Taking action on the volume-quality relationship: how long can we hide our heads in the colostomy bag?

There is remarkably consistent evidence that the more experience doctors or health care systems have with a procedure, the better the results. At last count, 123 of 128 published studies showed some evidence of a “volume–quality” relationship (1,2). For oncology, the evidence is mostly consistent and often provides striking examples of markedly better outcomes with higher volume (3). There is even some evidence that outcomes can be improved with standardized care and clinical practice guidelines, among other things (4). In this issue of the Journal, Hodgson et al. (5) identified 7257 patients with stage I–III rectal cancer from 1994 through 1997 in the California Cancer Registry to study the relationship of hospital volume with colostomy rates, 30-day mortality, and 2-year survival. The results look like a dose–response curve in chemotherapy: the higher the volume, the better the results, as shown in Table 1. The high-volume hurdle is not very high here—just 20 surgeries or more per year compared with less than seven per year. High volume is about two surgeries per month, and the lowest volume is one every 2 months. Of the 7257 patients treated during this time period, 1621 (22%) had their surgery done at centers (221 different hospitals) that did fewer than seven surgeries per year. There were important differences in processes of care that may explain some of these outcomes. Fewer lymph nodes were evaluated in the lower volume hospitals (six vs. eight lymph nodes, P<.001), which some have suggested is a surrogate for the adequacy of resection, and could lead to under use of adjuvant treatment. There were some minor differences in socioeconomic status among patients at different hospitals, but these were carefully controlled in multivariable analysis, and the conclusions were still valid. Not all differences were attributable to volume alone. The differences in colostomy rates among individual hospitals near the study’s average were actually greater than the differences among the volume categories. As the authors point out, small hospitals (or nonacademic hospitals or whatever category one chooses) may provide optimal care and have superior outcomes. But most do not. The results are the most up-to-date, current reflection of contemporary surgical practice in a large state. They are consistent with the wide and deep body of knowledge that shows a strong relationship between volume and quality, long recognized in other industries. How we react to this convincing body of knowledge is up to us. We have listed some of the possible actions in Table 2. Those who have followed this debate will recognize the uses of all of these options. These important issues do not arise in a vacuum, and the actions taken will have important consequences for patient outcomes, physician livelihoods and incomes, hospital service capabilities, and the communities that support or are supported by these hospitals. It is unlikely that anyone will mount large randomized clinical trials shortly, and there is not likely to be any more or better evidence that that which is in our hands already. That said, what is the appropriate action to take for patients, health care professionals, and payers (i.e., health insurance companies)? Of the options, asking doctors and health care systems to submit data now seems the most logical.