Number Needed to Screen: Appropriate Use of This New Basis for Screening Mammography Guidelines

Abstract

AJR:198, May 2012 trials were conducted in the United States and Europe. The trials, their dates, and patient ages at entry were as follows: the Health Insurance Plan of Greater New York trial, 1963–1969, ages 40–64 years; the Malmo, Sweden, trial, 1976–1986, ages 45–69 years; the Swedish Two-County trial, 1979–1988, ages 40–74 years; the Edinburgh, Scotland, trial, 1981– 1985, ages 45–64 years; and the Gothenburg, Sweden, trial, 1982–1988, ages 40–59 years. Relative reduction in breast cancer mortality differed among these trials and ranged from 20–32% [2]. None of these randomized controlled trials (RCTs) were designed to separately assess the effect of screening women ages 40–49 years. Early subset analyses of this younger age group found no statistically significant benefit. When mortality reduction finally became apparent on longer-term follow-up, it was at first smaller and appeared later after the screening process than that for older women. The explanations for these initially disparate results for younger women included their lower breast cancer incidence rate, insufficient numbers of younger women in the trials, their faster breast cancer growth rates requiring shorter screening intervals, and their denser breasts making mammographic detection more difficult. Adequate assessment of the benefit for younger women initially required pooling results from multiple trials in a meta-analysis and longer-term follow-up. Not until 1996 did an 11to 15-year follow-up of five Swedish RCTs find a statistically significant 29% mortality reduction for women ages 40–49 years, identical in magnitude to the benefit for screening older women [2]. Subsequently, several individual trials reported statistically significant breast cancer mortality reductions for these younger women: 24% in a reanalysis of the Health Insurance Plan of Greater New York data; 36% for the Malmo, Sweden, trial; and 45% for the Gothenburg, Sweden, trial Number Needed to Screen: Appropriate Use of This New Basis for Screening Mammography Guidelines