Objective: Follow-up skeletal surveys have been shown to improve the rate of fracture detection in suspected cases of non-accidental trauma (NAT). As these studies are performed in a particularly radiosensitive population, it is important to evaluate if all of the (approximately 20) radiographs obtained at repeat skeletal survey are clinically useful. Our goal was to evaluate if certain radiographs can be excluded at follow-up skeletal survey without compromising the clinical efficacy. Methods: This retrospective study included 22 cases of suspected NAT (average age 3.8 months, range 0.7-15 months) in which patients received both initial and follow-up bone surveys. The follow-up survey was performed an average of 16.7 days (range 11-29 days) after the initial survey. Radiographs were reviewed by 2 pediatric radiologists, with discrepancies resolved by consensus. In addition, we combined our data with data from all known previously published reports of follow-up skeletal surveys for NAT for meta-analysis. Results: A total of 36 fractures were found on the initial bone survey in 16/22 patients (73%). Six patients had no fractures detected at initial survey. Follow-up bone surveys demonstrated an additional 3 fractures (2 extremities and 1 rib) in 3/22 cases (14%); 1 was in a patient whose initial survey was negative. No additional fractures in the skull, spine, pelvis, feet, or hands were detected in any case. In combination with patients reported in the literature (194 patients total) no new fracture of the skull, spine, pelvis, or hands was detected at follow-up survey. The skull, spine and pelvis radiographs are the highest dose-exposure studies of the skeletal survey. Conclusion and practice implications: If no injury is detected or suspected in the pelvis, spine, hands, or skull at initial bone survey for suspected NAT, a limited follow-up skeletal survey which excludes the pelvis, lateral spine, hands, and skull should be considered to limit radiation exposure without limiting diagnostic information. (Contains 1 table.)