This presentation is part of: Opening Plenary Session (6)
COULD WE HAVE DONE BETTER? A RETROSPECTIVE COST-EFFECTIVENESS ANALYSIS OF ROUTINE SCREENING MAMMOGRAPHY
Natasha K. Stout, PhD1, Marjorie A. Rosenberg, PhD2, Amy Trentham-Dietz, PhD1, Maureen A. Smith, PhD, MD1, Stephen M. Robinson, PhD3, and Dennis G. Fryback, PhD1. (1) University of Wisconsin, Population Health Sciences, Madison, WI, (2) University of Wisconsin, Actuarial Science and Risk Management, and Biostatistics and Medical Informatics, Madison, WI, (3) University of Wisconsin, Industrial Engineering, Madison, WI
Purpose: Screening mammography is recommended every 1-2 years for women over age 40 and surveys indicate that over 70% of women now participate in routine screening. Few studies have examined the societal impact on cost and quality adjusted life years (QALYs) of screening practices over the past decade. To inform current and future practice we ask retrospectively if we could have done better.
Methods: Using a validated discrete-event simulation model of the epidemiology of breast cancer, the total costs and health effects of 60 hypothetical screening scenarios implemented for 1990-2000 were estimated. This analysis considered the effects for all women age 40 and older in the state of Wisconsin for the 10 year study time period. Screening scenarios varied by starting and ending age as well as the frequency of mammograms. Screening as it actually occurred during this time period was also included as a scenario. Accounting for the effects associated with screening use and breast cancer treatment, we compared the QALYs accrued and the total costs generated by each screening program. Costs and QALYs were discounted at 3%.
Results: The estimated total cost of screening and treating breast cancer in the state of Wisconsin from 1990-2000 was $3.15 billion. We estimate the total QALYs accrued were 18.2 million. The actual screening use consisted of a mixture of different screening patterns in the population. Scenarios consisting of fixed screening patterns dominated the actual screening patterns. Screening women aged 55-75 every three years accrued a similar number of QALYs at a cost of approximately $600 million less than the actual screening patterns. Alternatively, screening women age 50-75 every two years accrued 8,000 more QALYs at a similar cost compared with the actual screening patterns.
Conclusions: While the estimated total costs and QALYs represent the effects for the state of Wisconsin, the results can be extrapolated to the entire US population using our methods. Our population-based analysis is unique in that it accounts for screening use prior to 1990. By recommending less frequent mammograms and ensuring more women participate, we could have achieved more quality-adjusted life years at a lower cost compared with actual practice. This should lead to reconsideration of screening polices and implementation for the future.
