Statistical Methods
Sensitivity analyses restricted the data to first hospitalisations and used mixed-effects logistic regression with patient-level random intercepts. Mixed-effects logistic regression produced results consistent with the main GEE analysis bu…
3 sources - 15 claims
Sensitivity analyses restricted the data to first hospitalisations and used mixed-effects logistic regression with patient-level random intercepts. Mixed-effects logistic regression produced results consistent with the main GEE analysis but with wider confidence intervals. Stratified discharge-setting analyses assessed effect modification, and Wald tests evaluated interaction p-values.