Strategies for Reducing Population Surgical Costs in Medicare: Local Referrals to Low-cost Hospitals.

Abstract

We sought to assess hospital cost variation for elective inpatient surgical procedures within small geographic areas. Previous studies have documented cost variation for inpatient surgical procedures on a national basis, suggesting opportunities for savings. Cost variation within small geographic areas is more relevant to policymakers, providers, and patients, but it has not been studied. Using Medicare payment data, we identified elderly patients undergoing 1 of 7 elective inpatient surgical procedures during 2010-2012. We calculated 30-day surgical episode costs including payments for the index hospitalization, readmission, physician services, and post-acute care. Using hierarchical regression models, we identified hospitals with significantly higher average costs than the least expensive hospitals in their metropolitan statistical areas. The proportion of patients undergoing surgery at the lowest-cost hospitals in their metropolitan statistical areas ranged from 10% for hip replacement to 25% for coronary artery bypass grafting. In contrast, the proportion of patients undergoing surgery at significantly higher-cost hospitals ranged from 5.0% for bariatric surgery to 64% for hip replacement. These high-cost hospitals had higher complication and readmission rates than their lowest-cost peers. Surgery at high-cost hospitals resulted in Medicare expenditures that were $4427 to $10,417 higher than those at the lowest-cost hospitals, increasing episode costs by 25% to 47% per case. Significant excess expenditures are incurred due to care at hospitals with substantially higher average costs than their nearby peers. This finding highlights the potential for substantial savings without the need to refer patients over long distances. Some of the procedures studied may represent appropriate targets for future Medicare bundled payment initiatives.