Epidemiological fallacies of modern psychiatric research.

BACKGROUND: Psychiatric epidemiology is an important cornerstone of research in psychiatry and integral for the treatment and care of people suffering from psychiatric disorders. However, psychiatric epidemiology is a difficult science, which is often beset with methodological problems.

AIMS: In light of this, the current review sought to explore 13 of the common methodological issues in psychiatric epidemiology.

METHODS: Many methodological problems result from misunderstandings. As such, we sought to highlight these problems, provide evidence to counteract the myths surrounding these problems and subsequently provide recommendations to overcome these problems. To highlight and clarify these issues, examples are provided from current psychiatric literature.

RESULTS: Areas discussed in the review include problems with: taxonometry of disorders, sole reliance on self-reports, single-question diagnoses, baseline participation rates, measurement of lifetime prevalence, inconsistency of multiple informants, selection of covariates, testing of interactions, correction for multiple testing, the intermittent measurement of disorders during follow-up, evaluation of causal associations, data invalidation related to loss from follow-up and the publication of negative findings.

CONCLUSION: Many methodological myths prevail in the area of epidemiology and this review endeavoured to elucidate and clarify these. This review was developed as a teaching tool for students, clinicians and researchers.