Adjustment of risk ratios in case-base studies (hybrid epidemiologic designs)

The recent growth of occupational epidemiology has led to a resurgence of interest in the hybrid epidemiologic study design of Kupper et al., which has come to be termed a 'case-base' or 'case-cohort' design. This paper shows how the adjustment method for the risk ratio estimate given by Kupper et al. may have substantial bias, and also shows how one must modify conventional adjustment techniques (such as the Miettinen, Woolf and Mantel-Haenszel methods) when used with hybrid designs. This paper also provides the necessary modifications for tests and confidence intervals.