Prediction Models Developed to Predict Risk for Total Hip, Knee Replacements

Two new prediction equations have been developed and validated to quantify the absolute risks for future primary total hip replacement (THR) and total knee replacement (TKR) in patients newly diagnosed with hip pain or osteoarthritis (OA) or knee pain or OA in the primary care setting, according to findings published in Annals of the Rheumatic Diseases.

The investigators sought to design and validate 2 prediction models to estimate the future risk for primary THR and TKR among patients with early-stage OA, using clinical variables that are available in the United Kingdom primary care electronic health records. Two prospective open cohorts of patients ≥40 years from the UK Clinical Practice Research Datalink were identified and evaluated. Researchers identified candidate predictors by systematic review, panel consensus, and a novel hypothesis-free “Record-Wide Association Study” with replication. To derive risk algorithms for THR and TKR, they applied Cox proportional hazards models to account for the competing risk for death.

Overall, 45 predictors for THR and 53 predictors for TKR were identified, selected, and reviewed by the panel. A total of 301,052 and 416,030 patients newly consulting between 1992 and 2015 were identified in the hip and knee cohorts, respectively. The median follow-up time was 6 years. Internal-external cross-validation was applied over geographic regions to validate the 2 models.  

The final THR prediction model, with 20 predictors, was able to distinguish between patients with and without a primary THR with a C-statistic of 0.73 (95% CI, 0.72-0.73) over the 10-year follow-up period. Moreover, the final TKR prediction model, with 24 predictors, was able to differentiate between those with and without a primary TKR with a C-statistic of 0.79 (95% CI, 0.78-0.79) over the 10-year follow-up period. The calibration slope was 1.00 (95% CI, 0.98-1.02) and 1.00 (95% CI, 0.99-1.01) for THR and TKR, respectively.

The internal-external cross-validation C-statistics ranged between 0.70 and 0.74 for the THR model and between 0.76 and 0.82 for the TKR model. The internal-external cross-validation calibration slope ranged between 0.93 and 1.07 for the THR model and between 0.92 and 1.12 for the TKR model.

The investigators concluded that the 2 newly developed prediction models demonstrated good discrimination and calibration for estimating an individual’s risk for THR and TKR. These models have been validated in large-scale, nationally representative data and are readily automated in electronic patient records. The algorithms can be used to inform clinical decision making, such as targeting intensive, nonsurgical management of patients identified at high risk for future primary THR or TKR. Future research is warranted to assess the clinical outcomes and cost-effectiveness of using these risk equations in primary care.

follow @RheumAdvisor

Reference

Yu D, Jordan KP, Snell KIE, et al. Development and validation of prediction models to estimate risk of primary total hip and knee replacements using data from the UK: two prospective open cohorts using the UK Clinical Practice Research Datalink [published online October 18, 2018]. Ann Rheum Dis. doi:10.1136/annrheumdis-2018-213894