Quantifying The Proportion Of Cases Attributable To An Exposure

The attributable fraction and the average attributable fractions, which are commonly used to assess the relative effect of several exposures to the prevalence of a disease, do not represent the proportion of cases caused by each exposure. Furthermore, the sum of attributable fractions over all exposures generally exceeds not only the attributable fraction for all exposures taken together, but also 100%. Other measures are discussed here, including the directly attributable fraction and the confounding fraction, that may be more suitable in defining the fraction directly attributable to an exposure.