cent) between 1984 and 1995 to the increase in share of discharges at the high-volume provider, Johns Hopkins. During this period, Johns Hopkins increased its yearly share of pancreaticoduodenectomies from 21 percent to 59 percent.
Feuer et al. (1994) retrospectively compared survival of 133 patients with metastatic testicular cancer participating in a clinical trial at a large cancer center in New York (Memorial Sloan-Kettering Cancer Center) to the survival of 172 patients cared for at other hospitals identified from five SEER registries in 1978-1984. Although 89 percent of the SEER cases received chemotherapy and 95 percent of these used cisplatin regimens, the three-year survival was markedly better at the cancer center. For patients with minimal to moderate extent of disease the benefit was more striking (95 versus 73 percent three-year survival rate) than for advanced cases (52 versus 40 percent). The authors speculate that the differences could have been due to many factors including chemotherapy regimen, dose intensity, or institutional factors. Limitations of this study are its small sample size, the comparison of clinical trial participants with population controls, and inclusion of only one specialized center.
Radiation treatment planning and delivery for rectal and sigmoid cancers appeared to be better at academic than nonacademic hospitals according to a Patterns of Care study (Kline et al., 1997). The care received by 408 patients in 1989-1990 from 73 randomly selected facilities (21 academic, 24 hospital based, and 24 free standing) was audited according to consensus guidelines.
Davis et al. (1987) contrasted the survival of 3,607 Hodgkin's disease (HD) patients diagnosed in 1977-1982 and registered by SEER (community care) with that of 2,278 HD patients treated at one of 21 comprehensive cancer (university or referral) centers. Modest differences in patient age, histologic pattern, and frequency of Stage II disease between locations were seen and adjusted for in the comparisons. The mortality rate among SEER patients was higher (relative risk 1.5; 95 percent C.I. 1.3-1.7) than among those treated at comprehensive centers. The survival difference was consistently seen for all stages, histologic types, and patient ages. A strength of this study is the large number of cases of HD represented nationally. Underlying differences in the comparison groups may not have been accounted for in the analysis (e.g., comorbidity).
In a study by Lee-Feldstein et al. (1994), hospital teaching status had no effect on five-year survival of women with localized or regional breast cancer diagnosed 1988 and 1990. Three large studies suggest that relative to other hospitals, teaching hospitals are more likely to use breast conserving surgery (Ballard-Barbash and Potosky, 1996; Johantgen et al., 1995; Nattinger et al., 1992).
Several European and Canadian studies have assessed the effects of specialization on outcomes. In England, Basnett et al. (1992) compared survival of 999 women with breast cancer treated initially between 1982 and 1986 in two settings, an urban teaching hospital and a rural nonteaching one, both of which had radiation and chemotherapy capabilities. Numerous differences in process of care were seen. After adjusting for age and stage, the adjusted risks of relapse or death were worse (1.45 and 1.74, respectively) for the nonteaching hospitals. Multivariate analyses to explore the reasons for these differences were not reported. Chemotherapy was used in only 5-8 percent of women in both settings, but more often at teaching hospitals. It is difficult to attribute differences reported in this study to the teaching status of the hospital. There is also a rural-urban difference between hospitals. Furthermore, differences in staging by hospital type could confound the observed relationship.