Comparing traumatic brain injury diagnoses, intensive care utilization, length of stay, and costs between Department of Defense and Veterans Affairs hospitals: A long-term impact of military-relevant brain injury consortium study

Abstract: INTRODUCTION: The paucity of health economics studies on traumatic brain injury (TBI) was highlighted in a recent report from the National Academies of Sciences, Engineering, and Medicine. Resource and cost modeling of acute hospitalizations for Service Members and Veterans (SMVs) can be used to inform Department of Defense (DoD) and Department of Veterans Affairs (VA) policy. The objective of this study was to compare intensive care unit (ICU) utilization, length of stay (LOS), and cost between DoD and VA facility hospitalizations for SMVs with a primary diagnosis of TBI. MATERIALS AND METHODS: A unique database of ICU utilization, LOS, cost, and demographic variables was created from DoD and VA administrative databases for hospitalizations with a primary diagnosis code for TBI. Logistic and generalized linear models were used to estimate the adjusted differences in ICU utilization, LOS, and hospitalization cost between DoD and VA. All models were first adjusted for age, sex, and specific TBI diagnosis and then for clinical severity, as measured by the Medicare Severity Diagnosis-Related Group (MS-DRG). RESULTS: SMVs hospitalized in DoD facilities were younger (median age 44 vs 73) and more than double were female (10.7% vs 3.5%), relative to VA facilities. The most frequent diagnosis for SMVs in both systems was traumatic subdural hemorrhage without loss of consciousness (LOC) (15.2% in DoD vs 47.8% in VA). DoD-treated SMVs had a much higher frequency of LOC diagnoses: traumatic subdural hemorrhage with LOC, unspecified duration (14.4% vs 8.1%), traumatic subarachnoid hemorrhage with LOC, unspecified duration (11.5% vs 4.1%), and concussion with LOC, unspecified duration (8.2% vs 1.1%). The most frequent MS-DRG in DoD facilities was concussion without complications/comorbidities or major complications/comorbidities (16.1%), and traumatic stupor and coma <1 hour with complications/comorbidities in VA facilities (27.5%). In diagnosis-adjusted models, the odds of ICU utilization were higher (OR 2.19, 95% CI 1.47, 3.25), although LOS was lower (-3.67, 95% CI -5.43, -1.9) for SMVs treated in DoD facilities. Female SMVs had significantly lower diagnosis-adjusted LOS (-1.88, 95% CI -3.69, -.07). Unadjusted total hospitalization cost ($13,548 vs $23,084) was statistically significantly lower in DoD than VA facilities, but neither total nor daily marginal cost was statistically significantly different between the two systems in diagnosis-adjusted models. Similar results were found for MS-DRG adjusted models. CONCLUSION: This study found that DoD facilities treat SMVs with TBI more intensively and rapidly than the VA. SMVs hospitalized in DoD facilities were younger, more likely to be female, and had diagnoses associated with LOC. In contrast, SMVs hospitalized in VA facilities had much higher rates of MS-DRGs with complications and comorbidities. Female SMVs were found to have lower LOS, consistent with civilian hospitalizations. These findings are limited in that Service Members and Veterans are distinct groups in terms of the mechanism of injury and non-hospitalization comorbidities, which are not accounted for in hospitalization data. Despite limitations, the findings suggest that the DoD and VA, as two government systems, have similar marginal costs for treating TBI. Future studies should include SMV cohorts with data on the mechanism of injury, military characteristics, and non-hospitalization comorbidities.