Background: Many patients suffer from osteoarthritis (OA) in multiple joints, possibly resulting in multiple joint arthroplasties (MJAs). Primarily, we determined the cumulative incidence (C_in) of MJA in hip and knee joints up to 10 years. Secondly, we calculated the mean time between the first and subsequent joint arthroplasty, and evaluated the different MJA trajectories. Lastly, we compared patient characteristics and outcomes (functionality and pain) after surgery between MJA patients and single hip arthroplasty or knee arthroplasty (HA and KA) patients. Methods: Primary index (first) HA or KA for OA were extracted from the Dutch Arthroplasty Register. The 1, 2, 5, and 10-year C_in (including competing risk death) of MJA, mean time intervals, and MJA-trajectories were calculated and stratified for primary index HA or KA. Sex, preoperative age, and body mass index were compared using ordinal logistic regression. Outcomes, measured preoperatively, 3, 6, and 12 months postoperatively (function: Hip Disability or Knee Injury and OA Outcome Score; Pain: Numerical Rating Scale), were compared using linear regression. Results: A total of 140,406 HA-patients and 140,268 KA-patients were included. One, 2, 5, and 10-year C_in for a second arthroplasty were respectively 8.9% [95% confidence interval (CI): 8.7 to 9.0], 14.3% [95% CI: 14.1 to 14.5], 24.0% [95% CI: 23.7 to 24.2], and 32.7% [95% CI: 32.2 to 33.1] after index HA, and 9.5% [95% CI: 9.4 to 9.7], 16.0% [95% CI: 15.9 to 16.2], 26.4% [95% CI: 26.1 to 26.6], and 35.8% [95% CI: 35.4 to 36.3] after index KA. The 10-year C_in for > 2 arthroplasties were small in both the index HA and KA groups. Time-intervals from first to second, third, and fourth arthroplasty were 26 [95% CI: 26.1 to 26.7], 47 [95% CI: 46.4 to 48.4], and 58 [95% CI: 55.4 to 61.1] months after index HA, and 26 [95% CI: 25.9 to 26.3], 52 [95% CI: 50.8 to 52.7], and 61 [95% CI: 58.3 to 63.4] months after index KA. There were 83% of the second arthroplasties placed in the contralateral cognate joint (ie, knee or hip). Differences in postoperative functionality and pain between MJAs and single HAs and KAs were small. Conclusions: The 10-year C_in showed that about one-third of patients received a second arthroplasty after approximately 2 years, with the majority in the contralateral cognate joint. Few patients received > 2 arthroplasties within 10 years. Being a women, having a higher body mass index, and being younger increased the odds of MJA. Postoperative outcomes were slightly negatively affected by MJA.

@en

Background
Estimating the risk of revision after arthroplasty could inform patient and surgeon decision-making. However, there is a lack of well-performing prediction models assisting in this task, which may be due to current conventional modeling approaches such as traditional survivorship estimators (such as Kaplan-Meier) or competing risk estimators. Recent advances in machine learning survival analysis might improve decision support tools in this setting. Therefore, this study aimed to assess the performance of machine learning compared with that of conventional modeling to predict revision after arthroplasty.

Question/purpose
Does machine learning perform better than traditional regression models for estimating the risk of revision for patients undergoing hip or knee arthroplasty?

Methods
Eleven datasets from published studies from the Dutch Arthroplasty Register reporting on factors associated with revision or survival after partial or total knee and hip arthroplasty between 2018 and 2022 were included in our study. The 11 datasets were observational registry studies, with a sample size ranging from 3038 to 218,214 procedures. We developed a set of time-to-event models for each dataset, leading to 11 comparisons. A set of predictors (factors associated with revision surgery) was identified based on the variables that were selected in the included studies. We assessed the predictive performance of two state-of-the-art statistical time-to-event models for 1-, 2-, and 3-year follow-up: a Fine and Gray model (which models the cumulative incidence of revision) and a cause-specific Cox model (which models the hazard of revision). These were compared with a machine-learning approach (a random survival forest model, which is a decision tree–based machine-learning algorithm for time-to-event analysis). Performance was assessed according to discriminative ability (time-dependent area under the receiver operating curve), calibration (slope and intercept), and overall prediction error (scaled Brier score). Discrimination, known as the area under the receiver operating characteristic curve, measures the model’s ability to distinguish patients who achieved the outcomes from those who did not and ranges from 0.5 to 1.0, with 1.0 indicating the highest discrimination score and 0.50 the lowest. Calibration plots the predicted versus the observed probabilities; a perfect plot has an intercept of 0 and a slope of 1. The Brier score calculates a composite of discrimination and calibration, with 0 indicating perfect prediction and 1 the poorest. A scaled version of the Brier score, 1 – (model Brier score/null model Brier score), can be interpreted as the amount of overall prediction error.

