85). A more detailed description of the validation process is presented elsewhere (Isohanni et al., 1997; Moilanen et al., 2003). By the end of 1997, there were 153 subjects (90 men, 59%) in the FHDR with a known psychotic episode in their life. Of these cases, 128 (76 men, 59%) were given the diagnosis of schizophrenia spectrum in the validation process. Eleven (9 men, 82%) of these were deceased by the end of the year 2000 and four subjects were pensioned before the onset of psychosis: one female due to unspecified delay in development (ICD-9 diagnosis: 31599); one female due to mild mental retardation (ICD-9: 317); and one male and one female due to epilepsy (ICD-9: 345). These subjects were excluded from this study. The final sample included all living subjects with schizophrenic psychoses who have not been pensioned due to developmental or neurological disorder (n = 113 cases; 67 men, 59%). The sample included 85 schizophrenia cases and 28 other schizophrenia spectrum cases. 2.3. Assessment of occupational status We collected information on all disability pensions from the Social Insurance Institution of Finland. In Finland, disability pension is granted if the disorder is persisting and incapacitating for an indefinite or timelimited period, usually for 1–2 years. Subjects on disability pension are allowed to have small earnings. We also collected information concerning all work periods contributing to pension from the Central Pension Security Institute. All these data sets were available until the end of year 2000. We summarized positive occupational status to two dichotomized variables representing positive occupational outcome: Not on vs. on pension at the end of year 2000; and not on pension at the end of 2000 and working 50% of all days during the year 2000 vs. on pension or working less than 50% during the year 2000. 2.4. Predictors of occupational status We collected information on predictors from registers and hospital notes. Various socio-demographic factors as well as illness-related variables were analyzed: 2.4.1. Gender Male vs. female. Gender was also used as a covariate. 2.4.2. Pre-morbid factors Information on pre-morbid factors was collected with a retrospective review of hospital notes using the
J. Miettunen et al. / Psychiatry Research 150 (2007) 217–225
Operational Criteria Checklist for Psychotic Illness (OCCPI) (McGuffin et al., 1991). The variables referred to the time before or around onset of illness. Poor work adjustment (no/yes) around the illness onset was scored, e.g., if the patient was unable to keep any job for more than 6 months, had a history of frequent changes of job, or was only able to sustain a job well below that expected by his/her educational level. Poor pre-morbid social adjustment (no/yes) refers, for example, to difficulties in maintaining normal social relationships or to persistent social isolation. Pre-morbid personality disorder (no/yes) refers to any evidence of personality disorder present prior to the onset of psychosis. Alcohol abuse within 1 year of onset of psychotic symptoms (no/yes) was rated if quantity was excessive (rater judgement) and alcohol-related complications had occurred. Definite psychosocial stressor (no/yes) was scored if a severely threatening event had occurred prior to onset of disorder that was unlikely to have resulted from the subject's own behavior. Also, the following variables were used: mode of onset (acute or gradual/ insidious), marital status (married/single), and unemployed at illness onset (no/yes). OCCPI data were available in some form for 109 (96.5%) subjects of the sample, although some more data were missing in the single OCCPI items. The reasons for missing data were lack of hospital notes or insufficient information given in the notes. 2.4.3. Familial risk for psychosis First-degree relative having vs. not having psychosis. Assessed by hospital notes and by using data on mothers' and fathers' psychotic hospitalizations (FHDR) between 1972 and 2000. 2.4.4. Length of first psychiatric hospitalization The length of first hospitalization was used as a dichotomized variable (1–14 days, 15 days or more). This categorization is comparable with some earlier studies (Appleby et al., 1993; Saarento et al., 1998; Øiesvold et al., 2000); we have used this categorization in a previous study as well (Miettunen et al., 2006). 2.4.5. Proportion of time spent in psychiatric hospitals after onset of illness Information on psychiatric hospitalization was collected from the FHDR until the end of 2000 (Miettunen et al., 2006). The proportion of time spent in psychiatric hospitals after onset of illness was used as a covariate, as this variable is a proxy of the chronicity of illness and also takes into account different lengths of follow-up times of the subjects. Time in hospital is taken into
219
account also because it may hinder maintaining the contact with the world of work. As the variable was extremely skewed, we transformed the original variable taking a natural logarithm. 2.4.6. Educational level The information on education of the subjects was collected in 1997 from the National Educational Registry of Statistics, Finland. The variable includes three categories according to the length of education: primary level (less than 10 years), secondary level (10–12 years) and tertiary level (more than 12 years) education. 2.5. Comparison to general population We also report previously unpublished data on pension rates and work periods in the total sample of the Northern Finland 1966 Birth Cohort Study living in Finland 1997 and being alive at the end of follow-up (n = 10,748; 5452 males and 5296 females). This is done in order to make the presented pension and employment rates in schizophrenic psychoses comparable to the general population. In the total sample, we were able to study the effect of gender and education level on occupational status. 2.6. Statistical analyses Differences between outcome groups were analyzed with chi-square tests and odds ratios (dichotomized variables) and trend tests (linear-by-linear association statistic, ordinal variables). All tests were two-sided. Survival analysis with Log Rank test statistics was used to study the time to being pensioned when comparing patients by their marital status at the time of onset of illness. Logistic and Cox regression analyses were used to adjust for confounding. In these multivariate analyses gender, being unemployed at onset, educational level, and the proportion of time spent in psychiatric hospitals after onset of illness (natural logarithm) were used as covariates. Statistical analyses were made with SPSS 12.0 for Windows (SPSS Inc., 2003). 3. Results In the total Northern Finland 1966 Birth Cohort 2.3% were pensioned at the end of 2000 (247/10,736). Males (2.4%; 131/5442) were more commonly pensioned than females (2.2%; 116/5294). In the total sample, 69.0% (7421/10,748) were working at least 50% of days and were not pensioned during the year 2000. Males (77.6%; 4233/5452) were more commonly working than females
220
J. Miettunen et al. / Psychiatry Research 150 (2007) 217–225
Table 1 Socio-demographic and symptom variables of living subjects with schizophrenic psychoses by different occupational variables within the Northern Finland 1966 Birth Cohort Predictor variables (n = males/ females)
Pre-morbid factors
Male (n = 67)
Female (n = 46)
Not on any pension at the end of year 2000 a
At work b and not on any pension year 2000 a
Not on any pension at the end of year 2000 a
At work b and not on any pension year 2000 a
n
n
n
n
%
%
%
%
Unemployed at onset (n = 63/40) No 19/37 51.4 13/37 35.1 12/28 42.9 2/28 7.1 Yes 6/26 23.1 2/26 7.7 8/12 66.7 2/12 16.7 Poor work adjustment (n = 57/41) No 20/38 52.6 13/38 34.2 13/30 43.3 1/30 3.3 Yes 4/19 21.1 1/19 5.3 7/11 63.6 3/11 27.3 Mode of onset (n = 58/42) Acute 8/13 61.5 Gradual 9/20 45.0 Insidious 7/25 28.0 Marital status (n = 53/38) Married or 8/12 66.7 cohabiting Single 9/41 22.0
5/13 38.5 5/20 25.0 4/25 16.0
5/7 71.4 1/7 14.3 7/13 53.8 3/13 23.1 8/22 36.4 0/22 0.0
4/12 33.3 16/21 76.2 4/21 19.0 4/41
9.8
3/17 17.6 1/17
5.9
Poor pre-morbid social adjustment (n = 37/30) No 9/14 64.3 4/14 28.6 11/20 55.0 2/20 10.0 Yes 5/23 21.7 3/23 13.0 6/10 60.0 2/10 20.0 Pre-morbid personality disorder (n = 35/25) No 6/24 25.0 4/24 16.7 7/19 36.8 1/19 5.3 Yes 5/11 45.5 4/11 36.4 3/6 50.0 1/6 16.7 Alcohol abuse within 1 year of onset (n = 62/43) No 18/52 34.6 12/52 23.1 20/41 48.8 4/41 9.8 Yes 6/10 60.0 3/10 30.0 1/2 50.0 1/2 50.0 Psychosocial stressor prior to onset (n = 63/45) No 22/58 37.9 15/58 25.9 18/40 45.0 5/40 12.5 Yes 2/5 40.0 0/5 0.0 4/5 80.0 0/5 0.0 Familial risk for psychosis c (n = 67/46) No 22/54 40.7 14/54 25.9 15/34 44.1 4/34 11.8 Yes 6/13 46.2 3/13 23.1 7/12 58.3 1/12 8.3 Onset age before 21 years (n = 67/46) No 24/44 54.5 14/44 31.8 16/27 59.3 3/27 11.1 Yes 4/23 17.4 3/23 13.0 6/19 31.6 2/19 10.5 Length of first hospitalization (n = 66/46) 1–14 days 13/24 54.2 7/24 29.2 6/10 60.0 2/10 20.0 15 days or 14/42 33.3 9/42 21.4 16/36 44.4 3/36 8.3 more
Table 1 (continued) Predictor variables (n = males/ females)
Pre-morbid factors
Male (n = 67)
Female (n = 46)
Not on any pension at the end of year 2000 a
At work b and not on any pension year 2000 a
Not on any pension at the end of year 2000 a
At work b and not on any pension year 2000 a
n
n
n
n
%
%
%
%
Educational level at year 1997 (n = 67/46) Basic 3/14 21.4 2/14 14.3 1/7 14.3 0/7 0.0 Secondary 23/49 46.9 15/49 30.6 17/34 50.0 4/34 11.8 Tertiary 2/4 50.0 0/4 0.0 4/5 80.0 1/5 20.0 Statistically significant (P b 0.05; chi-square test or Fisher exact test as appropriate, or trend test if more than two categories) differences are in bold. a Information from the Social Insurance Institution of Finland. b At least 6 months during the year 2000, information from the Finnish Centre for Pensions, includes all work periods contributing to pension. c Collected from hospital notes for all first-degree relatives and additionally for parents from the Finnish Hospital Discharge Register until the end of 2000.
