1608 - Artificial Intelligence-Driven Insights and the Hidden Costs of Socioeconomic Barriers in Traumatic Brain Injury Rehabilitation: A Case Study

Case Diagnosis: The rehabilitation trajectory of a 28-year-old female who sustained a traumatic brain injury (TBI) from a 2020 motor vehicle accident, resulting in diffuse axonal injury (DAI) and left-sided hemiplegia, is evaluated juxtaposed with ScholarAI-predicted outcomes.

Case Description: Using the Section GG Self-Care (ADL) scale (6 = Independent, 5 = Setup, 2 = Substantial Assistance, 1 = Dependent), the patient's scores were as follows:

• 2021: Oral hygiene: 1, Upper body dressing: 2, Lower body dressing: 2, Shoes: 1, Toilet hygiene: 2, Bathing: 2, Toilet transfer: 2, Eating: 6.


• 2024: Oral hygiene: 5, Upper body dressing: 2, Lower body dressing: 1, Shoes: 1, Toilet hygiene: 2, Bathing: 2, Toilet transfer: 3, Eating: 6.

PT Assessments:

• 2021: Bed mobility: Supervision, Sit to stand: Max assist, Ambulation (10 ft, platform walker): Mod assist.


• 2024: Bed mobility: Mod assist, Sit to stand: Mod assist, Bed to chair transfer: Max assist, Ambulation (15 ft): Mod assist.

Using machine learning data from similar TBI cases, AI predicted setup assistance (score 5) for dressing, hygiene, and transfers by 2024 with consistent multidisciplinary rehabilitation. AI also projected ambulation distances of 50 ft with minimal assistance.

Discussions: AI-driven insights can offer promising recovery projections, yet reality can be shaped by socioeconomic barriers. Slower progress is demonstrated with a paired t-test revealing a significant 2-level gap on the Section GG scale for dressing and transfers (p < 0.05). Unlike much of the current literature, which primarily focuses on identifying disparities in access to care, this case suggests that AI could go beyond recognition to actively mitigate these disparities by dynamically adjusting treatment protocols.

Conclusions: AI has the potential to personalize rehabilitation, but this case demonstrates the disparity between predicted outcomes and real-world recovery. Bridging this gap requires AI-integrated rehabilitation models that adapt to patient needs in real-time while ensuring equitable access to advanced technologies.