Stratification Issues with Binary Endpoints

  • Devan V. MehrotraEmail author


This note addresses four interrelated issues for a stratified comparative trial with a binary endpoint: 1. How to define the true overall treatment effect parameter, 2. How the strata should be weighted when conducting inference and estimation involving the overall treatment effect, 3. How to (and how not to) test for a treatment by stratum (T × S) interaction, and 4. When, why, and how the outcome of the T × S test should influence the weights assigned to each stratum. Numerical examples are provided to reinforce the key points.

Key Words

Binomial Independent proportions Interaction Scale Weighting 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Mehrotra DV, Railkar R. Minimum risk weights for comparing treatments in stratified binomial trials. Stat Med. 2000;19:811–825.CrossRefGoogle Scholar
  2. 2.
    Radhakrishna S. Combination of results from several 2x2 contingency tables. Biometrics. 1965;21:86–98.CrossRefGoogle Scholar
  3. 3.
    Breslow NE, Day NE. The analysis of case-control studies. In Statistical Methods in Cancer Research, Vol. 1. Lyon, France: International Agency for Research on Cancer; 1980.Google Scholar
  4. 4.
    Gail M, Simon R. Testing for qualitative interactions between treatment effects and patient subsets. Biometrics. 1985;41:361–372.CrossRefGoogle Scholar
  5. 5.
    DerSimonian R, Laird NM. Meta-analysis in clinical trials. Control Clin Trials. 1986;7:177–188.CrossRefGoogle Scholar
  6. 6.
    Tarone RE. Homogeneity score tests with nuisance parameters. Commun Stat, Series A. 1988;17:1549–1556.CrossRefGoogle Scholar
  7. 7.
    Gart JJ, Nam J. Approximate interval estimation of the difference in binomial parameters: correction for skewness and extension to multiple tables. Biometrics. 1990;46:637–643.CrossRefGoogle Scholar
  8. 8.
    Lipsitz SR, Dear KBG, Laird NM, Molenberghs G. Tests for homogeneity of the risk difference when data are sparse. Biometrics. 1998;54:148–160.CrossRefGoogle Scholar
  9. 9.
    Lui KJ, Kelly C. A revisit on tests for homogeneity of the risk difference. Biometrics. 2000;56:309–315.CrossRefGoogle Scholar
  10. 10.
    Brown MB, Forsythe AB. The small sample behavior of some statistics which test the equality of several means. Technometrics. 1974;16:129–132.CrossRefGoogle Scholar
  11. 11.
    Mehrotra DV. Improving the Brown-Forsythe solution to the Generalized Behrens-Fisher problem. Commun Stat, Series B. 1997;23:1139–1145.CrossRefGoogle Scholar
  12. 12.
    Senn S. Statistical Issues in Drug Development. Chichester: Wiley; 1997.Google Scholar
  13. 13.
    Fleiss JL. Analysis of data from multicenter trials. Control Clin Trials. 1986;10:237–243.CrossRefGoogle Scholar
  14. 14.
    Pocock SJ. Clinical Trials: A Practical Approach. New York, NY: Wiley; 1983.Google Scholar

Copyright information

© Drug Information Association, Inc 2001

Authors and Affiliations

  1. 1.Scientific Staff, Clinical Biostatistics, UN-A102Merck Research LaboratoriesBlue BellUSA

Personalised recommendations