Figure 2. Primary and Per-Protocol Analyses in a Study of the Effect of High-Dose Baclofen on Agitation-Related Events Among Patients With Unhealthy Alcohol Use Receiving Mechanical Ventilation.

Trend lines were truncated when observations fell below 20%. Between-group differences were tested using Fine-Gray regression considering death as competing event for agitation. When baclofen was compared with placebo, the hazard ratio (HR) for agitation was 0.62 ([95% CI, 0.40-0.96]; P = .03) in the primary analysis and 0.36 ([95% CI, 0.17-0.79]; P = .01) in the per-protocol analysis. A hazard ratio less than 1 indicates that the probability of agitation was lower in the baclofen group at all times.