Search reports, ask a question, or jump to a page.
The problem
Intervention: Machine learning predictive models using 119 clinical and demographic predictors for firearm injury identification and 141 predictors for post-discharge mortality prediction
Who it was tested with: Children aged 0-17 years admitted to U.S. trauma centers with injuries (260,098 children aged 0-10 and 242,669 adolescents aged 11-17).
Children or communities are compared with a group that did not receive the intervention, or followed over years. This is the design that best separates what the intervention did from everything else that changed.
"We used ML to analyze 119 predictors and a primary outcome of firearm injury, and then 141 predictors to model a secondary outcome of death after discharge." The study used machine learning with multiple clinical and demographic predictors to develop risk stratification models, testing their performance on a large retrospective cohort with follow-up data.
The model shows moderate discriminative ability with relatively low sensitivity, meaning it may miss a substantial proportion of at-risk patients. The study is limited to injured patients presenting to trauma centers and cannot be applied to prevention in the general population.
Across countries and designs
3 studies · 3 countries
