Skewed distributions--new outcome measures

Abstract

The traditional measure of caries, the DMF index, either as prevalence or incidence of disease, has become highly positively skewed among children and young adults. Most discussion of skewed distributions has focused on the properties of statistical analyses using such data or the implications for sample sizes and subject selection in clinical trials. This paper examines the full range of epidemiologic studies, their aims and constitutive interest in order to identify the measurement problems associated with skewed DMF index data. Constitutive interests include: description; documentation; explanation and prediction; evaluation; advocacy; and, experimentation. 'New' outcome measures that would assist in reaching the aims and constitutive interests of the epidemiology of caries include caries severity grading, variants of prevalence, extent and severity and their combination into case definitions, and weighting of the components of the DMF index. Research questions for each area of 'new' outcome measures are identified as steps in the codifying of their use in the epidemiology of caries.