Microcomputer programs for complex epidemiologic procedures. II. Test for equality and trend of adjusted rates

The microcomputer program presented in Part I of this series (M. H. Dong, Comput. Biomed. Res. 22, 349 (1989)) is useful for computing adjusted excess rates. That program, however, does not provide a means for a quantitative analysis of the adjusted rates computed for the categories of a risk factor. At times it is crucial for us to determine whether after adjustment a factor exerts a dose-response effect that is statistically significant. Recently a hypothesis test for analyzing such an effect has been reported. The test is based on score statistics obtained directly from a product model assumed for the excess rate. The objective of this second part is to continue to make available a microcomputer program that has been written to implement the computation of these score statistics, which likewise entails a complex and lengthy iterative procedure. In this communication the test is described and the application of the microcomputer program is illustrated.