Anchor-based predictive modelling and receiver operating curve estimates of patient acceptable symptom state for the forgotten joint score in total knee arthroplasty patients stratified by age and gender

Tools

Abugu, Glory Uche and Holloway, Nicholas and Riches, Philip and Clarke, Jon and Giardini, Mario Ettore and Chopra, Swati (2025) Anchor-based predictive modelling and receiver operating curve estimates of patient acceptable symptom state for the forgotten joint score in total knee arthroplasty patients stratified by age and gender. Journal of Clinical Epidemiology, 185. 111897. ISSN 1878-5921 (https://doi.org/10.1016/j.jclinepi.2025.111897)

[thumbnail of Abugu-etal-JCE-2025-Anchor-based-predictive-modelling-and-receiver-operating-curve-estimates-of]
Preview
 Text. Filename: Abugu-etal-JCE-2025-Anchor-based-predictive-modelling-and-receiver-operating-curve-estimates-of.pdf
Final Published Version
License: Creative Commons Attribution 4.0 logo
Download (415kB)| Preview

Abstract

Objective: To estimate patient acceptable symptom state (PASS) thresholds in the Forgotten Joint Score (FJS) one year following primary total knee arthroplasty (PTKA) while investigating the impact of patients’ characteristics on PASS thresholds. Study Design and Setting: This cohort study used data from patients who underwent PTKA at a public hospital in Scotland between April 2021 and December 2022. Assessment of FJS (0-100, high-low knee awareness) was completed one year post-operatively. A single-item question about satisfaction with the operated knee was completed at one year and served as the anchor for estimating PASS thresholds. Anchor-based predictive modelling (adjusted and unadjusted) and receiver operating characteristic (ROC) curve methods were used to determine PASS thresholds. The impact of patient characteristics on PASS threshold values was investigated by calculating stratified PASS values based on gender and age groups. Results: A total of 1832 PTKAs was performed between April 2021 and December 2022, of which 1359/1832 (74%) had complete data comprising the study cohort. The median age and BMI of patients included in the study was 70 years and 31.2 kg⁄m2 with 54% being females. The proportion of satisfied patients was 84%. A moderate positive correlation between FJS and patient satisfaction was found (r = 0.64, P < .001) which supports the validity of the external anchor. PASS thresholds for the entire cohort were 31 (ROC method) and ~33 points (predictive modelling method). Larger PASS values were found for male patients and patients aged 70+ years compared to their female and younger counterparts. The adjusted predictive modelling estimate was 15.3, given that the data does not meet the assumption of normal distribution, we consider this threshold might be biased and must be interpreted with circumspection. Conclusion: A postoperative FJS of ≥ 33 points can be used as a reference guide to evaluate successful achievement of a “forgotten joint” in a Scottish population. Patients’ characteristics impact PASS estimates and should be considered when interpreting outcome scores.

ORCID iDs

Abugu, Glory Uche ORCID logoORCID: https://orcid.org/0000-0001-8750-6394, Holloway, Nicholas, Riches, Philip ORCID logoORCID: https://orcid.org/0000-0002-7708-4607, Clarke, Jon, Giardini, Mario Ettore and Chopra, Swati;
CORE (COnnecting REpositories)