Hi all,
I'd like to compare the risk-adjusted mortality of an observed group to that of a large registry dataset. The risk adjustment creates a score (I'll call it the Awesome score or A-score) that is computed from logistic regression and represents the expected mortality. I can calculate this score for each patient. In order to assess whether the entire cohort of observed folks differs from the expected mortality for that group I'd thought to use z-scores. In this case, however, it's not just asking whether one patient differs, but whether the whole cohort differs. Can I simply average my A-score of all the group and compare it to the registry group with a z-score?
I'd like to compare the risk-adjusted mortality of an observed group to that of a large registry dataset. The risk adjustment creates a score (I'll call it the Awesome score or A-score) that is computed from logistic regression and represents the expected mortality. I can calculate this score for each patient. In order to assess whether the entire cohort of observed folks differs from the expected mortality for that group I'd thought to use z-scores. In this case, however, it's not just asking whether one patient differs, but whether the whole cohort differs. Can I simply average my A-score of all the group and compare it to the registry group with a z-score?
