Recorded presentation

Charles P. Schade, MD, MPH and Karen L. Hannah, MBA. West Virginia Medical Institute, 3001 Chesterfield Place, Charleston, WV 25304, (304) 346-9864 x 2243, cschade@wvmi.org

Objectives. Increasing screening mammography use in women aged 50-69 is a key Medicare quality improvement objective. This retrospective analysis of Medicare screening mammography claims (1997-2002) in a single state measures the continuity of care among mammography service providers and determines whether there are variations not explained by patient characteristics. It also assess the potential impact of provider-based efforts to improve continuity rates and compliance with mammography recommendations.

Methods. We built a composite data table for each beneficiary, containing information on mammography procedures and residence for every year the beneficiary was in the state’s Medicare population. We identified successive pairs of mammography records. For each pair we determined the time between services and whether the same provider performed both services. We tabulated outcomes for all pairs in the six-year period. For providers with at least 100 claims in three successive baseline years and across the entire 6 year period, we examined continuity rate (number of patients who had a mammogram in a baseline period whose next mammogram was billed by the same provider) and loss rate (number of patients with a mammogram in a baseline period who were still in the denominator but had not had a subsequent claim within two years).

Results. Among 83 mammography providers who accounted for about 50,000 (90% of total mammograms) screening mammography claims per annum, median continuity and loss rates were nearly constant over the study period (65% and 20%, respectively). There was substantial variation among providers with more than twofold differences between 10th and 90th percentiles of both indicators. Women aged 50-67 were significantly more likely to have timely subsequent mammograms from the same provider than older or younger women, and nonwhite women were 30% more likely than white women to be late for a subsequent mammogram. Women obtaining an initial mammogram in a county not their residence were twice as likely to have their subsequent mammogram from a different provider as those whose initial mammogram was in the same county. Providers with high continuity rates had low loss rates; we estimate that nearly 6,000 additional women would receive timely mammograms each year in our state if all providers maintained continuity rates comparable to those of the best-performing providers.

Conclusions. Beneficiary characteristics influence repeat mammography rates, but there are unexplained variations in continuity among providers. Provider-based efforts to improve continuity could increase overall compliance rates substantially.

Learning Objectives: