OBJECTIVES

The evidence-based medicine pyramid considers systematic reviews as a lens of appraisal for other studies. But does this lens always provide the same image? If not, what can affect the conclusions of systematic reviews? In the set of three reviews, we analyzed the determinants of benefits, harms, cost-effectiveness, and uptake of screening mammography.

METHODS

We systematically searched Medline, Embase, Scopus, and Cochrane databases and supplementary sources, duplicating the search, selection, and evaluation processes and validating the data extraction. The quality of the reviews was assessed by the Assessing the Methodological Quality of Systematic Reviews checklist.

RESULTS

Screening of 9,976 (for benefits/harms), 7,768 (for cost-effectiveness) and 1,893 (for uptake) abstracts resulted to inclusion of 58, 30, and 29 reviews respectively. There was no consistency in the reviews’ conclusions on most of the main outcomes (benefits, harms, cost-effectiveness, and uptake rate), which was not related to quality of the reviews or publication date. Higher level of primary evidence tend to report lower benefits (15–25% of breast cancer mortality reduction) and harms (1–12% for overdiagnosis) related to mammography. No clear determinants of benefits and harms of mammography screening were identified. The determinants of cost-effectiveness are related to the geographic region (West vs Asia), national income level (high vs low income countries), screening age (50-69 years-old vs younger and older population), biennial compared to annual screening intervals, digital versus field mammography, and double versus single-reading using computer aided detection. Simple low-cost interventions, such as invitation letters, basic information on screening, and multiple reminders, could improve screening uptake. Previous positive experience with mammography screening would influence the repeated participation in screening programs.

CONCLUSIONS

Reviews, even of the high quality, bias the conclusions by personal opinion of authors. They still may be useful for decision makers to optimize the national programs.