Key result
Machine learning models using LASSO regression demonstrated good accuracy for predicting 30-day death (C-statistic 0.73; 95% CI 0.70-0.76) and cardiac complications after elective primary TJA.
Why the study?
Can machine learning methods produce accurate prediction models for 30-day mortality and complications after elective total joint arthroplasty?
Population
107,792 nonemergent primary THAs and TKAs in the 2013 to 2014 ACS-NSQIP
Comparison
Machine learning prediction models vs observed outcomes with VASQIP external validation
Design
Retrospective risk-prediction model development and validation study
Follow-up
30 days
Authors
Loading...
Current TJA risk models have limitations; leaves open whether new validated tools will improve preoperative decisions.
Observational (n=107,792)
Yes
Can machine learning methods produce accurate prediction models for 30-day mortality and complications after elective total joint arthroplasty?
Effect estimate: C-statistic 0.73 (for death) (95% CI 0.70-0.76)
Machine learning models using preoperative variables can accurately predict 30-day mortality and cardiac complications after elective primary total joint arthroplasty.
Harris et al. (2019) conducted an observational in Elective total joint arthroplasty (TJA) (n=107,792). Machine learning risk prediction models (LASSO regression) vs. VASQIP-derived models was evaluated on 30-day death and major complications (C-statistic 0.73 (for death), 95% CI 0.70-0.76). Machine learning models using LASSO regression demonstrated good accuracy for predicting 30-day death (C-statistic 0.73; 95% CI 0.70-0.76) and cardiac complications after elective primary TJA.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: