Application of Conditional Logistic Regression method in a prospective effectiveness comparative study among patients with Acute Ischemic Stroke

To compare the optimal and non-optimal intervention combinations among patients with Acute Ischemic Stroke (AIS) under matched states by applying Conditional Logistic Regression (CLR) method. Methods: Hospitalized patients with AIS managed by an established clinical pathway were enrolled. Patients with paired states on Day 1 or Day 7 were matched as N:K pairs, while these patients were assigned as two groups according to whether they receive an optimal combination of treatments or not. Results: 299 states at Day 1 and 275 states at Day 7 were collected, only those states with at least one of the patients were treated with optimal combination of treatments were included, resulting in 6 states at Day 1, as well as 7 states at Day 7. Two states at Day 1 were found to be significantly more effective than the non-optimal group. Conclusion: CLR is a valuable analytic technique in effectiveness comparative researches.