For example, all participants were prescribed diet, but few were prescribed insulin. Thus, for pill and insulin adherence, the hypotheses tests were considered to be run on sub-populations of those prescribed the treatment in addition to diet, physical activity, and glucose monitoring.