Bundled Payment Programs and Changes in Practice Patterns and Episode Spending in Major Gastrointestinal Surgery.

Objective: To evaluate the association between enrollment in the Bundled Payments for Care Improvement -Advanced (BPCI-A) program and changes in utilization of minimally invasive surgery and 90-day episode spending for patients undergoing major gastrointestinal surgery.

Study setting and design: We compared hospitals that voluntarily enrolled in BPCI-A to control hospitals that did not participate. We used entropy balancing to reweight controls to match the BPCI-A cohort based on observable patient and hospital characteristics. We then used a difference-in-differences approach to estimate the association between surgical approach and 90-day episode payments.

Data sources and analytic sample: We used Medicare claims and American Hospital Association data between 2013 and 2021 to evaluate whether hospital enrollment in the BPCI-A program was associated with changes in 90-day episode spending and utilization of minimally invasive surgical approaches. Using entropy balancing, we reweighted the control group to achieve covariate balance with beneficiaries who obtained care at BPCI-A program hospitals. We performed a difference-in-differences analysis using multivariable linear and generalized linear models, adjusting for patient demographics, comorbidities, and hospital characteristics, with standard errors clustered at the hospital-year level to evaluate these outcomes.

Principal findings: Changes in 90-day episode payments at BPCI-A program hospitals versus non-program hospitals were not significantly different (-$172, 95% CI: -$1104 to $760). In comparing trends at BPCI-A program and control hospitals, we identified no significant differences in utilization trends for minimally invasive surgical approaches (relative risk difference: -0.003, 95% CI: -0.10 to 0.04). The similarity in utilization trends between BPCI-A program and control hospitals was observed in the context of increasing overall utilization of MIS approaches from 40.3 to 38.4 to 43.9 to 42.9 during the study period, respectively.

Conclusions: We found no evidence that hospitals participating in BPCI-A's major bowel surgery episodes led to differences in episode spending or utilization of minimally invasive surgical approaches.