(60.2%; 3188/5296). Cohort subjects with tertiary education (72.8%; 1978/2718) and secondary education (71.4%; 4630/6483) were more commonly at work than subjects with basic education (52.5%; 806/1535). These differences in gender and education were statistically significant (chi-square tests, P b 0.001). In patients with schizophrenic psychoses, the mean (standard deviation) follow-up time of the sample was 10.6 (S.D. = 4.2) years. Sixty-three (55.8%) patients were pensioned at the end of the follow-up time. The average time from the first psychiatric admission to pension was 3.3 (S.D. = 3.1) years. Fifty (44.2%) patients were not pensioned at the end of the follow-up time and 22 (19.5%) also worked for at least 6 months during the year 2000. The corresponding figures among males were 28 (41.8%) and 17 (25.4%); and among females 22 (47.8%) and 5 (10.9%), respectively. The socio-demographic and clinical characteristics of the subjects with schizophrenic psychoses by gender are presented in Table 1 by the two occupational status variables. Among men being employed and married or cohabiting at the time of onset, good pre-morbid work and social adjustment, and earlier age at onset were associated (P b 0.05) with better occupational outcome at least in one of the outcome variables. Among women being single associated with being pensioned, and higher education, and prior good work adjustment were
J. Miettunen et al. / Psychiatry Research 150 (2007) 217–225
221
Table 2 Predicting working ability in subjects with schizophrenic psychoses within the Northern Finland 1966 Birth Cohort Not on any pension at the end of year 2000 a
Female gender Employed at onset Acute or gradual onset Married/cohabiting at onset Good work adjustment prior to onset Good pre-morbid social adjustment No pre-morbid personality disorder No alcohol abuse within one year of onset Psychosocial stressor prior to onset No familial risk for psychosis Late onset age (21 years or later) Short first hospitalization Secondary/tertiary educational level
At work b and not on any pension year 2000 a
OR (95% CI), P-value
OR (95% CI), P-value
1.51 (0.53–4.26), p = 0.44 1.07 (0.37–3.09), p = 0.90 1.72 (0.58–5.13), p = 0.33 6.51 (1.83–23.12), p = 0.004 1.57 (0.47–5.21), p = 0.46 1.93 (0.51–7.33), p = 0.34 0.20 (0.03–1.34), p = 0.10 0.08 (0.01–0.58), p = 0.01 10.01 (1.31–76.33), p = 0.03 0.93 (0.28–3.07), p = 0.91 3.74 (1.19–11.77), p = 0.02 2.26 (0.71–7.22), p = 0.17 3.23 (0.70–14.89), p = 0.13
0.26 (0.07–0.96), p = 0.04 1.95 (0.53–7.13), p = 0.31 3.12 (0.80–12.26), p = 0.10 2.32 (0.51–10.52), p = 0.27 1.05 (0.24–4.66), p = 0.95 0.68 (0.38–9.20), p = 0.63 0.14 (0.02–0.97), p = 0.05 0.16 (0.02–1.12), p = 0.07 – (too small sample) 3.43 (0.66–17.92), p = 0.15 1.41 (0.39–5.10), p = 0.60 1.25 (0.37–4.21), p = 0.72 1.49 (0.25–8.86), p = 0.66
Odds ratios are adjusted for gender, educational level, being unemployed at onset and the proportion of time spent in psychiatric hospitals after onset of illness (transformed using natural logarithm). Statistically significant (Wald chi-square test, degrees of freedom = 1, P b 0.05) differences are in bold. a Information from the Social Insurance Institution of Finland. b At least 6 months during 2000, information from the Finnish Centre for Pensions, includes all work periods contributing to pension.
associated with being at work and not on pension at the end of the follow-up. Table 2 presents adjusted odds ratios for the predictors of patterns of occupational status in this sample of schizophrenic psychoses. Due to the small sample size, we present results by genders pooled. We also present the effect of gender when predicting outcome and when studying other predictors gender was added to the model as a covariate. After adjusting for gender, being unemployed at onset, educational level, and number of hospital days, those who were married or cohabiting were less often on pension than those who were single at the time of onset of illness (OR = 6.51; 95% CI 1.83–23.12). Those
subjects who had their onset of illness after the age of 20 were at lower risk for being pensioned (OR = 3.74; 95% CI 1.19–11.77). Males were more commonly at work at the end of the follow-up. Other statistically significant predictors were found as well, but these were significant only after adjusted analyses, but not in crude analyses. In Fig. 1, we present the survival curves for time to pension for males and females by marital status at the time of onset of illness (single vs. married/cohabited). The subjects who were single were pensioned sooner (mean = 3.0 years, S.D. = 2.9) than others (mean = 4.0, S.D. = 3.8). The difference was statistically significant (Log Rank test = 18.40, P b 0.0001). When adjusted for
Fig. 1. Survival curves for time to pension by marital status in living subjects with schizophrenic psychoses in males (N = 53) and females (N = 38) within the Northern Finland 1966 Birth Cohort.
