ESTIMATING THE GENERALIZED IMPACT FRACTION FROM CASE-CONTROL DATA

The generalized impact fraction (also called the generalized attributa ble fraction) was introduced by Waiter (1980, American Journal of Epid emiology 112, 409-416) and Morgenstern and Bursic (1982, Journal of Co mmunity Health 7, 292-309) is a measure that generalizes the populatio n attributable fraction (attributable risk). It is defined as the frac tional reduction of a disease resulting from changing the current dist ribution of a risk factor to some modified distribution. We show that the point and variance estimator derived by Greenland and Drescher (19 93, Biometrics 49, 865-872) for fixed shift functions can be extended to situations where the shift is a probabilistic function of the actua l exposure value. The formulas are applicable for case-control designs where the cases are simply randomly selected and the controls are cho sen in one of three ways: simple random sampling, stratified random sa mpling, and frequency matching.