The exact reasons for the better outcomes are unclear, according to an article in the current issue of the British Journal of Surgery, but the correlation between high volume and better outcomes, especially in pancreatic, esophageal, gastric and rectal cancers, was consistent across most studies.
"All studies showed either an inverse relationship, of variable magnitude, between provider volume and mortality," according to authors S.D. Killeen and colleagues of the Cork University Hospital and University College in Ireland. "The majority of clinical studies revealed a statistically significant correlation between volume and outcome; no study demonstrated the opposite relationship."
The authors point out that the results make a good argument for centralizing cancer and possibly other treatments to high-volume medical centers. "Present findings support volume-based referral initiatives," they write. "Centralization of most, if not all, oncological procedures now seems appropriate."
The current review included 41 studies from 1984 to 2004. Because studies varied, no exact definition of low volume versus high volume for either providers or hospitals was established. Some studies measured hospital volume only, some provider volume only, and some both. Low-volume hospitals ranged from one case per year to fewer than 22 per year for a single procedure. Surgeons considered low-volume providers had rates varying from two, five, or 13 cases per year.
The measurement of success or failure was fairly was consistent across studies--thirty-day or inpatient mortality was the primary outcome used. However, authors note that the variability of criteria made interpreting the results difficult. "A note of caution is advisable before advocating policy changes based solely on currently available evidence," they write.
Teasing out whether the higher volume at the hospital level or the physician level provides better outcomes is difficult, according to the authors.
Previous studies have shown similar results in patients treated for heart attacks, breast cancer and other diseases -- those hospitals that treat a higher volume of chronically ill patients have lower mortality rates. Again, the reasons for the disparities in outcomes are unclear. Although some researchers postulate that "practice make perfect," other experts believe that physicians tend to refer their sickest patients to medical centers that have an outstanding reputation.
Also, so-called system-wide factors such as services and organization that enable larger medical centers to move a high-volume of patients may affect outcome. These factors include well-organized clinical teams, the use of recommended practice guidelines and sophisticated, high-tech equipment.
Not all experts agree that centralizing services to larger medical centers is the best policy. "Volume outcome of certain procedures don't always hold up when comparing one hospital or one surgeon with one other hospital or one other surgeon who treat higher volumes," says Nancy Foster, vice president for quality and patient policy at the American Hospital Association in Washington. "We need to consider patient needs such as proximity to home, social and family support."
Foster adds that rather than centralizing treatment, future studies should focus on determining which processes predict better outcomes for patients. "If we can identify those factors, we could spread the word and support smaller hospitals and physicians that serve smaller populations," she said.
However, conclude the authors, "… it appears that high-volume providers are associated with a significantly better outcome, at least after complex surgery for cancer. These findings support the centralization of oncology services."
By Lise Stevens, Contributing Writer
Health Behavior News Service
FOR MORE INFORMATION
Health Behavior News Service: (202) 387-2829 or www.hbns.org.
Interviews: H.P. Redmond at redmondhp@shb.ie.
British Journal of Surgery: Visit www.interscience.wiley.com/bjs
Killeen SD, et al. Provider volume and outcomes for oncological procedures. British Journal of Surgery, 92:389–402, 2005.
Journal
BJS (British Journal of Surgery)