Systematic review of machine-learning models in orthopaedic trauma - overview and quality assessment of 45 studies

Aug 2, 2026

Upload by Violet

Machine-learning prediction models in orthopaedic trauma offer promise for assisting clinicians with tasks such as personalized risk stratification, yet a synthesized overview and rigorous critical appraisal against reporting and bias standards has been missing. A systematic search identified 3,252 studies and included 45 ML-based prediction model studies. TRIPOD was used to assess completeness of reporting and PROBAST to evaluate risk of bias. Results show only modest completeness and a predominance of high or unclear bias, highlighting gaps before clinical adoption.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages11

Reading Time28 min

AccessFree to download

Aims

Methods

Results

Conclusion

Take home message

Introduction

Systematic review of machine-learning models in orthopaedic trauma an overview and quality assessment of 45 studies | Cite this article:
Bone Jt Open 2024;5(1): 9–19.
DOI: 10. 1302/2633-1462 .
51.BJO-2023-0095.R1
Correspondence should be sent to H. Dijkstra h.b. [dijkstra@umcg.nl](dijkstra@umcg.nl)
| H. Dijkstra,1,2 A. van de Kuit,1 T. M. de Groot,1,3 O. Canta,1 O. Q. Groot,4 J. H.F. Oosterho,f 5 J. N. Doornberg,1,6 On behalf of the Machine Learning Consortium
1 Department of Orthopaedic Surgery, University Medical Centre Groningen, Groningen, Netherlands
2 University Center for Geriatric Medicine, University of Groningen, University Medical Center Groningen, Groningen, Netherlands
3 Department of Orthopaedic Surgery, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, USA
4 Department of Orthopaedic Surgery, University Medical Centre Utrecht, University of Utrecht, Utrecht, Netherlands
5 Department of Engineering Systems & Services, Faculty Technology Policy and Management, Delft University of Technology, Delft, Netherlands
6 Department of Orthopaedic Trauma Surgery, Flinders Medical Center, Flinders University, Adelaide, Australia | | --- | --- | | | Aims
Machine-learning (ML) prediction models in orthopaedic trauma hold great promise in assisting clinicians in various tasks, such as personalized risk stratification. However, an overview of current applications and critical appraisal to peer-reviewed guidelines is lacking. The objectives of this study are to 1) provide an overview of current ML prediction models in orthopaedic trauma; 2) evaluate the completeness of reporting following the Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis (TRIPOD) statement; and 3) assess the risk of bias following the Prediction model Risk Of Bias Assessment Tool (PROBAST) tool.
Methods
A systematic search screening 3,252 studies identified45 ML-based prediction models in orthopaedic trauma up to January 2023. The TRIPOD statement assessed transparent reporting and the PROBAST tool the risk of bias.
Results
A total of 40 studies reported on training and internal validation; four studies performed both development and external validation, and one study performed only external validation. The most commonly reported outcomes were mortality (33%, 15/45) and length of hospital stay (9%, 4/45), and the majority of prediction models were developed in the hip fracture population (60%, 27/45) . The overall median completeness for the TRIPOD statement was 62%(interquartile range 30 to 81%). The overall risk of bias in the PROBAST tool was low in 24%(11/45), high in 69%(31/45), and unclear in 7%(3/45) of the studies. High risk of bias was mainly due to analysis domain concerns including small datasets with low number of outcomes, complete-case analysis in case of missing data, and no reporting of performance measures.
Conclusion
The results of this study showed that despite a myriad of potential clinically useful applications, a substantial part of ML studies in orthopaedic trauma lack transparent reporting, and areat high risk | Systematic review of machine-learning models in orthopaedic trauma 9 H. Dijkstra, A. van de Kuit, T. M. de Groot, et al. of bias. These problems must be resolved by following established guidelines to instil confidence in ML models among patients and clinicians. Otherwise, there will remain a sizeable gap between the development of ML prediction models and their clinical application in our day-to-day orthopaedic trauma practice. Take home message • Useful applications of machine-learning prediction models in orthopaedic trauma exist, but a substantial proportion lack external validation and transparent reporting, and are at high risk of bias. Introduction Machine learning (ML) has shown great potential in aiding clinicians with different tasks in orthopaedic trauma. 1,2 Specific ap

FAQs

What is the main aim of this systematic review?

The review provides an overview of machine-learning prediction models in orthopaedic trauma and evaluates both reporting completeness using TRIPOD and risk of bias using PROBAST.

How many studies were included and how were they identified?

A systematic search screened 3,252 studies, ultimately including 45 ML-based prediction model studies up to January 2023.

What did the review find about reporting completeness and risk of bias?

Overall median TRIPOD completeness was 62%, while PROBAST showed low risk of bias in 24% of studies, high risk in 69%, and unclear risk in 7%. High risk was mainly driven by analysis domain concerns.

Why is external validation important according to the review?

The review emphasizes that clinical implementation should involve external validation and prospective testing, since generalizability cannot be established from a single external validation study.

Aug 2, 2026

Upload by Violet