Fig. 3.

Survival probability according to the use of adjunctive corticosteroid therapy for PJP in the whole study population (N = 158). The upper panel shows the weighted Kaplan–Meier curves for each group obtained by propensity score-based inverse probability of treatment weighting (PS-IPW, see “Methods” for details). In contrast, the risk table (lower panel) shows raw numbers at risk for each group (i.e., not in the weighted population constructed by PS-IPW for statistical analysis). The P value is the one obtained by accelerated failure time modeling (see “Methods”) for the comparison of use of adjunctive corticosteroid therapy versus no use. Time zero was the time at which curative antibiotic treatment for PJP was started. Patients were right censored at Day 180 after the beginning of curative treatment. Note that one patient of the no corticosteroids group could be followed up only until the 178th day