228-2011: Matching-Adjusted Indirect Comparison Analysis Using Common SAS® 9.2 Procedures

This paper presents a novel matching-adjusted approach to indirectly compare survival estimates for competitive treatment options. Using patient-level data for the treatment arm and summary patient characteristics and survival outcomes for the comparator, matching variables prognostic for survival are chosen. A program involving an extension of a common SAS® 9.2 procedure, PROC SURVEYSELECT, is applied to select 1000 random repeated sub-samples from the original population with the same distribution of matched variables. Using Output Delivery System (ODS) and survival analysis procedures, the survival estimates are computed for each bootstrapped sample and a 95% confidence interval (CI) is inferred around the mean of the sampled survival estimates. These outcome measures may then be compared to those reported for the comparator treatment. In the absence of head-to-head randomized controlled trial (RCT) data, comparative effectiveness of treatment outcomes can be evaluated using a matching-adjusted indirect comparison, which accounts for observed differences between populations and results in an effect of treatment exposure on survival outcomes less likely due to confounders.