222
J. Miettunen et al. / Psychiatry Research 150 (2007) 217–225
gender, being unemployed at onset, educational level and proportion of time spent in psychiatric hospitals after onset of illness in Cox regression analysis, the difference between curves remained (Wald chi-square test = 6.60, P = 0.01). When the survival curves were analyzed separately for males and females, the crude test statistics remained statistically significant (Log Rank test: Males 4.53, P = 0.03; Females 12.40, P = 0.0004), but adjusted statistics were no longer statistically significant (Wald chi-square test: Males = 2.01, P = 0.16; Females = 2.69, P = 0.10). Based on our findings, those patients who are single at the time of onset of their illness need most support to retain their contacts to the world of work. These findings may have important implications, e.g., for vocational rehabilitation of these persons. 4. Discussion 4.1. Rates of employment and pensions The unemployment rate is relatively high in Finland, ranging from 6% in southern Finland to 14% in Lapland, December 2004 (Homepage of Statistics Finland, www. stat.fi). The employment rates of schizophrenia patients are not necessarily strongly correlated with rates in the general population. For instance, an UK study using data from London found that the general employment rate was about 95% both in 1990 and in 1999, but the employment rate among longer term mental health service users decreased from 20% to 8% during that period (Perkins and Rinaldi, 2002). The authors conclude that this difference may reflect the greater discrimination experienced by people with mental health problems. Mechanic et al. (2002) report data from three large US surveys; employment rates for people with no mental illness were between 76% and 84% and for people with schizophrenia or related disorders between 22% and 40%. In the large WHO International Study of Schizophrenia (ISoS), 37% of schizophrenia patients (n = 502) in incidence cohorts and 46% of schizophrenia patients (n = 87) in prevalence cohorts had been working most part of the past 2 years at the 15-year and 25-year follow-up, respectively (Harrison et al., 2001). In our sample, about 44% of the patients with schizophrenic psychoses were not pensioned at the end of the follow-up time. The large (n = 15,733) Finnish study by Suvisaari et al. (1998) reported quite similar findings, i.e., 40% of females and 36% of males were not on disability pension. Salokangas (1978) studied patients with schizophrenia or schizophreniform psychosis first admitted to hospital in 1965–1967; after a
7.5-year follow-up, two-thirds were still able to work. Achte et al. (1986) presented the results of a series of first admission cohorts of schizophrenic psychoses in Helsinki, in 1950, 1960, 1965 and 1970. The proportion of persons not on disability pension after 5 years of follow-up varied from 65% to 87%, decreasing in the older samples. Pakaslahti (1992) followed up 133 schizophrenia or schizophreniform disorder cases for 5.5 years; at the end of the study, 55% of them were not on disability pension. Our results seem to be somewhat more pessimistic than these previous comparable studies. However, it is difficult to compare the present results with previous studies because definition and general level of employment disability vary between countries and periods, and patient characteristics may also differ between studies. The main barriers to employment for people with serious mental illness described in the literature are stigmatization, economic disincentives, attitudes, and self-esteem of those with serious mental illness, and the response of mental health services to their need for support in obtaining and maintaining employment (Marwaha and Johnson, 2004). Maintaining working capacity in psychotic subjects is a major goal in the modern workplace, where high-level cognitive capacities and emotional facilities in human networks are needed; in addition to an ability to tolerate stress, a fast work pace, conflicts, uncertainties, traveling—all major challenges and relapse provokers for schizophrenic persons. 4.2. Predictors of occupational status 4.2.1. Marital status It is quite common that patients who develop schizophrenic psychoses are single at the time of onset of illness. In our sample, 51% of patients (77% of males and 45% of females) were single at the time of onset. A large Danish study found that schizophrenia patients have a very high risk for being unemployed and single both before and after the onset of psychosis (Agerbo et al., 2004). Our finding that being single predicted poor occupational outcome supports some earlier studies in schizophrenia. Salokangas et al. (2001) have reported that the quality of life of single men was poorer than that of others in almost all the areas in which it was measured, including working. Studies predicting outcome of rehabilitation programs have also found marital status to associate with better outcome (reviewed by Michon et al., 2005). In healthy population, Ek et al. (2005) have reported survey data from the Northern Finland 1966
J. Miettunen et al. / Psychiatry Research 150 (2007) 217–225
Birth Cohort, finding that among males, being single was associated with employment difficulties. 4.2.2. Onset age In general, earlier onset age has been associated with more severe psychotic illness (Suvisaari et al., 1998). We did find later onset age to correlate with better occupational outcome; however, in our study, this association is affected by the period effect. Because this is a birth cohort sample, it was not possible to study the period effect (changes in hospitalization and employment rates) separate from onset age. In a large US National Health Survey on Disability, the employment pattern of schizophrenia patients differed from the general population in that younger schizophrenia patients (ages 18–24) had odds of employment almost 3.5 times higher than persons aged 45–65 (Mechanic et al., 2002). In psychiatric vocational rehabilitation programs younger age has been associated with better outcome (Michon et al., 2005). 4.2.3. Other predictors In a large Finnish register study (Suvisaari et al., 1998), it was found that schizophrenic males and those with high familial loading were at higher risk for disability pension. We had no statistically significant association in this, possibly due to the small sample size. In our sample, males were more commonly retired, but they were also more commonly at work which contributed to pension, although the difference was not statistically significant. This association was not specific to schizophrenic subjects as this was the case in the whole cohort as well. In our sample about 10% of patients with schizophrenic psychoses had a definite psychosocial stressor prior to onset of illness; these subjects were more likely to have a positive occupational outcome. We have previously found out that psychosocial stressor is also associated with positive clinical course: only 30% of these patients were re-hospitalized in 2 years, while the percentage among other patients was 62% (Miettunen et al., 2006). We also had indications that premorbid alcohol abuse has positive effect on occupational outcome. This was probably due to chance, as this association was statistically significant only after adjusted for time spent in the hospital, indicating that these subjects had a less severe illness in our sample. We have earlier found out that a short first hospitalization constitutes a risk for re-hospitalization (Miettunen et al., 2006). As expected, the occupational outcome was better among those with shorter first hospitalization, probably indicating that these patients were more often able to maintain their contact with work life.
223
4.3. Strengths and limitations of the study In earlier studies on the topic, the selection of subjects and a high proportion of dropouts have been major problems, but this was not the case in our study. This study is rare in being both population- and register-based, and the study thereby gives new and more valid information compared to most of the earlier studies. We were also able to give estimates of pension and employment rates from our whole population-based sample for comparison with corresponding rates of patients with schizophrenic psychoses. One limitation in this study is the fact that we used several predictors and had a relatively small sample, so chance findings are possible. However, no correction for multiple testing was done as all variables were hypothesized to be associated with occupational outcome. On the other hand, it is also possible that the relationships between predictors and occupational status could be associated with other underlying factors, which we were not able to study. Furthermore, it is likely that similar patient characteristics contribute to both the likelihood of being single and employment status. Our diagnostics should be accurate; we used hospital diagnoses validated according to DSM-III-R criteria, because our previous studies (Isohanni et al., 1997; Moilanen et al., 2003) have shown that schizophrenia diagnoses in particular are not always given in hospitals, although the criteria for the diagnoses are fulfilled. We collected information concerning pre-morbid factors from hospital notes using the OCCPI procedure. Case notes do not always provide information that has been clearly and unambiguously entered, which may result in scoring difficulties (McGlashan et al., 1988). It is likely that in some hospital notes, the information related to the used OCCPI variables was not presented well enough, for example, to give the pre-morbid personality disorder diagnosis. The reliability of the used OCCPI variables was not studied in our sample; McGuffin et al. (1991) report reliability for some of these items, e.g., for marital status kappa values were between 0.81 and 0.96 for three rater pairs. We have missing data especially in some of the multivariate analyses. This may have decreased our power to detect significant associations between predictors and outcome variables. Unlike in most of the earlier studies, we were able to study the effect of familial risk on occupational status. Unfortunately, we did not have access to hospital diagnoses of siblings, in addition to which the FHDR is only available from 1972 onwards, so earlier hospital periods of the parents are missing. However, we completed the FHDR data with information collected from hospital notes of the patients.
224
J. Miettunen et al. / Psychiatry Research 150 (2007) 217–225
The use of register information on hospitalizations and employment is also a strength, as the data are unbiased and there is no problem regarding recall bias. Those who were not pensioned or unemployed were not necessarily working, for example they could be students. We used register information on work periods contributing to pension; this is a strength when compared with the earlier studies. The follow-up period used in this study was based on available information; other selections might have resulted in different results and conclusions. We used two outcome variables as the variables have differences especially regarding severity of employment disability and may therefore also have different predictors. To conclude, males and those who were married or cohabiting at the time of onset of illness were less commonly pensioned. Those who are single at time of their onset of illness need support to retain their contacts to work life. It is also important for the clinical outcome of the schizophrenia patients that they are able to maintain their working ability (Bell and Lysaker, 1997; Priebe et al., 1998). The employment status is also significantly related to the quality of life among schizophrenia patients (Eklund et al., 2001). Many studies have presented different types of intervention programs for this purpose and our findings may have important implications for vocational rehabilitation in persons with schizophrenic psychoses. Acknowledgements This work was supported by grants from the Finnish Academy, the Sigrid Juselius Foundation, the Oy H. Lundbeck Ab Finland, and the Stanley Medical Research Institute. References Achte, K., Lönnqvist, J., Kuusi, K., Piirtola, O., Niskanen, P., 1986. Outcome studies on schizophrenic psychoses in Helsinki. Psychopathology 19, 60–67. Agerbo, E., Byrne, M., Eaton, W.W., Mortensen, P.B., 2004. Marital and labor market status in the long run in schizophrenia. Archives of General Psychiatry 61, 28–33. American Psychiatric Association, 1987. DSM-III-R: Diagnostic and Statistical Manual of Mental Disorders, 3rd edn. American Psychiatric Association, Washington. revised. Appleby, L., Desai, P.N., Luchins, D.J., Gibbons, R.D., Hedeker, D.R., 1993. Length of stay and recidivism in schizophrenia: a study of public psychiatric hospital patients. American Journal of Psychiatry 150, 72–76. Bell, M.D., Lysaker, P.H., 1997. Clinical benefits of paid work activity in schizophrenia: 1-year followup. Schizophrenia Bulletin 23, 317–328. Ek, E., Sovio, U., Remes, J., Järvelin, M.-R., 2005. Social predictors for unsuccessful entrance into the labour market. Journal of Vocational Behavior 66, 471–486.
