Figure 2.
PS distribution before weighting, after weighting, and after matching. Panel A shows a PS distribution before weighting and shows trend for SCHOLAR-5 to have lower propensities and ZUMA-5 to have higher ones. Patients with near-zero PS were all in SCHOLAR-5, and patients with higher propensities were almost all in ZUMA-5. Panel B shows the results after application of SMR weighting, with comparable distributions. Panel C shows that in the matched analysis, patients with scores that were only present in 1 study (eg, scores >0.80 in ZUMA-5) were removed, leading to a better PS overlap but reduced common support data set.

PS distribution before weighting, after weighting, and after matching. Panel A shows a PS distribution before weighting and shows trend for SCHOLAR-5 to have lower propensities and ZUMA-5 to have higher ones. Patients with near-zero PS were all in SCHOLAR-5, and patients with higher propensities were almost all in ZUMA-5. Panel B shows the results after application of SMR weighting, with comparable distributions. Panel C shows that in the matched analysis, patients with scores that were only present in 1 study (eg, scores >0.80 in ZUMA-5) were removed, leading to a better PS overlap but reduced common support data set.

Close Modal

or Create an Account

Close Modal
Close Modal