Specification of the regression function and the assumption
of homoscedasticity of errors were checked by examining
plots of the residuals versus predicted values.28
We used PROC GENMOD in SAS for Windows, Version
9.3 (SAS Institute, Cary, NC) and R programming language
for all analyses. All models were estimated using generalized
estimating equations (GEE) to account for clustering of
performance measures at the nurse level.29 All tests were
two-sided and a p value of ≤ 0.05 was defined as
statistically significant; we did not adjust for multiple
comparisons.