Logistic regression for two-stage case-control data

From MaRDI portal
Publication:3774797





In an epidemiological study the exposure of primary interest and the disease outcome usually are known for a large number of subjects but the covariable data needed to adjust for confounding effects, for instance because of their high cost or other reasons, can only be collected for a much smaller number. In the paper the case is regarded where samples of diseased cases and nondiseased controls are drawn at random from the population at risk in a first stage and after classification according to the exposure of interest, subsamples are selected for purposes of covariable ascertainment in a second stage. By balancing the numbers of exposed and nonexposed for whom covariance information is ascertained and by modifying the logistic regression analysis, by this procedure some efficiency may be gained over the single-stage approach. This is especially true if exposure is rare and relative risks associated with the covariables are large. The parameter vectors of the quantal response function are estimated by the extension of the conditional maximum likelihood method developed elsewhere.




Cited in
(only showing first 100 items - show all)








This page was built for publication: Logistic regression for two-stage case-control data

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3774797)