Age does not affect reliability of the Radiographic Union Score for HUmeral fractures score

Background: A key challenge of humeral shaft fracture management is nonunion after nonsurgical treatment. The Radiographic Union Score for HUmeral fractures (RUSHU) assesses early fracture healing; however, the effect of patient age on its reliability remains unclear. Methods: In this validation study, 143 adult patients with nonsurgically managed humeral shaft fracture were included (nonelderly 18-64 years, n = 71; elderly ≥65 years, n = 72). Six raters with varying experience (medical students, residents, consultants) independently scored 6-week radiographs using RUSHU. Intra-rater and inter-rater reliability were evaluated for total scores, a binary RUSHU (<8 vs. ≥ 8), and individual cortices using weighted Cohen's kappa, Fleiss' kappa, and intraclass correlation coefficients. Results: Mean RUSHU scores were similar between nonelderly (6.9 ± 2.0) and elderly patients (7.0 ± 1.9; P = .95). Intra-rater reliability was excellent for total RUSHU scores (κ = 0.92, 95% confidence interval (CI): 0.88-0.95) and moderate to excellent for the binary RUSHU (κ = 0.82, 95% CI: 0.72-0.91), with no age-related differences. Inter-rater reliability for total RUSHU was fair by Fleiss' kappa (κ = 0.24, 95% CI: 0.22-0.26) and moderate to good by intraclass correlation coefficient (0.69, 95% CI: 0.52-0.77). Binary scoring improved agreement (κ = 0.59-0.61). Inter-rater reliability also increased significantly with Cohen's kappa across rater pairs (κ = 0.45-0.82), experience levels (κ = 0.66-0.81), and within individual cortices (κ = 0.33-0.54). No significant differences were observed between age groups by experience level or individual cortices. Conclusion: Age did not influence RUSHU reliability. Intra-rater reliability was moderate to excellent, and inter-rater reliability was fair to excellent depending on the statistical method utilized. Agreement improved with binary RUSHU and stratifying raters by experience and pairs. This indicated that RUSHU had measurable reliability. Reliabilities were consistent across age groups, cortices, and levels of clinical experience.