Results
Using machine learning survivorship analysis, we found no differences between the competing risks estimator and traditional regression models for patients undergoing arthroplasty in terms of discriminative ability (patients who received a revision compared with those who did not). We found no consistent differences between the validated performance (time-dependent area under the receiver operating characteristic curve) of different modeling approaches because these values ranged between -0.04 and 0.03 across the 11 datasets (the time-dependent area under the receiver operating characteristic curve of the models across 11 datasets ranged between 0.52 to 0.68). In addition, the calibration metrics and scaled Brier scores produced comparable estimates, showing no advantage of machine learning over traditional regression models.

Conclusion
Machine learning did not outperform traditional regression models.

Clinical Relevance
Neither machine learning modeling nor traditional regression methods were sufficiently accurate in order to offer prognostic information when predicting revision arthroplasty. The benefit of these modeling approaches may be limited in this context.

@en

Background and purpose — Most arthroplasty registers give hospital-specific feedback on revision rates after total hip and knee arthroplasties (THA/TKA). However, due to the low number of events per hospital, multiple years of data are required to reliably detect worsening performance, and any single indicator provides only part of the quality of care delivered. Therefore, we developed an ordered composite outcome including revision, readmission, complications, and long length-of-stay (LOS) for a more comprehensive view on quality of care and assessed the ability to reliably differentiate between hospitals in their performance (rankability) with fewer years of data. Methods — All THA and TKA performed between 2017 and 2019 in 20 Dutch hospitals were included. All combi-nations of the 4 indicators were ranked from best to worst to create the ordinal composite outcome for THA and TKA separately. Between-hospital variation for the composite outcome was compared with individual indicators standard-ized for case-mix differences, and we calculated the statisti-cal rankability using fixed and random effects models. Results — 22,908 THA and 20,423 TKA were included. Between-hospital variation for the THA and TKA composite outcomes was larger when compared with revision, readmis-sion, and complications, and similar to long LOS. Rankabili-ties for the composite outcomes were above 80% even with 1 year of data, meaning that largely true hospital differences were detected rather than random variation. Interpretation — The ordinal composite outcome gives a more comprehensive overview of quality of delivered care and can reliably differentiate between hospitals in their performance using 1 year of data, thereby allowing earlier intro-duction of quality improvement initiatives.@en

Background and purpose — A lifetime perspective on revision risks is needed for optimal timing of arthroplasty in osteoarthritis (OA) patients, weighing the benefit of total hip arthroplasty/total knee arthroplasty (THA/TKA) against the risk of revision, after which outcomes are less favorable. Therefore, we provide population-based 10-year cumulative revision risks stratified by joint, sex, fixation type, and age. Patients and methods — Data from the Dutch Arthroplasty Register (LROI) was used. Primary THAs and TKAs for OA between 2007 and 2018 were included, except metal-on-metal prostheses or hybrid/reversed hybrid fixation. Revision surgery was defined as any change of 1 or more prosthesis components. The 10-year cumulative revision risks were calculated stratified by joint, age, sex, at primary arthroplasty, and fixation type (cemented/uncemented), taking into account mortality as a competing risk. We estimated the percentage of potentially avoidable revisions assuming all OA patients aged < 75 received primary THA/TKA 5 years later while keeping age-specific 10-year revision risks constant. Results — 214,638 primary THAs and 211,099 TKAs were included, of which 31% of THAs and 95% of TKAs were cemented. The 10-year cumulative revision risk varied between 1.6% and 13%, with higher risks in younger age categories. Delaying prosthesis placement by 5 years could potentially avoid 23 (3%) THA and 162 (17%) TKA revisions. Interpretation — Cumulative 10- year revision risk varied considerably by age in both fixation groups, which may be communicated to patients and used to guide timing of surgery.@en

