The danger of applying uniform clinical policies across populations: the case of breast cancer in American Indians.

Abstract

This study examined the implications of annual screening mammography for cost and mortality in American Indian populations with differing baseline breast cancer rates. A decision tree compared annual screening mammography and screening clinical breast examination with referral for diagnostic mammography when appropriate. The decision tree was constructed to examine the effect of different base-line cancer rates, stage at diagnosis, and stage-specific survival. Outcomes included 5-year relative survival, deaths prevented at 5 years, cost per death prevented, and total costs. The findings suggest that the total cost of breast cancer is 3.6 times higher with the screening mammography program but results in a 27.9% reduction in breast cancer deaths over the first 5 years of the program. Both costs and deaths prevented are sensitive to the incidence of breast cancer in the population and are less favorable in the range of incidence seen in American Indians. The cost and impact of a given strategy for cancer screening vary among communities with different disease incidence, stage at diagnosis, and stage-specific survival, as seen in American Indian populations.