1:Exploring data and unadjusted (univariate) effect estimates

The current simple logistic regression modeling approach being used by AHRQ in the risk-adjusted model fitting assumes that all patient responses are independent and identically distributed. However, it is likely that responses of patients from within the same hospital may be correlated, even after adjusting for the effects of age, gender ... ................
................