Eklund, M., Hansson, L., Bejerholm, U., 2001. Relationships between satisfaction with occupational factors and health-related variables in schizophrenia outpatients. Social Psychiatry and Psychiatric Epidemiology 36, 79–85. Harrison, G., Croudace, T., Mason, P., Glazebrook, C., Medley, I., 1996. Predicting the long-term outcome of schizophrenia. Psychological Medicine 26, 697–705. Harrison, G., Hopper, K., Craig, T., Laska, E., Siegel, C., Wanderling, J., Dube, K.C., Ganev, K., Giel, R., der Heiden, W., Holmberg, S.K., Janca, A., Lee, P.W., Leon, C.A., Malhotra, S., Marsella, A.J., Nakane, Y., Sartorius, N., Shen, Y., Skoda, C., Thara, R., Tsirkin, S.J., Varma, V.K., Walsh, D., Wiersma, D., 2001. Recovery from psychotic illness: a 15- and 25-year international follow-up study. British Journal of Psychiatry 178, 506–517. Isohanni, M., Mäkikyrö, T., Moring, J., Räsänen, P., Hakko, H., Partanen, U., Koiranen, M., Jones, P., 1997. A comparison of clinical and research DSM-III-R diagnoses of schizophrenia in a Finnish national birth cohort. Social Psychiatry and Psychiatric Epidemiology 32, 303–308. McGlashan, T.H., Carpenter, W.T., Bartko, J.J., 1988. Issues of design and methodology in long-term followup studies. Schizophrenia Bulletin 14, 569–574. McGuffin, P., Farmer, A.E., Harvey, I., 1991. A polydiagnostic application of operational criteria in psychotic illness: development and reliability of the OPCRIT system. Archives of General Psychiatry 48, 764–770. Marwaha, S., Johnson, S., 2004. Schizophrenia and employment. A review. Social Psychiatry and Psychiatric Epidemiology 39, 337–349. Mechanic, D., Bilder, S., McAlpine, D.D., 2002. Employing persons with serious mental illness. Health Affairs 21, 242–253. Michon, H.W.C., van Heeghel, J., Kroon, H., Schene, A.H., 2005. Person-related predictors of employment outcomes after participation in psychiatric vocational rehabilitation programmes. A systematic review. Social Psychiatry and Psychiatric Epidemiology 40, 408–416. Miettunen, J., Lauronen, E., Veijola, J., Koponen, H., Saarento, O., Isohanni, M., 2006. Patterns of psychiatric hospitalizations in schizophrenic psychoses within the Northern Finland 1966 Birth Cohort. Nordic Journal of Psychiatry 56, 223–228. Moilanen, K., Veijola, J., Läksy, K., Mäkikyrö, T., Miettunen, J., Kantojärvi, L., Kokkonen, P., Karvonen, J.T., Herva, A., Joukamaa, M., Järvelin, M.-R., Moring, J., Jones, P.B., Isohanni, M., 2003. Reasons for the diagnostic discordance between clinicians and researchers in schizophrenia in the Northern Finland 1966 Birth Cohort. Social Psychiatry and Psychiatric Epidemiology 38, 305–310. Mueser, K.T., Salyers, M.P., Mueser, P.R., 2001. A prospective analysis of work in schizophrenia. Schizophrenia Bulletin 27, 281–296. Øiesvold, T., Saarento, O., Sytema, S., Vinding, H., Göstas, G., Lönnerberg, O., Muus, S., Sandlund, M., Hansson, L., 2000. Predictors for readmission risk of new patients: The Nordic Comparative Study on Sectorized Psychiatry. Acta Psychiatrica Scandinavica 101, 367–373. Pakaslahti, A., 1992. Prediction of Working Disability in Schizophrenia: A 5-Year Prospective Study of a Representative Cohort of First-Admissions. Social Insurance Institution, Helsinki. Perkins, R., Rinaldi, M., 2002. Unemployment rates among patients with long-term mental health problems. Psychiatric Bulletin 26, 295–298. Priebe, S., Warner, R., Hubschmid, T., Eckle, I., 1998. Employment, attitudes toward work, and quality of life among people with
J. Miettunen et al. / Psychiatry Research 150 (2007) 217–225 schizophrenia in three countries. Schizophrenia Bulletin 24, 469–477. Rantakallio, P., 1969. Groups at risk in low birth weight infants and perinatal mortality. Acta Paediatrica Scandinavica 193, 1–71 (Suppl). Saarento, O., Øiesvold, T., Sytema, S., Göstas, G., Kastrup, M., Lönnerberg, O., Muus, S., Sandlund, M., Hansson, L., 1998. The Nordic Comparative Study on Sectorized Psychiatry: continuity of care related to characteristics of the psychiatric services and the patients. Social Psychiatry and Psychiatric Epidemiology 33, 521–527. Salokangas, R.K., 1978. Psychosocial Prognosis in Schizophrenia. Annales Universitatis Turkuensis. University of Turku, Turku.
225
Salokangas, R.K.R., Honkonen, T., Stengård, E., Koivisto, A.-M., 2001. To be or not to be married—that is the question of quality of life in men with schizophrenia. Social Psychiatry and Psychiatric Epidemiology 36, 381–390. SPSS Inc., 2003. SPSS Base 12.0 User's Guide. SPSS Inc., Chicago. Suvisaari, J.M., Haukka, J., Tanskanen, A., Lönnqvist, J.K., 1998. Age at onset and outcome in schizophrenia are related to the degree of familial loading. British Journal of Psychiatry 173, 494–500. Tsang, H., Lam, P., Ng, B., Leung, P., 2000. Predictors of employment outcomes for people with psychiatric disabilities: a review of literature since the mid '80s. Journal of Rehabilitation 66, 19–31.
Socio-demographic and clinical predictors of occupational status in schizophrenic psychoses—follow-up within the Northern Finland 1966 Birth Cohort Jouko Miettunen a,⁎, Erika Lauronen a , Juha Veijola a,b , Hannu Koponen b,c , Outi Saarento d , Anja Taanila e , Matti Isohanni a a
Department of Psychiatry, University of Oulu, P.O.Box 5000, 90014 Oulun Yliopisto, Finland b Academy of Finland, Helsinki, Finland c Department of Psychiatry, University of Kuopio, Kuopio, Finland d Department of Psychiatry, Oulu University Hospital, Oulu, Finland e Department of Public Health and General Practice, University of Oulu, Oulu, Finland Received 3 March 2006; received in revised form 16 June 2006; accepted 30 August 2006
Abstract We studied occupational status of persons with schizophrenic psychoses by age 34 in a longitudinal population-based cohort and predicted which demographic and illness-related factors could support the patients to maintain their occupational capacity. Subjects of the Northern Finland 1966 Birth Cohort with the diagnosis of DSM-III-R schizophrenic psychoses (n = 113) by the year 1997 were followed until the end of year 2000. Various illness and socio-demographic factors at the time of onset of illness were used as predictors. At the end of the follow-up time 50 (44%) of patients were not pensioned and 22 (20%) were also working at least half of the time during year 2000. After adjusting for gender, being unemployed at onset, educational level and proportion of time spent in psychiatric hospitals, those who were married or cohabiting at the time of onset of illness were less often on pension than those who were single (OR 6.51; 95% CI 1.83–23.12). Thus, nearly half of the patients with schizophrenic psychoses were not pensioned after an average 10 years follow-up. Based on our findings, those who were single at time of their onset of illness probably need most support to retain their contacts to work life. © 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Employment; Pension; Psychotic disorders; Schizophrenia; Functioning; Marital status
1. Introduction Schizophrenia is a severe, often life-long disease, which may cause severe functional decline and unemployment in the patients. Maintaining working ability is
⁎ Corresponding author. Tel.: +358 8 3156923; fax: +358 8 333167. E-mail address: [email protected] (J. Miettunen).
