Table 3

Test statistics for risk adjustment regression models

DiagnosisOutcome
In-hospital mortalityLong staysReadmissions all causesReadmissions same principal diagnosis
AMI = acute myocardial infarction.
c  =  “c” statistic corresponding to the area under the receiver operating characteristics (ROC) curve. Statistic values >0.7 indicate the model discriminates well; value of 0.5 indicates discrimination no better than chance.
A p value of >0.05 indicates the model provides a good fit for the data (Hosmer-Lemeshow goodness of fit test).
Heart failurec = 0.719c = 0.683c = 0.601c = 0.617
p = 0.146p = 0.125p = 0.824p = 0.289
AMIc = 0.793c = 0.808c = 0.645c = 0.567
p = 0.006p = 0.220p = 0.371p = 0.440
Strokec = 0.682c = 0.654c = 0.636c = 0.627
p = 0.075p = 0.393p = 0.817p = 0.759