Comparison of a computer based method and the classical manual method for radiographic joint space width assessment in hip osteoarthritis

Objective. To compare the intraobserver reliability and the sensitivity to change of 2 techniques evaluating the cartilage breakdown in hip osteoarthritis (OA). Methods. Design: 3 year longitudinal study. Participants: patients with painful hip OA. Outcome: coxofemoral joint space width (JSW) at baseline and at 3 year followup was measured on anteroposterior weight-bearing radiographs by 2 methods: a manual method to obtain JSW at the narrowest point (minimal JSW) using a 0.1 mm graduated magnifying glass and a computer based method to obtain minimal and average JSW. Statistical analysis: Two assessments, at one month interval, of each pair of films; evaluation of the intraobserver reliability using the intraclass coefficient of correlation, and the Bland and Altman approach, obtaining the smallest detectable difference (SDD). For each technique, percentage of progressors was evaluated, i.e., the percentage of patients with change greater than the SDD. Evaluation of sensitivity to change was performed using the standardized response mean (SRM), with 95% CI calculated using the jackknife method. Results. Twenty-five pairs of films were evaluated. There were no statistically significant differences between the 3 analyses in the calculated intraclass coefficients of correlation, percentage of progressors, or SRM. Conclusion. These results suggest that both techniques are reliable and sensitive enough to detect changes in a relevant percentage of patients with hip OA after a 3 year followup.