an important goal during the course of schizophrenia (Bell and Lysaker, 1997; Priebe et al., 1998). Many studies have reported low rates of employment in schizophrenia patient. A recent study by Marwaha and Johnson (2004) reviewed 22 studies with patients having schizophrenia. They reported wide variation in employment rates (between 4% and 90%). Most of the high rates were from the International Pilot Study of Schizophrenia (IPSS), where study samples excluded
0165-1781/$ - see front matter © 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.psychres.2006.08.011
218
J. Miettunen et al. / Psychiatry Research 150 (2007) 217–225
people with chronic schizophrenia with poorer outcome. In the same review, the employment rates were between 13% and 65% in first-episode psychosis studies. There have been some studies which have found predictors for occupational outcome (employment or disability pension) in schizophrenia subjects. Chronicity of illness (Tsang et al., 2000; Mueser et al., 2001) and male gender (Harrison et al., 1996; Agerbo et al., 2004) has been associated with poor occupational functioning in several studies. Also prior work experience, patient's and mother's higher educational level, and good social functioning have predicted work ability in people with schizophrenia (Mueser et al., 2001). In this article, we explore occupational status in schizophrenic psychoses within an early middle-aged, general population-based sample of the Northern Finland 1966 Birth Cohort. By using various demographic and illness-related variables, we also predict which of them could support the patients in maintaining their employment. We use register and hospital notes-based data, which minimizes the typical problem of drop-outs seen in most of the previous studies. Our special focus is on the variables concerning the time of the early phase of psychosis and first psychiatric hospitalization. 2. Methods 2.1. Data collection The Northern Finland 1966 Birth Cohort is an unselected, general population birth cohort ascertained during mid-pregnancy. It is based upon 12,058 live-born children in the provinces of Lapland and Oulu (Rantakallio, 1969). Permission to gather data was obtained from the Ministry of Social and Health Affairs and the study design is under review by the Ethical Committee of the Northern Ostrobothnia Hospital District. Altogether, 83 subjects have forbidden the use of their data and have been excluded. The current data include all subjects living in Finland at the age 16 years and still alive at the end of the year 2000 (n = 10,748). 2.2. Subjects of the study The nationwide Finnish Hospital Discharge Register (FHDR) covers all mental and general hospitals as well as beds in local health centers and private hospitals nationwide. All cohort members over 16 years appearing on the FHDR until the end of 1997 for any mental disorder were identified and their diagnoses were rechecked twice by professionals using DSM-III-R criteria (APA, 1987). The reliability for schizophrenia diagnoses
of this procedure was good (kappa = 0.85). A more detailed description of the validation process is presented elsewhere (Isohanni et al., 1997; Moilanen et al., 2003). By the end of 1997, there were 153 subjects (90 men, 59%) in the FHDR with a known psychotic episode in their life. Of these cases, 128 (76 men, 59%) were given the diagnosis of schizophrenia spectrum in the validation process. Eleven (9 men, 82%) of these were deceased by the end of the year 2000 and four subjects were pensioned before the onset of psychosis: one female due to unspecified delay in development (ICD-9 diagnosis: 31599); one female due to mild mental retardation (ICD-9: 317); and one male and one female due to epilepsy (ICD-9: 345). These subjects were excluded from this study. The final sample included all living subjects with schizophrenic psychoses who have not been pensioned due to developmental or neurological disorder (n = 113 cases; 67 men, 59%). The sample included 85 schizophrenia cases and 28 other schizophrenia spectrum cases. 2.3. Assessment of occupational status We collected information on all disability pensions from the Social Insurance Institution of Finland. In Finland, disability pension is granted if the disorder is persisting and incapacitating for an indefinite or timelimited period, usually for 1–2 years. Subjects on disability pension are allowed to have small earnings. We also collected information concerning all work periods contributing to pension from the Central Pension Security Institute. All these data sets were available until the end of year 2000. We summarized positive occupational status to two dichotomized variables representing positive occupational outcome: Not on vs. on pension at the end of year 2000; and not on pension at the end of 2000 and working 50% of all days during the year 2000 vs. on pension or working less than 50% during the year 2000. 2.4. Predictors of occupational status We collected information on predictors from registers and hospital notes. Various socio-demographic factors as well as illness-related variables were analyzed: 2.4.1. Gender Male vs. female. Gender was also used as a covariate. 2.4.2. Pre-morbid factors Information on pre-morbid factors was collected with a retrospective review of hospital notes using the
J. Miettunen et al. / Psychiatry Research 150 (2007) 217–225
Operational Criteria Checklist for Psychotic Illness (OCCPI) (McGuffin et al., 1991). The variables referred to the time before or around onset of illness. Poor work adjustment (no/yes) around the illness onset was scored, e.g., if the patient was unable to keep any job for more than 6 months, had a history of frequent changes of job, or was only able to sustain a job well below that expected by his/her educational level. Poor pre-morbid social adjustment (no/yes) refers, for example, to difficulties in maintaining normal social relationships or to persistent social isolation. Pre-morbid personality disorder (no/yes) refers to any evidence of personality disorder present prior to the onset of psychosis. Alcohol abuse within 1 year of onset of psychotic symptoms (no/yes) was rated if quantity was excessive (rater judgement) and alcohol-related complications had occurred. Definite psychosocial stressor (no/yes) was scored if a severely threatening event had occurred prior to onset of disorder that was unlikely to have resulted from the subject's own behavior. Also, the following variables were used: mode of onset (acute or gradual/ insidious), marital status (married/single), and unemployed at illness onset (no/yes). OCCPI data were available in some form for 109 (96.5%) subjects of the sample, although some more data were missing in the single OCCPI items. The reasons for missing data were lack of hospital notes or insufficient information given in the notes. 2.4.3. Familial risk for psychosis First-degree relative having vs. not having psychosis. Assessed by hospital notes and by using data on mothers' and fathers' psychotic hospitalizations (FHDR) between 1972 and 2000. 2.4.4. Length of first psychiatric hospitalization The length of first hospitalization was used as a dichotomized variable (1–14 days, 15 days or more). This categorization is comparable with some earlier studies (Appleby et al., 1993; Saarento et al., 1998; Øiesvold et al., 2000); we have used this categorization in a previous study as well (Miettunen et al., 2006). 2.4.5. Proportion of time spent in psychiatric hospitals after onset of illness Information on psychiatric hospitalization was collected from the FHDR until the end of 2000 (Miettunen et al., 2006). The proportion of time spent in psychiatric hospitals after onset of illness was used as a covariate, as this variable is a proxy of the chronicity of illness and also takes into account different lengths of follow-up times of the subjects. Time in hospital is taken into
219
account also because it may hinder maintaining the contact with the world of work. As the variable was extremely skewed, we transformed the original variable taking a natural logarithm. 2.4.6. Educational level The information on education of the subjects was collected in 1997 from the National Educational Registry of Statistics, Finland. The variable includes three categories according to the length of education: primary level (less than 10 years), secondary level (10–12 years) and tertiary level (more than 12 years) education. 2.5. Comparison to general population We also report previously unpublished data on pension rates and work periods in the total sample of the Northern Finland 1966 Birth Cohort Study living in Finland 1997 and being alive at the end of follow-up (n = 10,748; 5452 males and 5296 females). This is done in order to make the presented pension and employment rates in schizophrenic psychoses comparable to the general population. In the total sample, we were able to study the effect of gender and education level on occupational status. 2.6. Statistical analyses Differences between outcome groups were analyzed with chi-square tests and odds ratios (dichotomized variables) and trend tests (linear-by-linear association statistic, ordinal variables). All tests were two-sided. Survival analysis with Log Rank test statistics was used to study the time to being pensioned when comparing patients by their marital status at the time of onset of illness. Logistic and Cox regression analyses were used to adjust for confounding. In these multivariate analyses gender, being unemployed at onset, educational level, and the proportion of time spent in psychiatric hospitals after onset of illness (natural logarithm) were used as covariates. Statistical analyses were made with SPSS 12.0 for Windows (SPSS Inc., 2003). 3. Results In the total Northern Finland 1966 Birth Cohort 2.3% were pensioned at the end of 2000 (247/10,736). Males (2.4%; 131/5442) were more commonly pensioned than females (2.2%; 116/5294). In the total sample, 69.0% (7421/10,748) were working at least 50% of days and were not pensioned during the year 2000. Males (77.6%; 4233/5452) were more commonly working than females
