Possible consequences of regionally based bundled payments for diabetic amputations for safety net hospitals in Texas

Objective: Ongoing health reform in the United States encourages quality-based reimbursement methods such as bundled payments for surgery. The effect of such changes on high-risk procedures is unknown, especially at safety net hospitals. This study quantified the burden of diabetes-related amputation and the potential financial effect of bundled payments at safety net hospitals in Texas. Methods: We performed a cross-sectional analysis of diabetic amputation burden and charges using publically available data from Centers for Medicare and Medicaid and the Texas Department of Health from 2008 to 2012. Using hospital referral region (HRR)-level analysis, we categorized the proportion of safety net hospitals within each region as very low (0%-9%), low (10%-20%), average (20%-33%), and high (> 33%) and compared amputation rates across regions using nonparametric tests of trend. We then used charge data to create reimbursement rates based on HRR to estimate financial losses. Results: We identified 51 adult hospitals as safety nets in Texas. Regions varied in the proportion of safety net hospitals from 0% in Victoria to 65% in Harlingen. Among beneficiaries aged > 65, amputation rates correlated to the proportion of safety net hospitals in each region; for example, patients in the lowest quartile of safety net had a yearly rate of 300 amputations per 100,000 beneficiaries, whereas those in the highest quartile had a yearly rate of 472 per 100,000 (P = .007). Charges for diabetic amputation-related admissions varied almost 200-fold, from $5000 to $1.4 million. Using reimbursement based on HRR to estimate a bundled payment, we noted net losses would be higher at safety net vs nonsafety net hospitals ($180 million vs $163million), representing a per-hospital loss of $1.6 million at safety nets vs $700,000 at nonsafety nets (P<.001). Conclusions: Regions with a high proportion of safety net hospitals perform almost half of the diabetic amputations in Texas. Changes to traditional payment models should account for the disproportionate burden of high-risk procedures performed by these hospitals.