Does change in language change the properties of a shortened score previously validated in its complete version? Validation of the French versions of the HOOS-12 and KOOS-12 scores in primary knee and hip arthroplasties

Abstract

BackgroundThe HOOS and KOOS scoring questionnaires comprise respectively 40 and 42 items; a shorter 12-item version was recently developed, but remains to be validated in a French-speaking population. We therefore conducted a prospective study: 1) to determine whether the new 12-item versions in French are equivalent to the longer HOOS and KOOS versions, and 2) to validate the French-language HOOS-12 and KOOS-12 patient-reported outcome measures in a population of primary total hip and knee arthroplasty: validity, reliability, and responsiveness. HypothesisThe change in language in a score already validated in its long version does not alter its properties in the short version. Material and methodsOne hundred patients (59 males, 41 females) undergoing primary total hip arthroplasty and 100 patients (43 males, 57 females) undergoing primary total knee arthroplasty were prospectively included. They filled out the original HOOS or KOOS questionnaires, their simplified versions (PS: Physical function Short form; JR: Joint Replacement) and the short HOOS-12 and KOOS-12 versions, and also the Oxford-12 score assessing the affected joint, preoperatively, then at 6–12 months. ResultsThe 100% response rate confirmed ease of use. There were no redundant items. There were strong correlations between the 12-item and longer versions (>0.9). The HOOS-12 and KOOS-12 scores were reliable and valid: 1) there were no ceiling or floor effects for pre- or postoperative KOOS-12 scores, although a ceiling effect was found for HOOS-12 postoperatively (20% of patients having maximum scores of 100); 2) internal consistency was confirmed, with Cronbach alpha>0.8; 3) external consistency between Oxford-12 and HOOS-12/KOOS-12 was excellent, with Pearson correlation coefficient>0.8. Sensitivity to pre-/postoperative change was confirmed, with effect size>0.8. DiscussionThe present study confirmed the usefulness of this new 12-item form for HOOS and KOOS. Properties were identical between the French- and English-language versions, authorising everyday use of these simpler versions. Level of evidenceIV; prospective study without control group.