Associations between preoperative Oxford hip and knee scores and costs and quality of life of patients undergoing primary total joint replacement in the NHS England: an observational study

Abstract

Objectives: To assess how costs and quality of life (measured by EuroQoL-5 Dimensions (EQ-5D)) before and after total hip replacement (THR) and total knee replacement (TKR) vary with age, gender and preoperative Oxford hip score (OHS) and Oxford knee score (OKS).

Design: Regression analyses using prospectively collected data from clinical trials, cohort studies and administrative data bases.

Setting: UK secondary care.

Participants: Men and women undergoing primary THR or TKR. The Hospital Episode Statistics data linked to patient-reported outcome measures included 602 176 patients undergoing hip or knee replacement who were followed up for up to 6 years. The Knee Arthroplasty Trial included 2217 patients undergoing TKR who were followed up for 12 years. The Clinical Outcomes in Arthroplasty Study cohort included 806 patients undergoing THR and 484 patients undergoing TKR who were observed for 1 year.

Outcome measures: EQ-5D-3L quality of life before and after surgery, costs of primary arthroplasty, costs of revision arthroplasty and the costs of hospital readmissions and ambulatory costs in the year before and up to 12 years after joint replacement.

Results: Average postoperative utility for patients at the 5th percentile of the OHS/OKS distribution was 0.61/0.5 for THR/TKR and 0.89/0.85 for patients at the 95th percentile. The difference between postoperative and preoperative EQ-5D utility was highest for patients with preoperative OHS/OKS lower than 10. However, postoperative EQ-5D utility was higher than preoperative utility for all patients with OHS≤46 and those with OKS≤44. In contrast, costs were generally higher for patients with low preoperative OHS/OKS than those with high OHS/OKS. For example, costs of hospital readmissions within 12 months after primary THR/TKR were £740/£888 for patients at the 5th percentile compared with £314/£404 at the 95th percentile of the OHS/OKS distribution.

Conclusions: Our findings suggest that costs and quality of life associated with total joint replacement vary systematically with preoperative symptoms measured by OHS/OKS.

Keywords: costs; hip replacement; knee replacement; oxford hip score; oxford knee score; quality-of-life.

Figure 1

Associations between OHS/OKS and quality-of-life, costs of primary and revision TJR. Costs are measured in 2014 pound sterling (£). Lines show predicted values averaged over all observations in the sample with a given Oxford hip score (OHS)/Oxford knee score (OKS). All models included preoperative OHS or OKS score, age at operation and sex of the patient as independent variables, with the exception of the model for preoperative EQ-5D of total hip replacement (THR) patients, which did not include an indicator for female patients. Full regression results are shown in online supplementary appendix C. The grey areas show 95% CIs obtained from 1000 bootstrap replications. National Health Service patient-reported outcome measures and Hospital Episode Statistics data, 2009– 2015; own calculations. EQ-5D, EuroQoL-5 Dimensions; TJR, total joint replacement; TKR, total knee replacement.

Figure 2

Associations between OHS/OKS and costs before TJR, ambulatory and hospital readmission costs <12 months after TJR. Costs are measured in 2014 pound sterling (£). Lines show predicted values averaged over all observations in the sample with a given Oxford hip score (OHS)/Oxford knee score (OKS). All models included preoperative OHS or OKS score and age at operation as independent variables. The models for readmission costs also included a binary indicator for female patients. Full regression results are shown in online supplementary appendix C. The grey areas show 95% CIs obtained from 1000 bootstrap replications. National Health Service patient-reported outcome measures, Hospital Episode Statistics, Knee Arthroplasty Trial and Clinical Outcomes in Arthroplasty Study data, 2009–2015; own calculations. THR, total hip replacement; TJR, total joint replacement; TKR, total knee replacement.

Figure 3

Associations between OHS/OKS and hospital readmission costs >12 months after THR. Costs are measured in 2014 pound sterling (£). Lines show predicted values averaged over all observations in the sample with a given Oxford hip score (OHS)/Oxford knee score (OKS). All models included preoperative OHS or OKS score as independent variables. The model for annual readmission costs between 1 and 6 years after total hip replacement (THR) additionally included current age and an indicator for time since primary surgery. The corresponding model for total knee replacement (TKR) included age at operation, time since primary surgery and an indicator for sex of the patient as independent variables. The models in (B) additionally included age at operation, time since primary surgery and sex of the patient as covariates. The model for annual readmission costs after revision TKR surgery included current age and time since primary surgery as additional covariates. Full regression results are shown in online supplementary appendix C. The grey areas show 95% CIs obtained from 1000 bootstrap replications. National Health Service patient-reported outcome measures and Hospital Episode Statistics data, 2009–2015; own calculations.