John D. Keen MD, MBA, Presenter: Nothing to Disclose

James Edward Keen DVM, Abstract Co-Author: Nothing to Disclose

The recommended recall rate of 10% for screening mammography has no quantitative justification. We wanted to find the age dependence of the optimal decision thresholds for a single baseline and subsequent screening mammogram, and to determine if current recall rates are reasonable.

Receiver operating characteristic (ROC) analysis predicts that the optimal false positive fraction (FPF) for a screening exam is a function of disease prevalence. We obtained age dependent breast cancer incidence and prevalence data from a national database. We used a previously developed decision model to calculate the cost of a false positive decision for a single baseline and subsequent mammogram. The false positive fraction (FPF) is close to the recall rate in a screening setting. We used previously developed life expectancy estimates and the life saving rate of mammography at ages 40, 50, and 60 to determine the utility of a true positive decision. We valued years of life saved at $50K/life year (LY) and $100K/LY and applied a discount rate of 3%. We ignored excess potential lawsuit costs for a false negative decision. ROC and average recall performance data were based on Breast Cancer Surveillance Consortium data.

The optimal decision thresholds depend on age and the value assumed for a year of life saved. For baseline mammography at age 50, the optimal FPF/TPF pair is about 7%/79% at $50K/LY and 14%/88% at $100K/LY. For 100K/LY, the FPF decreases to 9% at age 40, and increases to 17% at age 60. The national FPF/TPF/recall average is 14.1/89%/14.7%. For subsequent mammography at age 50, the operating points are 2%/57% at 50K/LY, and 5%/70% at 100K/LY. For $100K/LY the thresholds are 2%/47% at age 40, and 6%/73% at age 60. The national averages are 7.5%/77%/7.8%.

Baseline mammography recall rates for women over 50 are reasonable assuming a high society value per year of life saved. However, for subsequent mammography and lower life year values, recall rates are above optimal, especially for women under 50.

Advocates of higher recall rates underestimate false positive costs, especially in younger women. Economic theory predicts higher income women would prefer a higher recall rate.

Keen, J, Keen, J, Beyond the 10% Rule: How Does the Optimal Recall Rate for a Single Screening Mammogram Vary with a Woman's Age?.  Radiological Society of North America 2007 Scientific Assembly and Annual Meeting, November 25 - November 30, 2007 ,Chicago IL. http://archive.rsna.org/2007/5000833.html