A statistical analysis of hip scores

It is suggested that, for outcome assessment, only three variables need to be recorded: pain, walking distance and range of hip flexion, which is misleading.

Abstract

Thirteen methods of hip scoring were applied in the postoperative assessment of 47 hip arthroplasties. Their results were found to be inconsistent, often giving contrary measures of success in the same patient. Ten variables were measured during the postoperative review of 256 hip arthroplasties and the data were submitted to multivariate factor analysis. This revealed that the ten variables could be reduced to three factors: pain, which correlated poorly with any other variable (Spearman correlation, r 0.9). We suggest that, for outcome assessment, only three variables need to be recorded: pain, walking distance and range of hip flexion. The combination of these three measures into a single hip score is misleading.