Our study examined whether there would be a change in prescription rates following an intervention. We had 236 patients. 109 before intervention & 127 after the intervention. We were interested in subgroups. Because we were interested in rates, we used simple difference in population statistics and generated p-values/CI's for each of our subgroups. These formulas take n into account, but I read that there may be certain assumptions about the size of the sample and the appropriateness of the p-value. I don't know what those assumptions are. Any insight into a) if the p-value is appropriate given our small subgroups, b) if a one-tailed or two-tailed z test was appropriate. Thanks in advance.