Understanding Surgeon Performance and Improving Patient Outcomes

Some surgeons have better outcomes than others. Although the implications of this assertion make clinicians uncomfortable, there should be little doubt that it is true. For example, O’Connor et al demonstrated more than a decade ago, a six-fold variation in operative mortality rates among surgeons performing coronary artery bypass surgery in northern New England, despite adjusting for illness severity and other patient characteristics. Variation in performance is related to several surgeon characteristics, including how often they perform a given procedure (volume), subspecialty certification, and the hospital setting in which they operate. Although evidence linking these characteristics to operative mortality is particularly strong, surgeon factors also predict rates of postoperative complications and even cancer outcomes after selected surgical procedures. In this issue of the Journal Of Clinical Oncology, Herr et al add to the literature, suggesting that subspecialty surgeons have better outcomes. In the context of a randomized, cooperative group trial assessing the effectiveness of neoadjuvant chemotherapy, the investigators studied late outcomes in patients undergoing radical cystectomy and the role of surgeon-related factors. In a subgroup analysis, patients who underwent surgery by urologic oncologists had substantially lower rates of local tumor recurrence than those who were operated on by general urologists (6% v 23%, respectively; P .06). They also had higher 5-year survival rates (58% v 48%, respectively; P .053). Although not quite statistically significant, the magnitudes of these differences are clinically compelling. No oncologist would dismiss a new chemotherapy regimen or radiation treatment that produced a 10% absolute gain in long-term survival. One obvious response to such data is evidence-based referral— direct more patients to surgeons or hospitals likely to achieve the best outcomes (eg, subspecialists, highvolume surgeons or hospitals). This could occur simply by providing patients or their referring physicians with more information to guide their decisions about where and by whom to undergo surgery. Although most sites do not currently provide data on surgical treatment for bladder cancer, numerous Internet sites, including http://www. healthgrades.com, provide hospital-specific data about mortality rates and procedures volumes with many other operations. Evidence-based referral can also be leveraged by payers. For example, the Leapfrog Group, a large coalition of public and private employers covering more than 40 million patients, is using a variety of financial and other incentives to steer patients to high-quality hospitals for coronary artery bypass surgery, percutaneous coronary interventions, elective aortic aneurysm repair, esophagectomy, and pancreatic resection. Although such efforts may be the right place to start, evidence-based referral is ultimately limited in its ability to improve surgical outcomes. First, our ability to correctly identify high-quality surgeons or hospitals is suspect. With the notable exception of cardiac surgery, adequately precise, risk-adjusted, and publicly available provider-specific outcome data remain scarce for most procedures. Indirect measures of quality, including surgeon volume or subspecialty certification, may be better than no information at all. However, such measures often poorly predict the performance of individual surgeons or hospitals. Second, even if quality could be reliably assessed, the likelihood of redirecting all patients to the highest quality providers is remote. Unlike Canada’s single-payer, government-run system, the US health care system is fractured and decentralized, such that no payer group or regulatory body has sufficient leverage to implement such far-reaching policies. Given patient preferences, geography, capacity at tertiary care centers, provider financial incentives, and other factors, many patients will undoubtedly continue to receive their care from surgeons or hospitals with suboptimal outcomes. JOURNAL OF CLINICAL ONCOLOGY COMMENTS AND CONTROVERSIES VOLUME 22 NUMBER 14 JULY 15 2004