Aims: Higher rates of psychiatric morbidity among non-participants may lead to biased estimates of prevalence and incidence in epidemiological studies of psychiatric disorders. We had a unique opportunity to explore psychiatric morbidity and non-participation in a large epidemiological survey including questionnaires and a clinical examination.
Methods: Members of the Northern Finland 1966 Birth Cohort were included in the study. In phase I, a postal questionnaire was mailed to all those with a known address in 1997 (N=11,540). In phase II, all subjects living in northern Finland or the Helsinki area (N=8,463) were invited to a clinical examination. In phase III, clinical examination participants were given a questionnaire with psychological subscales to be filled in at home and returned by mail. The data on hospital-treated psychiatric disorders were obtained from the Finnish Hospital Discharge Register. Educational level was obtained from Statistics Finland.
Results: The participation rates were 76%, 71% and 61% in phases I, II and III, respectively. Subjects with any psychiatric disorder participated less actively than those without any psychiatric disorder in all phases, in both genders and at all educational levels. Participation was not found to vary across specific disorders. Gender or education did not explain the association of psychiatric disorders with participation.
Conclusions: Owing to non-participation, the true prevalence of psychiatric disorders may be higher than the prevalence estimated from epidemiological field surveys.