220
J. Miettunen et al. / Psychiatry Research 150 (2007) 217–225
Table 1 Socio-demographic and symptom variables of living subjects with schizophrenic psychoses by different occupational variables within the Northern Finland 1966 Birth Cohort Predictor variables (n = males/ females)
Pre-morbid factors
Male (n = 67)
Female (n = 46)
Not on any pension at the end of year 2000 a
At work b and not on any pension year 2000 a
Not on any pension at the end of year 2000 a
At work b and not on any pension year 2000 a
n
n
n
n
%
%
%
%
Unemployed at onset (n = 63/40) No 19/37 51.4 13/37 35.1 12/28 42.9 2/28 7.1 Yes 6/26 23.1 2/26 7.7 8/12 66.7 2/12 16.7 Poor work adjustment (n = 57/41) No 20/38 52.6 13/38 34.2 13/30 43.3 1/30 3.3 Yes 4/19 21.1 1/19 5.3 7/11 63.6 3/11 27.3 Mode of onset (n = 58/42) Acute 8/13 61.5 Gradual 9/20 45.0 Insidious 7/25 28.0 Marital status (n = 53/38) Married or 8/12 66.7 cohabiting Single 9/41 22.0
5/13 38.5 5/20 25.0 4/25 16.0
5/7 71.4 1/7 14.3 7/13 53.8 3/13 23.1 8/22 36.4 0/22 0.0
4/12 33.3 16/21 76.2 4/21 19.0 4/41
9.8
3/17 17.6 1/17
5.9
Poor pre-morbid social adjustment (n = 37/30) No 9/14 64.3 4/14 28.6 11/20 55.0 2/20 10.0 Yes 5/23 21.7 3/23 13.0 6/10 60.0 2/10 20.0 Pre-morbid personality disorder (n = 35/25) No 6/24 25.0 4/24 16.7 7/19 36.8 1/19 5.3 Yes 5/11 45.5 4/11 36.4 3/6 50.0 1/6 16.7 Alcohol abuse within 1 year of onset (n = 62/43) No 18/52 34.6 12/52 23.1 20/41 48.8 4/41 9.8 Yes 6/10 60.0 3/10 30.0 1/2 50.0 1/2 50.0 Psychosocial stressor prior to onset (n = 63/45) No 22/58 37.9 15/58 25.9 18/40 45.0 5/40 12.5 Yes 2/5 40.0 0/5 0.0 4/5 80.0 0/5 0.0 Familial risk for psychosis c (n = 67/46) No 22/54 40.7 14/54 25.9 15/34 44.1 4/34 11.8 Yes 6/13 46.2 3/13 23.1 7/12 58.3 1/12 8.3 Onset age before 21 years (n = 67/46) No 24/44 54.5 14/44 31.8 16/27 59.3 3/27 11.1 Yes 4/23 17.4 3/23 13.0 6/19 31.6 2/19 10.5 Length of first hospitalization (n = 66/46) 1–14 days 13/24 54.2 7/24 29.2 6/10 60.0 2/10 20.0 15 days or 14/42 33.3 9/42 21.4 16/36 44.4 3/36 8.3 more
Table 1 (continued) Predictor variables (n = males/ females)
Pre-morbid factors
Male (n = 67)
Female (n = 46)
Not on any pension at the end of year 2000 a
At work b and not on any pension year 2000 a
Not on any pension at the end of year 2000 a
At work b and not on any pension year 2000 a
n
n
n
n
%
%
%
%
Educational level at year 1997 (n = 67/46) Basic 3/14 21.4 2/14 14.3 1/7 14.3 0/7 0.0 Secondary 23/49 46.9 15/49 30.6 17/34 50.0 4/34 11.8 Tertiary 2/4 50.0 0/4 0.0 4/5 80.0 1/5 20.0 Statistically significant (P b 0.05; chi-square test or Fisher exact test as appropriate, or trend test if more than two categories) differences are in bold. a Information from the Social Insurance Institution of Finland. b At least 6 months during the year 2000, information from the Finnish Centre for Pensions, includes all work periods contributing to pension. c Collected from hospital notes for all first-degree relatives and additionally for parents from the Finnish Hospital Discharge Register until the end of 2000.
(60.2%; 3188/5296). Cohort subjects with tertiary education (72.8%; 1978/2718) and secondary education (71.4%; 4630/6483) were more commonly at work than subjects with basic education (52.5%; 806/1535). These differences in gender and education were statistically significant (chi-square tests, P b 0.001). In patients with schizophrenic psychoses, the mean (standard deviation) follow-up time of the sample was 10.6 (S.D. = 4.2) years. Sixty-three (55.8%) patients were pensioned at the end of the follow-up time. The average time from the first psychiatric admission to pension was 3.3 (S.D. = 3.1) years. Fifty (44.2%) patients were not pensioned at the end of the follow-up time and 22 (19.5%) also worked for at least 6 months during the year 2000. The corresponding figures among males were 28 (41.8%) and 17 (25.4%); and among females 22 (47.8%) and 5 (10.9%), respectively. The socio-demographic and clinical characteristics of the subjects with schizophrenic psychoses by gender are presented in Table 1 by the two occupational status variables. Among men being employed and married or cohabiting at the time of onset, good pre-morbid work and social adjustment, and earlier age at onset were associated (P b 0.05) with better occupational outcome at least in one of the outcome variables. Among women being single associated with being pensioned, and higher education, and prior good work adjustment were
J. Miettunen et al. / Psychiatry Research 150 (2007) 217–225
221
Table 2 Predicting working ability in subjects with schizophrenic psychoses within the Northern Finland 1966 Birth Cohort Not on any pension at the end of year 2000 a
Female gender Employed at onset Acute or gradual onset Married/cohabiting at onset Good work adjustment prior to onset Good pre-morbid social adjustment No pre-morbid personality disorder No alcohol abuse within one year of onset Psychosocial stressor prior to onset No familial risk for psychosis Late onset age (21 years or later) Short first hospitalization Secondary/tertiary educational level
At work b and not on any pension year 2000 a
OR (95% CI), P-value
OR (95% CI), P-value
1.51 (0.53–4.26), p = 0.44 1.07 (0.37–3.09), p = 0.90 1.72 (0.58–5.13), p = 0.33 6.51 (1.83–23.12), p = 0.004 1.57 (0.47–5.21), p = 0.46 1.93 (0.51–7.33), p = 0.34 0.20 (0.03–1.34), p = 0.10 0.08 (0.01–0.58), p = 0.01 10.01 (1.31–76.33), p = 0.03 0.93 (0.28–3.07), p = 0.91 3.74 (1.19–11.77), p = 0.02 2.26 (0.71–7.22), p = 0.17 3.23 (0.70–14.89), p = 0.13
0.26 (0.07–0.96), p = 0.04 1.95 (0.53–7.13), p = 0.31 3.12 (0.80–12.26), p = 0.10 2.32 (0.51–10.52), p = 0.27 1.05 (0.24–4.66), p = 0.95 0.68 (0.38–9.20), p = 0.63 0.14 (0.02–0.97), p = 0.05 0.16 (0.02–1.12), p = 0.07 – (too small sample) 3.43 (0.66–17.92), p = 0.15 1.41 (0.39–5.10), p = 0.60 1.25 (0.37–4.21), p = 0.72 1.49 (0.25–8.86), p = 0.66
Odds ratios are adjusted for gender, educational level, being unemployed at onset and the proportion of time spent in psychiatric hospitals after onset of illness (transformed using natural logarithm). Statistically significant (Wald chi-square test, degrees of freedom = 1, P b 0.05) differences are in bold. a Information from the Social Insurance Institution of Finland. b At least 6 months during 2000, information from the Finnish Centre for Pensions, includes all work periods contributing to pension.
associated with being at work and not on pension at the end of the follow-up. Table 2 presents adjusted odds ratios for the predictors of patterns of occupational status in this sample of schizophrenic psychoses. Due to the small sample size, we present results by genders pooled. We also present the effect of gender when predicting outcome and when studying other predictors gender was added to the model as a covariate. After adjusting for gender, being unemployed at onset, educational level, and number of hospital days, those who were married or cohabiting were less often on pension than those who were single at the time of onset of illness (OR = 6.51; 95% CI 1.83–23.12). Those
subjects who had their onset of illness after the age of 20 were at lower risk for being pensioned (OR = 3.74; 95% CI 1.19–11.77). Males were more commonly at work at the end of the follow-up. Other statistically significant predictors were found as well, but these were significant only after adjusted analyses, but not in crude analyses. In Fig. 1, we present the survival curves for time to pension for males and females by marital status at the time of onset of illness (single vs. married/cohabited). The subjects who were single were pensioned sooner (mean = 3.0 years, S.D. = 2.9) than others (mean = 4.0, S.D. = 3.8). The difference was statistically significant (Log Rank test = 18.40, P b 0.0001). When adjusted for
Fig. 1. Survival curves for time to pension by marital status in living subjects with schizophrenic psychoses in males (N = 53) and females (N = 38) within the Northern Finland 1966 Birth Cohort.
