Stochastic curtailing and conditional power in matched case-control studies

Authors

  • Sally Hunsberger,

    1. Biostatistics Research Branch and Field Studies and Biometry Branch, Division of Epidemiology and Clinical Applications, National Heart, Lung and Blood Institute, Bethesda, MD 20892, U.S.A.
    Search for more papers by this author
  • Paul Sorlie,

    1. Biostatistics Research Branch and Field Studies and Biometry Branch, Division of Epidemiology and Clinical Applications, National Heart, Lung and Blood Institute, Bethesda, MD 20892, U.S.A.
    Search for more papers by this author
  • Nancy L. Geller

    1. Biostatistics Research Branch and Field Studies and Biometry Branch, Division of Epidemiology and Clinical Applications, National Heart, Lung and Blood Institute, Bethesda, MD 20892, U.S.A.
    Search for more papers by this author

Abstract

New methodology for early stopping of observational epidemiologic studies is developed. To indicate the likelihood of a positive result at the end of a study, the conditional power (CP) is calculated using an approach similar to that of Halperin et al. Given the current result, the conditional probability of a positive result for the entire study is determined. An expression relating the new approach to an approach given by Lan and Wittes is derived. As an example, the CP is calculated based on information from the first third of a matched case-control study embedded in an epidemiological study.

Ancillary