Machine Learning in the Prediction of Trauma Outcomes: A Systematic Review

Zhang, Timothy; Nikouline, Anton; Lightfoot, David; Nolan, Brodie

doi:10.1016/j.annemergmed.2022.05.011

Cited by 10 publications

(8 citation statements)

References 90 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Previous research successfully applied some of the learners used in our study for predicting mortality and other more homogeneous outcomes in trauma, were equal performance where often found using logistic regression compared to other learners (35). While OFI presents a more dilJicult prediction challenge, our study showed similar results where logistic regression had average performance.…”

Section: Discussionmentioning

confidence: 70%

“…A systematic review from 2022 on the use of ML for other trauma related predictions found a similar trend (18). They identified 25 studies comparing the performance of LR to other learners in which performance where similar in twelve studies, better for non LR learners in ten, and worse in three.…”

Section: Discussionmentioning

confidence: 79%

“…In our study all ML learners showed relatively similar results however newer models showed a small, but significant, performance boost compared to more established learners such as k-NN, SVM or LR. A systematic review from 2022 on the use of ML for other trauma related predictions found a similar trend (18). They identified 25 studies comparing the performance of LR to other learners in which performance where similar in twelve studies, better for non LR learners in ten, and worse in three.…”

Section: Discussionmentioning

confidence: 79%

“…While previous research successfully apply some of the learners used in our study to predict mortality and other more homogeneous outcomes (18), the heterogeneous nature of OFI presents a more difficult prediction challenge. It is possible that higher quality data with 11 reduced missing frequency and higher resolution, eg vital signs series instead of single values, additional predictors and complex machine learning systems such as ensembles are needed for more precise and better performing models.…”

Section: Interpretation and Implicationsmentioning

confidence: 99%

“…Different machine learning (ML) based prediction models have been used for several outcomes in trauma care (18), but the use of ML for OFI prediction has never been investigated. Therefore, our aim was to develop, validate, and compare the performance of different machine learning models for predicting OFI.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Predicting Opportunities for Improvement in Trauma Care: A Registry-Based Cohort Study

Attergrim

Szolnoky

Strömmer

et al. 2023

Preprint

View full text Add to dashboard Cite

Rationale The multidisciplinary mortality and morbidity conference is the core of programs that aim to improve the quality of trauma care and is used to identify and address opportunities for improvement based on reviewing patient cases. Current systems rely on audit filters when selecting patients for review, a process that is hampered by high frequencies of false positives. Objectives To develop, validate, and compare the performance of different machine learning models for predicting opportunities for improvement. Methods We conducted a registry based study using all patients from the Karolinska university hospital that had been reviewed regarding the presence of opportunity for improvement, a binary consensus decision from the mortality and morbidity conference. We developed eight binary classification models using 45 predictors. Training used an 80%-20% train-test split and 1000 resamples without replacement estimated confidence intervals. Performance was also compared to current audit filters. Measurements and Main Results The dataset included 6310 patients where opportunities for improvement were present among 431 (7%) patients. The audit filters (Area under the receiver operating characteristics curve: 0.624) was outperformed by all machine learning models. The best performing model was LightGBM (Area under the receiver operating characteristics curve: 0.789). Conclusions Machine learning models outperform the currently used audit filters and could prove to be valuable additions in the screening for opportunities for improvement. More research is needed on how to increase model performance and how to incorporate these models into trauma quality improvement programs.

show abstract

Section: Discussionmentioning

confidence: 70%

Section: Discussionmentioning

confidence: 79%

Section: Discussionmentioning

confidence: 79%

Section: Interpretation and Implicationsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations