What is an adverse effect? A possible resolution of clinical and epidemiological perspectives on neurobehavioral toxicity

Abstract

The sizes of the effects observed in studies that rely on neurobehavioral endpoints are often small. Because the mean deficits implied are more modest in magnitude than are those that correspond to the clinical criteria used to diagnose "disease," some observers dismiss them as inconsequential. Other observers argue that the mean deficits take on greater import when viewed as effects on a population rather than on individual members of the population. Several considerations germane to an effort to reconcile these perspectives are discussed: (1) the relative sensitivity of clinical diagnoses and continuously distributed scores on neurobehavioral tests as indices of adverse effect, (2) the syndromal nature of many diagnoses in pediatric neurology and neuropsychology and the implications of shifting nosology, (3) neurobehavioral test-score changes as surrogates or as prodromes for clinically significant deficits, (4) the distinction between individual risk and population risk, and (5) the tendency of the distribution of a risk factor in a population to move up and down as a whole. The clinical and epidemiological perspectives are complementary rather than incompatible.