Interaction and its solution in individual matching case-control study.

OBJECTIVE: To indicate the deficiency of the classical method for analyzing data on individual matching case-control study in consideration of the interaction between the study factor (exposure) and the matching factor, and to find out a proper method for handling this deficiency. METHOD: First, experimental data with 50 pairs of cases and controls were used for strata analysis according to the values of a matching factor to illustrate the possible interaction between a risk factor (exposure) and the matching factor. Second, a detailed procedure was proposed for analyzing such data. RESULTS: Interaction between the study factor and matching factor was demonstrated by using strata analysis and unconditional logistic regression analysis. Therefore the results from the classical analysis for such data might be incorrect. CONCLUSION: Data from individual matching case-control study design should be dealt with strata analysis or multivariate analysis to explore and evaluate the possible interaction between the study factor and matching factor. The conclusion would be valid only after such analysis is conducted.