222
J. Miettunen et al. / Psychiatry Research 150 (2007) 217–225
gender, being unemployed at onset, educational level and proportion of time spent in psychiatric hospitals after onset of illness in Cox regression analysis, the difference between curves remained (Wald chi-square test = 6.60, P = 0.01). When the survival curves were analyzed separately for males and females, the crude test statistics remained statistically significant (Log Rank test: Males 4.53, P = 0.03; Females 12.40, P = 0.0004), but adjusted statistics were no longer statistically significant (Wald chi-square test: Males = 2.01, P = 0.16; Females = 2.69, P = 0.10). Based on our findings, those patients who are single at the time of onset of their illness need most support to retain their contacts to the world of work. These findings may have important implications, e.g., for vocational rehabilitation of these persons. 4. Discussion 4.1. Rates of employment and pensions The unemployment rate is relatively high in Finland, ranging from 6% in southern Finland to 14% in Lapland, December 2004 (Homepage of Statistics Finland, www. stat.fi). The employment rates of schizophrenia patients are not necessarily strongly correlated with rates in the general population. For instance, an UK study using data from London found that the general employment rate was about 95% both in 1990 and in 1999, but the employment rate among longer term mental health service users decreased from 20% to 8% during that period (Perkins and Rinaldi, 2002). The authors conclude that this difference may reflect the greater discrimination experienced by people with mental health problems. Mechanic et al. (2002) report data from three large US surveys; employment rates for people with no mental illness were between 76% and 84% and for people with schizophrenia or related disorders between 22% and 40%. In the large WHO International Study of Schizophrenia (ISoS), 37% of schizophrenia patients (n = 502) in incidence cohorts and 46% of schizophrenia patients (n = 87) in prevalence cohorts had been working most part of the past 2 years at the 15-year and 25-year follow-up, respectively (Harrison et al., 2001). In our sample, about 44% of the patients with schizophrenic psychoses were not pensioned at the end of the follow-up time. The large (n = 15,733) Finnish study by Suvisaari et al. (1998) reported quite similar findings, i.e., 40% of females and 36% of males were not on disability pension. Salokangas (1978) studied patients with schizophrenia or schizophreniform psychosis first admitted to hospital in 1965–1967; after a
7.5-year follow-up, two-thirds were still able to work. Achte et al. (1986) presented the results of a series of first admission cohorts of schizophrenic psychoses in Helsinki, in 1950, 1960, 1965 and 1970. The proportion of persons not on disability pension after 5 years of follow-up varied from 65% to 87%, decreasing in the older samples. Pakaslahti (1992) followed up 133 schizophrenia or schizophreniform disorder cases for 5.5 years; at the end of the study, 55% of them were not on disability pension. Our results seem to be somewhat more pessimistic than these previous comparable studies. However, it is difficult to compare the present results with previous studies because definition and general level of employment disability vary between countries and periods, and patient characteristics may also differ between studies. The main barriers to employment for people with serious mental illness described in the literature are stigmatization, economic disincentives, attitudes, and self-esteem of those with serious mental illness, and the response of mental health services to their need for support in obtaining and maintaining employment (Marwaha and Johnson, 2004). Maintaining working capacity in psychotic subjects is a major goal in the modern workplace, where high-level cognitive capacities and emotional facilities in human networks are needed; in addition to an ability to tolerate stress, a fast work pace, conflicts, uncertainties, traveling—all major challenges and relapse provokers for schizophrenic persons. 4.2. Predictors of occupational status 4.2.1. Marital status It is quite common that patients who develop schizophrenic psychoses are single at the time of onset of illness. In our sample, 51% of patients (77% of males and 45% of females) were single at the time of onset. A large Danish study found that schizophrenia patients have a very high risk for being unemployed and single both before and after the onset of psychosis (Agerbo et al., 2004). Our finding that being single predicted poor occupational outcome supports some earlier studies in schizophrenia. Salokangas et al. (2001) have reported that the quality of life of single men was poorer than that of others in almost all the areas in which it was measured, including working. Studies predicting outcome of rehabilitation programs have also found marital status to associate with better outcome (reviewed by Michon et al., 2005). In healthy population, Ek et al. (2005) have reported survey data from the Northern Finland 1966
J. Miettunen et al. / Psychiatry Research 150 (2007) 217–225
Birth Cohort, finding that among males, being single was associated with employment difficulties. 4.2.2. Onset age In general, earlier onset age has been associated with more severe psychotic illness (Suvisaari et al., 1998). We did find later onset age to correlate with better occupational outcome; however, in our study, this association is affected by the period effect. Because this is a birth cohort sample, it was not possible to study the period effect (changes in hospitalization and employment rates) separate from onset age. In a large US National Health Survey on Disability, the employment pattern of schizophrenia patients differed from the general population in that younger schizophrenia patients (ages 18–24) had odds of employment almost 3.5 times higher than persons aged 45–65 (Mechanic et al., 2002). In psychiatric vocational rehabilitation programs younger age has been associated with better outcome (Michon et al., 2005). 4.2.3. Other predictors In a large Finnish register study (Suvisaari et al., 1998), it was found that schizophrenic males and those with high familial loading were at higher risk for disability pension. We had no statistically significant association in this, possibly due to the small sample size. In our sample, males were more commonly retired, but they were also more commonly at work which contributed to pension, although the difference was not statistically significant. This association was not specific to schizophrenic subjects as this was the case in the whole cohort as well. In our sample about 10% of patients with schizophrenic psychoses had a definite psychosocial stressor prior to onset of illness; these subjects were more likely to have a positive occupational outcome. We have previously found out that psychosocial stressor is also associated with positive clinical course: only 30% of these patients were re-hospitalized in 2 years, while the percentage among other patients was 62% (Miettunen et al., 2006). We also had indications that premorbid alcohol abuse has positive effect on occupational outcome. This was probably due to chance, as this association was statistically significant only after adjusted for time spent in the hospital, indicating that these subjects had a less severe illness in our sample. We have earlier found out that a short first hospitalization constitutes a risk for re-hospitalization (Miettunen et al., 2006). As expected, the occupational outcome was better among those with shorter first hospitalization, probably indicating that these patients were more often able to maintain their contact with work life.
223
4.3. Strengths and limitations of the study In earlier studies on the topic, the selection of subjects and a high proportion of dropouts have been major problems, but this was not the case in our study. This study is rare in being both population- and register-based, and the study thereby gives new and more valid information compared to most of the earlier studies. We were also able to give estimates of pension and employment rates from our whole population-based sample for comparison with corresponding rates of patients with schizophrenic psychoses. One limitation in this study is the fact that we used several predictors and had a relatively small sample, so chance findings are possible. However, no correction for multiple testing was done as all variables were hypothesized to be associated with occupational outcome. On the other hand, it is also possible that the relationships between predictors and occupational status could be associated with other underlying factors, which we were not able to study. Furthermore, it is likely that similar patient characteristics contribute to both the likelihood of being single and employment status. Our diagnostics should be accurate; we used hospital diagnoses validated according to DSM-III-R criteria, because our previous studies (Isohanni et al., 1997; Moilanen et al., 2003) have shown that schizophrenia diagnoses in particular are not always given in hospitals, although the criteria for the diagnoses are fulfilled. We collected information concerning pre-morbid factors from hospital notes using the OCCPI procedure. Case notes do not always provide information that has been clearly and unambiguously entered, which may result in scoring difficulties (McGlashan et al., 1988). It is likely that in some hospital notes, the information related to the used OCCPI variables was not presented well enough, for example, to give the pre-morbid personality disorder diagnosis. The reliability of the used OCCPI variables was not studied in our sample; McGuffin et al. (1991) report reliability for some of these items, e.g., for marital status kappa values were between 0.81 and 0.96 for three rater pairs. We have missing data especially in some of the multivariate analyses. This may have decreased our power to detect significant associations between predictors and outcome variables. Unlike in most of the earlier studies, we were able to study the effect of familial risk on occupational status. Unfortunately, we did not have access to hospital diagnoses of siblings, in addition to which the FHDR is only available from 1972 onwards, so earlier hospital periods of the parents are missing. However, we completed the FHDR data with information collected from hospital notes of the patients.
