Tabrez Alam, Scott D Olson, Charles S Cox
Among older adults with mainly mild TBI, preinjury frailty and cognitive-mediated function were independently associated with poor 6-month outcome, yet these preinjury measures did not significantly improve prediction accuracy. Instruments such as the FAQ and GFI capture clinically meaningful preinjury heterogeneity and may inform the development of more precise oaTBI prognostic models.
The Federal Interagency Traumatic Brain Injury Research Informatics System (FITBIR) was designed as a centralized repository for all federally funded traumatic brain injury (TBI) datasets. FITBIR is intended to feature pre-harmonized data using specific common data elements, which can then be applied uniformly by all contributors. Yet, FITBIR is seldom used for exposure and outcomes research. This study evaluated whether FITBIR data could identify the impact of pre-injury immunomodulator exposure on post-injury outcomes. FITBIR forms were downloaded and found to contain a significant number of nonharmonized entries across studies, requiring extensive manual review and harmonization across covariates, exposures, and outcomes. Outcomes included functional, cognitive, psychiatric, and global measures. Among 52,897 participants, 204 had pre-injury immunomodulator use, and 90 had sufficient data for further analysis. All 90 exposed individuals were matched to controls, yielding a cohort of 180 participants. Least absolute shrinkage and selection operator and Firth regression were used to generate a propensity-matched cohort, with standardized mean differences of 0.62 (0.078 vs. 0.032) and 0.07 (0.078 vs. 0.073), respectively, due to the high dimensionality of the available demographic and historical covariates. Across all outcomes, except length of stay (+3.04 days in the immunomodulator group, p = 0.0454), no statistically significant differences were observed between groups for neurocognitive, psychiatric, and quality-of-life outcomes. This proof-of-concept analysis demonstrates that national TBI data can support the construction of matched cohorts and the evaluation of multiple outcome domains. Nonetheless, fragmented medication data, inconsistent coding, and missing longitudinal outcomes limited analytic power, underscoring the potential of multicenter databases for future research.