Advanced Therapy in Traumatic Brain Injury Inpatient Rehabilitation: Effects on Outcomes During the First Year after Discharge

Abstract Objective To use causal inference methods to determine if receipt of a greater proportion of inpatient rehabilitation treatment focused on higher level functions, for example, executive functions, ambulating over uneven surfaces (advanced therapy [AdvTx]), results in better rehabilitation outcomes. Design A cohort study using propensity score methods applied to the traumatic brain injury practice-based evidence (TBI-PBE) database, a database consisting of multisite, prospective, longitudinal observational data. Setting Acute inpatient rehabilitation facilities. Participants Patients enrolled in the TBI-PBE study (N=1843), aged 14 years or older, who sustained a severe, moderate, or complicated mild TBI, receiving their first inpatient rehabilitation facility admission to 1 of 9 sites in the United States, and consented to follow-up 3 and 9 months postdischarge from inpatient rehabilitation. Interventions Not applicable. Main Outcome Measures Participation Assessment with Recombined Tools-Objective-17, FIM motor and cognitive scores, Satisfaction with Life Scale, and Patient Health Questionnaire-9. Results Controlling for measured potential confounders, increasing the percentage of AdvTx during inpatient TBI rehabilitation was found to be associated with better community participation, functional independence, life satisfaction, and decreased likelihood of depression during the year after discharge from inpatient rehabilitation. Participants who began rehabilitation with greater disability experienced larger gains on some outcomes than those who began rehabilitation with more intact abilities. Conclusions Increasing the proportion of treatment targeting higher level functions appears to have no detrimental and a small, beneficial effect on outcome. Caution should be exercised when inferring causality given that a large number of potential confounders could not be completely controlled with propensity score methods. Further, the extent to which unmeasured confounders influenced the findings is not known and could be of particular concern due to the potential for the patient’s recovery trajectory to influence therapists’ decisions to provide a greater amount of AdvTx.