Recall Laterality and Bilaterality: Possible New Screening Mammography Quality Metrics

Abstract

Purpose: Current screening mammography quality metrics are important and helpful, but do not address all quality concerns. An individual screening mammography reader may be systematically insensitive to findings present in the breast of one side, laterality bias, evidenced by left versus right difference in advised immediate recalls. Current metrics are not designed to detect laterality bias. Whether a reader exhibits laterality bias, or what an appropriate ratio/range of bilateral versus unilateral recalls should be, have never been discussed in medical literature. Methods: As a trainee quality project, five attending (‘consultant’ in Europe) radiologists’ screening mammography reports over 2 years at an academically affiliated, public hospital were tallied with regard to laterality of recommended recall, and with respect to unilateral versus bilateral recalls advised. The chi-square (χ²) statistic was applied to reports advising unilateral recall. Findings: No group laterality bias was discovered. One radiologist (the most experienced) evidenced a consistent laterality bias over 2 years (p=0.07) against left-breast findings. Of reports recommending recall, the radiologists’ single-year range for recall regarding both breasts was 10.2–23.3%; for both years combined, the individual radiologists ranged from 13.6–17.9%. The group, 2-year mean recommending bilateral recall was 16.5%. Conclusion: A radiologist may exhibit laterality bias, favoring detection of findings in one breast over the other, a concern never before considered. Audit to discern such bias leads simultaneously to assessment of bilateral recall bias. Possible causes of these biases are discussed, and research regarding them as possible quality metrics is encouraged.