Two studies from Scotland address the effects of referral to specialty centers on processes and outcomes. Clarke et al. (1995) determined that 92 percent of testicular cancer was referred to specialty centers between 1983 and 1990, and that the quality of care at these specialty centers varied. Harding et al. (1993) performed a population-based audit of the management of 440 men diagnosed in 1975-1989 with non-seminomatous germ cell tumors (NSGCT) in western Scotland. All but 11 patients were treated at tertiary referral centers; 235 were treated at a single unit (unit 1) and 194 at four other units (units 2-5). Independent prognostic factors for NSGCT survival were extent of tumor at diagnosis, five-year period of diagnosis (from 1975-1979 to 1985-1989), and treatment unit (unit 1 versus units 2-5). Unit 1 had the best survival rates, had treated the most patients overall (53 percent), and had treated the majority (70 percent) with the worst prognosis (e.g., metastatic disease). Receipt of care according to the nationally agreed-upon protocol treatment varied: 97 percent at unit 1 versus 61 percent at units 2-5. After adjusting for known prognostic factors and limiting the analysis to those treated according to protocol, the relative risk of death outside unit 1 was 2.8 (C.I. 1.5-5.2). The authors conclude, "These findings suggest that centralization of treatment for NSGCT improves outcome; the benefit seems to be additional to any advantage resulting from protocol treatment."
Aass et al. (1991) noted a similar trend, although not as dramatic. They studied 193 patients with metastatic testicular cancer treated at 14 Swedish or Norwegian centers between 1981 and 1986 who entered a clinical trial. If all care had been given at the lead institution, which treated 46 percent of cases, the chance of dying after controlling for known prognostic factors would have been reduced by 28 percent. A limitation of this study is the relatively small sample size.
Junor et al. (1994) reported one of the few studies addressing the benefits of multidisciplinary clinical care. In 1987, 533 cases of ovarian cancer were diagnosed and 479 records were available for audit in Scotland; 27 percent of cases were referred postoperatively to a multidisciplinary clinic. After adjusting for clinical factors and the use of platinum chemotherapy (about 50 percent of patients younger than 65 received chemotherapy), the relative risk of death was 0.73 (27 percent reduction) if patients received care at the multidisciplinary clinic.
Intuition would dictate that being cared for by a physician who has extensive training and experience in cancer care would improve outcomes. Only one U.S. study and several studies conducted in Great Britain or Canada have assessed the relationship between specialization of individual physicians and processes or outcomes. Specialization is defined in different ways: receipt of specialty training, professional identification as a specialist, or university affiliation.
Physician specialty affected the processes and outcomes of care of women with ovarian cancer according to a study by Nguyen and colleagues (1993), conducted as part of the U.S. National Survey of Ovarian Carcinoma. Detailed data were requested on 25 consecutive patients from 1,230 hospitals with cancer programs across the United States from 1983 and 1988. A total of 904 hospitals provided data on 5,156 patients for 1983 and 7,160 patients for 1988. About 96 percent of patients had exploratory surgery as their initial management. The breakdown of physician's specialty was 19 percent general surgeons, 25 percent gynecologic oncologists, 43 percent