Prevalences of Depression Among Adults: Comparative Analysis of a Nationwide Survey and Routine Data

Abstract

There are differences in the prevalence estimates of depressive disorders based on primary and secondary data. The reasons for this are, for instance, the use of divergent indicators and varying observation periods. This study examines the prevalence of depressive disorders using survey and routine data for a comparable survey period and age range. Effects of differences between data sources and indicators are estimated. For 2010, 3 indicators are compared: in a population survey collected a) self-reported medical diagnosis of depression, b) diagnosis of depressive disorders identified by clinical interviews and c) administrative depression diagnoses collected from routine data of a statutory health insurance. In sensitivity analyses, privately insured participants of the population survey were excluded, and insured persons with care needs were excluded from routine data. The definition of administrative depression diagnosis was varied depending on the frequency of coded diagnoses and the specificity of the diagnoses. The highest prevalence (9.8%) was found for depression diagnoses from administrative data, the lowest prevalence (5.9%) in self-reported medical diagnoses of depression in the population survey. The prevalence of depression identified by clinical interviews was 8.4%. Differing age and gender-related courses of illness were found. The administrative prevalence dropped significantly if unspecific diagnoses (F3x.8, F3x.9) were excluded. Depending on the definition of depression diagnoses used in administrative data, there was a reduction in differences of prevalence compared to the self-reported medical diagnoses. Differences in prevalence based on a diagnosis of a depressive disorder identified in a clinical interview remained stable, which indicates different groups of persons.