PubMed2000
PURPOSE: To conduct a comprehensive review of the health services literature to search for evidence that hospital or physician volume or specialty affects the outcome of cancer care. METHODS: We reviewed the 1988 to 1999 MEDLINE literature that considered the hypothesis that higher volume or specialization equals better outcome in processes or outcomes of cancer treatments. RESULTS: An extensive, consistent literature that supported a volume-outcome relationship was found for cancers treated with technologically complex surgical procedures, eg, most intra-abdominal and lung cancers. These studies predominantly measured in-hospital or 30-day mortality and used the hospital as the unit of analysis. For cancer primarily treated with low-risk surgery, there were fewer studies. An association with hospital and surgeon volume in colon cancer varied with the volume threshold. For breast cancer, British studies found that physician specialty and volume were associated with improved long-term outcomes, and the single American report showed an association between hospital volume of initial surgery and better 5-year survival. Studies of nonsurgical cancers, principally lymphomas and testicular cancer, were few but consistently showed better long-term outcomes associated with larger hospital volume or specialty focus. Studies in recurrent or metastatic cancer were absent. Across studies, the absolute benefit from care at high-volume centers exceeds the benefit from break-through treatments. CONCLUSION: Although these reports are all retrospective, rely on registries with dated data, rarely have predefined hypotheses, and may have publication and self-interest biases, most support a positive volume-outcome relationship in initial cancer treatment. Given the public fear of cancer, its well-defined first identification, and the tumor-node-metastasis taxonomy, actual cancer care should and can be prospectively measured, assessed, and benchmarked. The literature suggests that, for all forms of cancer, efforts to concentrate its initial care would be appropriate.