Artificial intelligence and machine learning: an important new set of tools for clinical shoulder arthroplasty research.

BACKGROUND: Artificial intelligence (AI) and machine learning (ML) have increasingly transformative potential in orthopedic surgery, enhancing precision, efficiency, and outcomes. We aim to raise surgeon and researcher awareness of the Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis (TRIPOD+AI) reporting guidelines for developing and applying AI/ML by conducting a gap analysis on manuscripts published prior or immediately after the release of the guidelines. Our goal is to help improve future reporting and transparency of AI/ML studies regarding shoulder arthroplasty and similar reconstructive procedures. METHODS: The (TRIPOD+AI) reporting guidelines are presented. A systematic search yielded 2,454 studies sourced from the PubMed, Embase, and Cochrane databases. Articles met inclusion criteria if they discussed the development or evaluation of an ML model or an algorithm for use in orthopedic shoulder arthroplasty. After screening, 32 total articles across all 3 databases met inclusion criteria. The transparency of reporting for each study was assessed according to the TRIPOD+AI statement. Three reviewers extracted data from the included studies to assess their adherence to TRIPOD+AI. RESULTS: It was found that 3 (9.4%) studies referenced TRIPOD+AI. The median adherence to the TRIPOD+AI guidelines was 56.5% across all studies and the mean adherence was 55.8%. Risk of bias of the included studies was rated overall as low in 40.6% (13/32), high in 34.4% (11/32), and unclear in 25.0% (8/32) according to The Prediction model Risk of Bias Assessment Tool. CONCLUSION: Overall, we found that recent shoulder AI/ML arthroplasty studies demonstrate an encouraging baseline adherence to TRIPOD+AI guidelines from which awareness and future improvement can be established. This review serves as an important first step in evaluating the completeness of transparent reporting for ML prediction models in shoulder surgery, offering a baseline and stimulus for future improvements.