I performed propensity score matching and thus generated a quasi-randomized study.

Now I got a risk for Akute Kidney Injury of

Drug1: 15.6 % Drug 2: 11.2 % Sample size = 4711 control + 4711 treated

I also tested for renal replacement therapy and got Drug 1: 6.6 % Drug 2: 6.2 %

and mortality Drug 1: 6.5 % Drug 2: 6.3 %

Now intuitively, the risk difference in renal replacement and mortality is NOT statistically significant, but the risk for AKI is.

My question is this: Which statistical test should I use to determine if the difference is statistically significant? Would a simple t.test suffice considering I calculated the risks using the mean()-function, but the risk is not "really" a mean on a ratio scale?



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.