Some methodologic criteria for studies in developmental neuropsychiatry.

Studies which have sought to address the question of how brain dysfunction might be an etiological factor in the development of psychiatric disorders in childhood are reviewed from the point of view of the methodological issues which arise. It is argued that common forms of behavioral disturbance are more likely to have a psychosocial rather than a neurological origin, whereas epidemiologically unusual conditions are more likely to have neuro-developmental causal factors. Possible study designs are reviewed with the recommendation that cohort studies most closely approximate the ideal of the randomized control trial, whereas case control studies are weakest. The assessment of the reliability and validity of measures of brain dysfunction is discussed. Criteria for attributing causation after an association has been identified include temporal sequence, a 'dose-response' relationship, and the specificity of the cause. The clinical as opposed to the statistical significance of the association must be assessed. Confounding variables in such studies, variables which account in part for the association identified, but which undermine the etiological significance which can be attributed, are discussed. The particular importance of psychosocial disadvantage in this field is stressed as a factor to be taken into account in the critical evaluation of studies.