Soc Psychiatry Psyehiatr Epidemiol (1991) 26:273 280 Social Psychiatry and 9 Springer-Verlag 1991 PsychiatricEpidemiology Occupational characteristics and the occurrence of psychotic disorders C. Muntaner, A. Y. Tien, W. W. Eaton, and R. Garrison Department of Mental Hygiene,The Johns Hopkins UniversitySchoolof Hygiene and PublicHealth, Baltimore, Maryland, USA Accepted: August 12,1991 Summary. This study was undertaken to investigate whether individuals working under various occupational stressors are at increased risk of three forms of psychotic conditions. This paper presents prospective analyses of antecedent occupational stressors in psychotic conditions with interview data from a community sample in five US metropolitan areas - the NIMH Epidemiologic Catch- ment Area Program. Three non-overlapping conditions were defined using DSM-III definitions as assessed by the Diagnostic Interview Schedule (DIS): (1) DSM-III schizophrenia criterion A; (2) full schizophrenia; and (3) criterion A and affective episode. Artistic (RR= 3.32, 95 % CI 1.08-10.14) and construction trades occupations (RR = 2.58, 95 % CI 1.15-5.77), "noisome working con- ditions" occupations (RR = t.20, 95 % CI 0.99-1.47) and physically demanding occupations (RR=1.39, 95% CI 1.05-1.72) were associated with increased risk of developing DIS/DSM-III schizophrenia criterion A, even after adjustment for sociodemographic and psycho- pathology factors including alcohol and marijuana use. Psychologically demanding occupations (RR=0.85, 95 % CI 0.75-0.95) were associated with decreased risk of developing DIS/DSM-III schizophrenia. This finding is supported by results from experimental studies on the arousability of pre-schizophrenics. Finally, teachers (RR=11.35, 95% CI 2.56-50.38), sales occupations (RR = 4.16, 95 % CI 1.00-2t.30) and occupations charac- terized by low control over work were associated with in- creased risk of developing DIS criterion A and affective episode, resembling previous findings on occupational stressors and depression. Overall, our results replicate and extend previous work on occupational stressors and psychotic conditions through use of prospective data, several psychotic conditions, multiple assessment of oc- cupational stressors and adjustment for potential con- founders. The present study investigates whether individuals work- ing under various occupational stressors are at increased risk of developing three different psychotic conditions: DSM-III schizophrenia criterion A (delusions and hallu- cinations), DSM-III schizophrenia, and DSM-III schizo- phrenia criterion A with affective episode (psychotic af- fective syndrome). Empirical evidence on the association between occu- pation and psychotic conditions has been a part of the lit- erature on social class and schizophrenia (Liberatos et al. 1988). The inverse relationship of social class to risk of schizophrenia is one of the most replicated findings in epidemiological psychiatry (Neugebauer et al. 1980; Eaton et al. 1988). However, it has been difficult to ex- plain the relationship between macro-level variables such as social class and individual psychopathology. As a consequence, research has increasingly focussed on the study of social class-linked micro-level risk factors, such as occupational stress (Dohrenwend 1990; Johnsson 1990). In a study using treatment cases and community con- trols, Link et al. (1986) showed that first full-time occupa- tions expose schizophrenic patients to "noisome" envi- ronments (hazards, noise, heat, humidity, fumes and cold), which are found mostly in working class occupations. The authors controlled for some possible confounders includ- ing sociodemographic variables, prestige of first full-thne occupation and family history of mental illness. Link et al. concluded that a particular type of occupational stress linked to social class could be an etiologic factor for schi- zophrenia. Despite that finding, the relationship between occupa- tional stress and schizophrenia could still be artifactual for three reasons. First, a study based on treatment samples may suffer from selection bias (B erkson 1946). This is par- ticularly true for schizophrenia which often goes un- treated (Von Korff et al. 1985). Second, potential confounders such as unemployment which is associated with psychotic conditions (Brenner 1973) and with working class occupations, were not con- sidered. Other potential confounders include alcohol and marijuana use, which produce psychotic symptomatology and are correlated with certain occupations (NIAAA 1981; Gust 1989).