Accuracy and Validity of Computer Adaptive Testing for Outcome Assessment in Patients Undergoing Total Hip Arthroplasty

Abstract

BackgroundProbability-based computer algorithms that reduce patient burden are currently in high demand. These computer adaptive testing (CAT) methods improve workflow and reduce patient frustration, while achieving high measurement precision. In this study, we evaluated the accuracy and validity of the CAT Hip Disability and Osteoarthritis Outcome Score (HOOS) and the Hip Disability and Osteoarthritis Outcome Score Joint Replacement (HOOS-JR) by comparing them to the full version of these scoring systems in a subset of patients who had undergone total hip arthroplasties. MethodsA previously developed CAT HOOS and HOOS-JR was applied to 354 and 1547 HOOS and HOOS-JR patient responses, respectively. Mean, standard deviations, Pearson’s correlation coefficients, interclass correlation coefficients, frequency distribution plots, and Bland-Altman plots were used to compare the precision, validity, and accuracy between CAT scores and full-form scores. ResultsBy modifying the questions to past responses, the CAT HOOS demonstrated a mean reduction of 30% of questions (28 vs 40 questions). There were no significant differences between the full HOOS and CAT HOOS with respect to pain (P = .73), symptoms (P = .94), quality of life (P = .99), activities of daily living (P = .82), and sports (P = .99). There were strong linear relationships between the CAT versions and the standard questionnaires (r > 0.99). The Bland-Altman plot showed that differences between CAT HOOS and full HOOS were independent of the overall scores. ConclusionThe CAT HOOS and HOOS-JR have high correlation and require fewer questions to finish compared to the standard full-form questionnaires. This may represent a reliable and practical alternative that may be less burdensome to patients and may help improve compliance for reporting outcome metrics.