Towards precision rehabilitation medicine after acquired brain injury: Exploring the prediction of patient independence using structured clinical data

Background: Acquired Brain Injury (ABI) affects millions globally each year, often resulting in complex and heterogeneous impairments. This diversity highlights the need for precision medicine in neurorehabilitation to tailor care to individual patient profiles. Structured clinical data may support the development of prediction models to guide such personalized care. Objective: This study uses structured clinical data from a measurement feedback system (MFS) to develop prediction models for patient outcome and treatment response in a specialized neurorehabilitation setting. Methods: Clinical data were prospectively collected using an MFS implemented in a specialized center for intensive neurorehabilitation in young patients with severe ABI and were re-used to develop cross-validated multivariate regression models. These models predicted levels of independence (Barthel Index) at admission, at three months post-admission, and the change in independence within this timeframe. Results: The study included a sample of 100 patients (38 % female, mean age 24.4 ± 5.6 years), mainly with TBI (62 %) and CVA (24 %). As expected, independence increased significantly within the first three months of neurorehabilitation (p < 0.001, Cohen's d = 0.79). The prediction models showed considerable performance in the prediction of independence at admission (R² = 65.7 %), after three months (R² = 59.3 %) and the change in independence (R² = 76.3 %). Conclusions: Structured clinical data derived from MFS integration provides a solid foundation for the development of representative in-house developed models to inform and shape care for the target population. This approach advances precision prognosis, an important component of precision medicine.