Objectives: To compare reporting outcomes of radiographic complications conducted by an independent review board and the responsible on-site study personnel in a multicenter study about locking plate fixation of proximal humeral fractures.
Design: Prospective, multicenter study; setting: 9 level I trauma centers.
Patients: One hundred fifty patients (age 50-90) with a radiographically confirmed displaced proximal humeral fracture fixed with a locking plate were included in the study.
Intervention: All radiographic data were reevaluated by an independent review board according to predefined criteria.
Main outcome measurements: Differences in outcomes between the review board and the on-site assessment were analyzed with a paired t test. Interrater agreements between the central review board and on-site assessments were estimated by means of kappa statistics.
Results: The review board revealed significantly more radiographic complications than the on-site assessment (P = 0.006), except for the complication "head necrosis." The interrater agreement was slight to moderate in all calculated categories.
Conclusions: Implementation of a complication review board using predefined criteria is recommended for clinical studies to prevent underreporting of radiographic complications by on-site assessment.