224
J. Miettunen et al. / Psychiatry Research 150 (2007) 217–225
The use of register information on hospitalizations and employment is also a strength, as the data are unbiased and there is no problem regarding recall bias. Those who were not pensioned or unemployed were not necessarily working, for example they could be students. We used register information on work periods contributing to pension; this is a strength when compared with the earlier studies. The follow-up period used in this study was based on available information; other selections might have resulted in different results and conclusions. We used two outcome variables as the variables have differences especially regarding severity of employment disability and may therefore also have different predictors. To conclude, males and those who were married or cohabiting at the time of onset of illness were less commonly pensioned. Those who are single at time of their onset of illness need support to retain their contacts to work life. It is also important for the clinical outcome of the schizophrenia patients that they are able to maintain their working ability (Bell and Lysaker, 1997; Priebe et al., 1998). The employment status is also significantly related to the quality of life among schizophrenia patients (Eklund et al., 2001). Many studies have presented different types of intervention programs for this purpose and our findings may have important implications for vocational rehabilitation in persons with schizophrenic psychoses. Acknowledgements This work was supported by grants from the Finnish Academy, the Sigrid Juselius Foundation, the Oy H. Lundbeck Ab Finland, and the Stanley Medical Research Institute. References Achte, K., Lönnqvist, J., Kuusi, K., Piirtola, O., Niskanen, P., 1986. Outcome studies on schizophrenic psychoses in Helsinki. Psychopathology 19, 60–67. Agerbo, E., Byrne, M., Eaton, W.W., Mortensen, P.B., 2004. Marital and labor market status in the long run in schizophrenia. Archives of General Psychiatry 61, 28–33. American Psychiatric Association, 1987. DSM-III-R: Diagnostic and Statistical Manual of Mental Disorders, 3rd edn. American Psychiatric Association, Washington. revised. Appleby, L., Desai, P.N., Luchins, D.J., Gibbons, R.D., Hedeker, D.R., 1993. Length of stay and recidivism in schizophrenia: a study of public psychiatric hospital patients. American Journal of Psychiatry 150, 72–76. Bell, M.D., Lysaker, P.H., 1997. Clinical benefits of paid work activity in schizophrenia: 1-year followup. Schizophrenia Bulletin 23, 317–328. Ek, E., Sovio, U., Remes, J., Järvelin, M.-R., 2005. Social predictors for unsuccessful entrance into the labour market. Journal of Vocational Behavior 66, 471–486.
Eklund, M., Hansson, L., Bejerholm, U., 2001. Relationships between satisfaction with occupational factors and health-related variables in schizophrenia outpatients. Social Psychiatry and Psychiatric Epidemiology 36, 79–85. Harrison, G., Croudace, T., Mason, P., Glazebrook, C., Medley, I., 1996. Predicting the long-term outcome of schizophrenia. Psychological Medicine 26, 697–705. Harrison, G., Hopper, K., Craig, T., Laska, E., Siegel, C., Wanderling, J., Dube, K.C., Ganev, K., Giel, R., der Heiden, W., Holmberg, S.K., Janca, A., Lee, P.W., Leon, C.A., Malhotra, S., Marsella, A.J., Nakane, Y., Sartorius, N., Shen, Y., Skoda, C., Thara, R., Tsirkin, S.J., Varma, V.K., Walsh, D., Wiersma, D., 2001. Recovery from psychotic illness: a 15- and 25-year international follow-up study. British Journal of Psychiatry 178, 506–517. Isohanni, M., Mäkikyrö, T., Moring, J., Räsänen, P., Hakko, H., Partanen, U., Koiranen, M., Jones, P., 1997. A comparison of clinical and research DSM-III-R diagnoses of schizophrenia in a Finnish national birth cohort. Social Psychiatry and Psychiatric Epidemiology 32, 303–308. McGlashan, T.H., Carpenter, W.T., Bartko, J.J., 1988. Issues of design and methodology in long-term followup studies. Schizophrenia Bulletin 14, 569–574. McGuffin, P., Farmer, A.E., Harvey, I., 1991. A polydiagnostic application of operational criteria in psychotic illness: development and reliability of the OPCRIT system. Archives of General Psychiatry 48, 764–770. Marwaha, S., Johnson, S., 2004. Schizophrenia and employment. A review. Social Psychiatry and Psychiatric Epidemiology 39, 337–349. Mechanic, D., Bilder, S., McAlpine, D.D., 2002. Employing persons with serious mental illness. Health Affairs 21, 242–253. Michon, H.W.C., van Heeghel, J., Kroon, H., Schene, A.H., 2005. Person-related predictors of employment outcomes after participation in psychiatric vocational rehabilitation programmes. A systematic review. Social Psychiatry and Psychiatric Epidemiology 40, 408–416. Miettunen, J., Lauronen, E., Veijola, J., Koponen, H., Saarento, O., Isohanni, M., 2006. Patterns of psychiatric hospitalizations in schizophrenic psychoses within the Northern Finland 1966 Birth Cohort. Nordic Journal of Psychiatry 56, 223–228. Moilanen, K., Veijola, J., Läksy, K., Mäkikyrö, T., Miettunen, J., Kantojärvi, L., Kokkonen, P., Karvonen, J.T., Herva, A., Joukamaa, M., Järvelin, M.-R., Moring, J., Jones, P.B., Isohanni, M., 2003. Reasons for the diagnostic discordance between clinicians and researchers in schizophrenia in the Northern Finland 1966 Birth Cohort. Social Psychiatry and Psychiatric Epidemiology 38, 305–310. Mueser, K.T., Salyers, M.P., Mueser, P.R., 2001. A prospective analysis of work in schizophrenia. Schizophrenia Bulletin 27, 281–296. Øiesvold, T., Saarento, O., Sytema, S., Vinding, H., Göstas, G., Lönnerberg, O., Muus, S., Sandlund, M., Hansson, L., 2000. Predictors for readmission risk of new patients: The Nordic Comparative Study on Sectorized Psychiatry. Acta Psychiatrica Scandinavica 101, 367–373. Pakaslahti, A., 1992. Prediction of Working Disability in Schizophrenia: A 5-Year Prospective Study of a Representative Cohort of First-Admissions. Social Insurance Institution, Helsinki. Perkins, R., Rinaldi, M., 2002. Unemployment rates among patients with long-term mental health problems. Psychiatric Bulletin 26, 295–298. Priebe, S., Warner, R., Hubschmid, T., Eckle, I., 1998. Employment, attitudes toward work, and quality of life among people with
J. Miettunen et al. / Psychiatry Research 150 (2007) 217–225 schizophrenia in three countries. Schizophrenia Bulletin 24, 469–477. Rantakallio, P., 1969. Groups at risk in low birth weight infants and perinatal mortality. Acta Paediatrica Scandinavica 193, 1–71 (Suppl). Saarento, O., Øiesvold, T., Sytema, S., Göstas, G., Kastrup, M., Lönnerberg, O., Muus, S., Sandlund, M., Hansson, L., 1998. The Nordic Comparative Study on Sectorized Psychiatry: continuity of care related to characteristics of the psychiatric services and the patients. Social Psychiatry and Psychiatric Epidemiology 33, 521–527. Salokangas, R.K., 1978. Psychosocial Prognosis in Schizophrenia. Annales Universitatis Turkuensis. University of Turku, Turku.
225
Salokangas, R.K.R., Honkonen, T., Stengård, E., Koivisto, A.-M., 2001. To be or not to be married—that is the question of quality of life in men with schizophrenia. Social Psychiatry and Psychiatric Epidemiology 36, 381–390. SPSS Inc., 2003. SPSS Base 12.0 User's Guide. SPSS Inc., Chicago. Suvisaari, J.M., Haukka, J., Tanskanen, A., Lönnqvist, J.K., 1998. Age at onset and outcome in schizophrenia are related to the degree of familial loading. British Journal of Psychiatry 173, 494–500. Tsang, H., Lam, P., Ng, B., Leung, P., 2000. Predictors of employment outcomes for people with psychiatric disabilities: a review of literature since the mid '80s. Journal of Rehabilitation 66, 19–31.