Volume-Outcome in Cancer Surgery: Why has the Data Not Affected Policy Change?

Excerpt

It has been over 30 years since the report by Dr. Harold Luft and colleagues in the New England Journal of Medicine demonstrating an inverse relationship between hospital volume and mortality for selected surgical procedures.1 Almost 20 years later the Leapfrog Group, a consortium of large corporations and public agencies that purchase health care, suggested minimum volume standards for five high-risk surgical procedures as part of a broader, value-based purchasing initiative.2 However, use of volume as a quality metric remains controversial, as it is not a direct indicator of quality of care, but is often associated with quality and can be easily calculated. Over several decades, countless publications have strengthened the reproducible association of higher hospital volume and improved outcome (operative mortality), particularly with regard to pancreatic surgery. It is inescapable that some minimum experience (patient volume) is necessary for both the acquisition and maintenance of surgical skill and perhaps of equal or greater importance, the development of multidisciplinary teams which enhance all aspects of patient evaluation and treatment. Multidisciplinary teams likely impact multiple fundamental factors which affect outcomes including patient selection, perioperative treatment, and postoperative care. Not surprisingly, the benefit of high volume centers extends beyond the postoperative period and is associated with improved long-term patient survival.3 Patient volume and the clinical experience that it generates combined with outcome assessment (Hawthorne effect) makes for improved results. The published data in support of a volume-outcome relationship has resulted in some degree of regionalization of care in other countries,4,5 but a general lack of progress in the United States where a host of explanations have been brought forth to explain the lack of progress: patients may not want to travel or cannot afford the cost of travel; loss of experience for physicians at low volume hospitals (a concern of possibly greater consequence when dealing with patients who may require emergent surgery); financial implications of the possible loss of pancreatic cancer patients—likely not a significant concern for the surgery-related revenue (at low volume institutions), but to the extent that all pancreatic disease may leave (especially if lost from the hospital system), the downstream revenue for the system including infusion, pharmacy, and diagnostic imaging may be significant; and finally, the opinion that general surgeons are well trained in pancreatic and upper abdominal surgery and referral to specialty programs is unnecessary. In this issue of the Annals of Surgical Oncology, Drs. Swanson and colleagues report findings from the National Cancer Data Base, again supporting the volume-outcome (mortality) relationship for major pancreatic resection and they also find that 90-day mortality is twice that of 30-day mortality, even at high volume hospitals.6 The authors conclude that hospitals should be aware of their annual volume, the mortality rates at 30 and 90 days, and be benchmarked to high volume hospitals. The manuscript raises important questions: What other outcomes should be measured beyond 30-day mortality? With the excess of data supporting the volume-outcome relationship, what barriers exist which prevent regionalization of care and how can we move the needle in the direction of policy change? …