Background:Given the low early revision rate after total hip arthroplasty (THA) and total knee arthroplasty (TKA), hospital performance is typically compared using 3 years of data. The purpose of this study was to assess how much earlier worsening hospital performance in 1-year revision rates after THA and TKA can be detected.Methods:All 86,468 THA and 73,077 TKA procedures performed from 2014 to 2016 and recorded in the Dutch Arthroplasty Register were included. Negative outlier hospitals were identified by significantly higher O/E (observed divided by expected) 1-year revision rates in a funnel plot. Monthly Shewhart p-charts (with 2 and 3-sigma control limits) and cumulative sum (CUSUM) charts (with 3.5 and 5 control limits) were constructed to detect a doubling of revisions (odds ratio of 2), generating a signal when the control limit was reached. The median number of months until generation of a first signal for negative outliers and the number of false signals for non-negative outliers were calculated. Sensitivity, specificity, and accuracy were calculated for all charts and control limit settings using outlier status in the funnel plot as the gold standard.Results:The funnel plot showed that 13 of 97 hospitals had significantly higher O/E 1-year revision rates and were negative outliers for THA and 7 of 98 hospitals had significantly higher O/E 1-year revision rates and were negative outliers for TKA. The Shewhart p-chart with the 3-sigma control limit generated 68 signals (34 false-positive) for THA and 85 signals (63 false-positive) for TKA. The sensitivity for THA and TKA was 92% and 100%, respectively; the specificity was 69% and 51%, respectively; and the accuracy was 72% and 54%, respectively. The CUSUM chart with a 5 control limit generated 18 signals (1 false-positive) for THA and 7 (1 false-positive) for TKA. The sensitivity was 85% and 71% for THA and TKA, respectively; the specificity was 99% for both; and the accuracy was 97% for both. The Shewhart p-chart with a 3-sigma control limit generated the first signal for negative outliers after a median of 10 months (interquartile range [IQR] = 2 to 18) for THA and 13 months (IQR = 5 to 18) for TKA. The CUSUM chart with a 5 control limit generated the first signal after a median of 18 months (IQR = 7 to 22) for THA and 21 months (IQR = 9 to 25) for TKA.Conclusions:Monthly monitoring using CUSUM charts with a 5 control limit enables earlier detection of worsening 1-year revision rates with accuracy so that initiatives to improve care can start earlier. @en

Purpose: Total knee arthroplasty (TKA) is usually effective, although not all patients have satisfactory outcomes. This assumes distinct recovery patterns might exist. Little attention has been paid to determine which patients have worse outcomes. This study attempts to distinguish specific recovery patterns using the Oxford knee score (OKS) during the first postoperative year. The secondary aim was to explore predictors of less favourable recovery patterns. Methods: Analysis of patients in the Dutch Arthroplasty Register (LROI) with unilateral primary TKA. Data collected up to one year postoperative was used. To identify subgroups of patients based on OKS, latent class growth modeling (LCGM) was used. Moreover, multivariable multinomial logistic regression analysis was used to explore predictors of class membership. Results: 809 Patients completed three OKS during the first year postoperative and were included. LCGM identified 3 groups of patients; ‘high risers’ (most improvement during first 6-months, good 12-month scores 77%), ‘gradual progressors’ (continuous improvement during the first year 13%) and ‘non responders’ (initial improvement and subsequent deterioration to baseline score 10%). Predictors of least favourable class membership (OR, 95%CI) are EQ-5D items: VAS health score (0.83, 0.73–0.95), selfcare (2.22, 1.09–4.54) and anxiety/depression (2.45, 1.33–4.52). Conclusion: Three recovery patterns after TKA were distinguished; ‘high risers', ‘gradual progressors' and ‘non responders'. Worse score on EQ-5D items VAS health, selfcare, and anxiety/depression were correlated with the least favourable ‘non responders’ recovery pattern.

@en

BACKGROUND: Variation in 1-year revision rates between Dutch hospitals after primary total hip and knee arthroplasty (THA and TKA) may direct quality-improvement initiatives if this variation accurately reflects true hospital differences. The aim of the present study was to assess the extent of variation, both overall and for specific indications, as well as the statistical reliability of ranking hospitals. METHODS: All primary THAs and TKAs that were performed between January 2014 and December 2016 were included. Observed/expected (O/E) ratios regarding 1-year revision rates were depicted in a funnel plot with 95% control limits to identify outliers based on 1 or 3 years of data, both overall and by specific indication for revision. The expected number was calculated on the basis of patient mix with use of logistic regression models. The statistical reliability of ranking hospitals (rankability) on these outcomes indicates the percentage of total variation that is explained by "true" hospital differences rather than chance. Rankability was evaluated using fixed and random effects models, for overall revisions and specific indications for revision, including 1 versus 3 years of data. RESULTS: The present study included 86,468 THAs and 73,077 TKAs from 97 and 98 hospitals, respectively. Thirteen hospitals performing THAs were identified as negative outliers (median O/E ratio, 1.9; interquartile range [IQR], 1.5-2.5), with 5 hospitals as outliers in multiple years. Eight negative outliers were identified for periprosthetic joint infection; 4, for dislocation; and 2, for prosthesis loosening. Seven hospitals performing TKAs were identified as negative outliers (median O/E ratio, 2.3; IQR, 2.2-2.8), with 2 hospitals as outliers in multiple years. Two negative outlier hospitals were identified for periprosthetic joint infection and 1 was identified for technical failures. The rankability for overall revisions was 62% (moderate) for THA and 46% (low) for TKA. CONCLUSIONS: There was large between-hospital variation in 1-year revision rates after primary THA and TKA. For most outlier hospitals, a specific indication for revision could be identified as contributing to worse performance, particularly for THA; these findings are starting points for quality-improvement initiatives.@en