It is interesting to consider the ways in which the low- and high-use groups differ from each other. Are they different in most respects or only in some? We calculate the Wald statistic to test the null hypothesis that the population parameters of the two groups are equal for each of the following three subsets of covariates: insurance, demographics and health status. For an arbitrary subset, denoted β, the Wald test is based on the quadratic form of the estimated differences between subsets of coefficients, denoted βˆ , for thetwo groups.