Predicting whether patients will achieve minimal clinically important differences following hip or knee arthroplasty

Abstract

Aims: A substantial fraction of patients undergoing knee arthroplasty (KA) or hip arthroplasty (HA) do not achieve an improvement as high as the minimal clinically important difference (MCID), i.e. do not achieve a meaningful improvement. Using three patient-reported outcome measures (PROMs), our aim was: 1) to assess machine learning (ML), the simple pre-surgery PROM score, and logistic-regression (LR)-derived performance in their prediction of whether patients undergoing HA or KA achieve an improvement as high or higher than a calculated MCID; and 2) to test whether ML is able to outperform LR or pre-surgery PROM scores in predictive performance.

Methods: MCIDs were derived using the change difference method in a sample of 1,843 HA and 1,546 KA patients. An artificial neural network, a gradient boosting machine, least absolute shrinkage and selection operator (LASSO) regression, ridge regression, elastic net, random forest, LR, and pre-surgery PROM scores were applied to predict MCID for the following PROMs: EuroQol five-dimension, five-level questionnaire (EQ-5D-5L), EQ visual analogue scale (EQ-VAS), Hip disability and Osteoarthritis Outcome Score-Physical Function Short-form (HOOS-PS), and Knee injury and Osteoarthritis Outcome Score-Physical Function Short-form (KOOS-PS).

Results: Predictive performance of the best models per outcome ranged from 0.71 for HOOS-PS to 0.84 for EQ-VAS (HA sample). ML statistically significantly outperformed LR and pre-surgery PROM scores in two out of six cases.

Conclusion: MCIDs can be predicted with reasonable performance. ML was able to outperform traditional methods, although only in a minority of cases.

Fig. 1

Flowchart of patient enrolment for this study.

Fig. 2

Graphical illustration of the decision-making support given by the prediction models for practical application. Once relevant data are gathered before surgery (1), trained models are fed with the data and make a prediction (2) about whether surgery is recommended for the respective patient given their input variables. Finally, at the time of (potential) surgery (3), patients recommended to undergo surgery do so, while patients not recommended to be operated do not. PROMs, patient-reported outcome measures.

Fig. 3

Receiver operating curves for all models, indications, and patient-reported outcome scores (PROMs). AUC, area under the receiver operating characteristic curve; EQ-5D-5L, EuroQol five-dimension five-level questionnaire; HOOS-PS, Hip disability and Osteoarthritis Outcome Score-Physical Function Short Form; KOOS-PS, Knee injury Osteoarthritis Outcome Score-Physical Function Short Form; LASSO, least absolute shrinkage and selection operator; VAS, visual analogue scale.

Fig. 4

Shapley Additive exPlanations (SHAP) analysis results for knee arthroplasty (KA) and hip arthroplasty (HA) patients and all patient-reported outcome measures (PROMs). Numbers in PROM names (e.g. KOOS_3_2) represent dummies for response options (e.g. response option 2 in KOOS_3 is KOOS_3_2) and the domain of the PROM (i.e. the third domain in KOOS is KOOS_3_2). EQ-5D-5L, EuroQol five-dimension five-level questionnaire; EQ-VAS, EuroQol visual analogue scale; HOOS-PS, Hip disability and Osteoarthritis Outcome Score-Physical Function Short Form; KOOS-PS, Knee injury and Osteoarthritis Outcome Score-Physical Function Short Form; PQ_back, self-reported back pain; PROMIS, patient-reported outcome measurement information system.

Fig. 5

Partial dependence plots for hip and knee arthroplasty patients and all patient-reported outcome measures. EQ-5D-5L, EuroQol five-dimension five-level questionnaire; EQ-VAS, EuroQol visual analogue scale; HOOS-PS, Hip disability and Osteoarthritis Outcome Score-Physical Function Short Form; KOOS-PS, Knee injury and Osteoarthritis Outcome Score-Physical Function Short Form.