LAUSR

D elivery of health care in the United States is the subject of extensive attention and debate, with specific regard to variability in quality of care, and associated rising costs. Michael Porter from Harvard Business School has defined an approach to value-based healthcare delivery, formulated as health outcomes achieved per dollar spent. Porter’s strategic agenda is built around 6 key tenets: first to organize care around patient medical conditions, second to measure outcomes and costs for every patient, third to move to bundled payments for care cycles, fourth to integrate care delivery systems, fifth to expand geographical reach, and sixth to build an enabling integrated information technology platform. In essence, care should be delivered by dedicated, multidisciplinary teams, for the full cycle of care (ie, outpatient, inpatient, and rehabilitative), with a common measurement platform and joint accountability for outcomes and costs. The extensive work of Birkmeyer et al regarding volume-outcome relationships, based upon provider and hospital volume status, underpins this notion, whereby morbidity and mortality rates are lower when clinicians and/or health facilities have greater caseloads. These powerful data have sought to drive forward the amalgamation of care provision, especially for complex and infrequently performed surgical cases such as esophageal, pancreatic, or lung resection. In May, 2015, leaders at Dartmouth-Hitchcock, Johns Hopkins, and University of Michigan health systems publicly announced a ‘‘Take the Volume Pledge’’ to restrict the performance of 10 surgical procedures to hospitals and surgeons that perform more than a given minimum number. Hospital and surgeon volume were defined by case, for example, 20 and 5 for esophagectomy, 20 and 10 for mitral valve repair, and 40 and 20 for bariatric surgery, respectively. Whereas the Pledge has potential to improve clinical care and outcomes, its acceptance has been controversial and highly emotive. In 2004, Urbach and Baxter asked an intriguing question— does it matter what a hospital is high volume for? Like others, they reported 30-day mortality for esophageal, pancreatic, or major lung resection, and repair of an unruptured abdominal aortic aneurysm, to be inversely related to hospital volume. Interestingly, 30-day mortality for these procedures was also related to that of most of the other procedures they considered too, such that the inverse association between high procedure volume and mortality within 30 days was not specific to the volume of the procedure being studied. Funk et al explored this further, to define system characteristics that were indicative of better outcomes at low-volume hospitals for esophagectomy. Low-volume hospitals with 3 out of 5 system characteristics (ie, high nurse ratios, lung transplantation services, complex medical oncology services, bariatric surgery services, and positron emission tomography scanners) had lower mortality rates than similarly low-volume hospitals with none of these characteristics, and approached the mortality rates of medium-volume hospitals. There is considerable debate regarding which measures can and should be apportioned to surgical quality. In this issue of Annals of Surgery, Fry et al have undertaken a retrospective review of Medicare data with respect to the relationship between reduction in surgical complications for 37,329 patients undergoing bariatric surgery at 562 hospitals, and concomitant costs, over a 9-year period. Their intent was to determine the impact of reducing serious postoperative complications on Medicare payments, comparing data from 2005 to 2006 and 2013 to 2014, to align with quality improvement in bariatric surgery. Hospitals were ranked into quintiles based upon absolute change in risk and reliability-adjusted 30-day rates of serious complications, and corresponding changes calculated in average price standardized payments for each quintile of hospitals. Serious complications were those that could be included into 1 of 12 categories (eg, shock, anastomotic leak, pulmonary, genitourinary, hemorrhage, etc), occurring within 30 days of index operation, and led to a procedure-specific extended length of stay. In 2005 to 2006, the risk-adjusted serious complication rate varied from 3.8% in the top