Injuries/Violence
Model-based data dashboard for estimating mild traumatic brain injury and musculoskeletal injury risk in United States service members Alexander Ivan B. Posis* Alexander Ivan Posis Posis Posis Posis Posis Posis Posis Posis Posis Military and Veterans Health Solutions, Leidos, Inc., San Diego, CA, USA; Warfighter Performance Department, Naval Health Research Center, San Diego, CA, USA
Background: Mild traumatic brain injury (mTBI) and musculoskeletal injury (MSKI) have negative impacts on service member (SM) readiness. There is a need for effective tools to understand injury risk, and interaction of risk between mTBI and MSKI. We developed a data dashboard for model-based estimation of mTBI and MSKI in US SMs.
Methods: This data dashboard used data from the Military Health System Data Repository to create a career archival cohort of 777,811 US SMs who joined between 2016 and 2020, with follow-up data to 2023. Mild TBI and MSKI events were identified using ICD-10 codes. We fit logistic regression models to estimate bidirectional risk of MSKI or mTBI based on prior injury, and demographic and military-relevant covariates. MSKI predictions were conditional estimates in a propensity score-matched sample of 26,784. We then estimated marginal population-level probabilities from bidirectional models. We used the Python tkinter package to develop a data dashboard of model-based estimates of mTBI or MSKI risk, conditional on user-defined parameters such as age and branch of service.
Results: The average age of the cohort was 20.6±3.4 years, 19% were female, 2.4% had an mTBI, and 69.9% had an MSKI. Median and interquartile ranges (IQRs) were used as cut points for “low,” “medium,” and “high” risk of injury in the data dashboard. For the MSKI-to-mTBI model, median probabilities were 0.023 (IQR = 0.016-0.032). For the mTBI-to-MSKI model, median probabilities were 0.500 (IQR= 0.402-0.598). Figure 1 provides a data dashboard example of high risk of mTBI, after providing covariate values.
Conclusions: We developed a data dashboard for model-based estimation of mTBI and MSKI risk in US SMs. Our methods may be adapted for other injury surveillance and medical planning purposes. Data dashboards for injury risk can be flexible and scalable tools for supporting injury prevention strategies and informing patient counseling.
