The Cochran-Armitage trend test is used to evaluate the association between a binary outcome (e.g., disease or no disease) and a categorical variable with ordered levels (e.g., age groups).
Note that this test can incoporate any weighting scheme that might be relevant to the question at hand, much like contrasts in regression or ANOVA settings (e.g., to assess the existence of a linear or simple monotonic relationships). It is widely used in clinical studies, as well as in genetic epidemiology (e.g., allelic dosage vs. dominant or recessive models of copy number variations or single nucleotide polymorphisms).
This test is different from the Cochran-Mantel-Haenszel chi-squared test where the cross-classification is performed between a binary outcome and a categorical variable used for stratification purpose (e.g., recruitment center).
