Citation

Arango-Lasprilla JC, Ketchum JM, Cifu D, Hammond F, Castillo C, Nicholls E, Watanabe T, Lequerica A, Deng X. Arch. Phys. Med. Rehabil. 2010; 91(10): 1495-1504.

Affiliation

Department of Physical Medicine and Rehabilitation, Virginia Commonwealth University, Richmond, VA; Center for Rehabilitation Sciences and Engineering, Virginia Commonwealth University, Richmond, VA.

Copyright

(Copyright © 2010, Elsevier Publishing)

DOI

10.1016/j.apmr.2010.07.010

PMID

20875505

Abstract

Arango-Lasprilla JC, Ketchum JM, Cifu D, Hammond F, Castillo C, Nicholls E, Watanabe T, Lequerica A, Deng X. Predictors of extended rehabilitation length of stay after traumatic brain injury. OBJECTIVE: To develop a prediction rule for acutely identifying patients at risk for extended rehabilitation length of stay (LOS) after traumatic brain injury (TBI) by using demographic and injury characteristics. DESIGN: Retrospective cohort study. SETTING: Traumatic Brain Injury Model Systems. PARTICIPANTS: Sample of TBI survivors (N=7284) with injuries occurring between 1999 and 2009. INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURES: Extended rehabilitation LOS defined as 67 days or longer. RESULTS: A multivariable model was built containing FIM motor and cognitive scores at admission, preinjury level of education, cause of injury, punctate/petechial hemorrhage, acute-care LOS, and primary payor source. The model had good calibration, excellent discrimination (area under the receiver operating characteristic curve = .875), and validated well. Based on this model, a formula for determining the probability of extended rehabilitation LOS and a prediction rule that classifies patients with predicted probabilities greater than 4.9% as at risk for extended rehabilitation LOS were developed. CONCLUSIONS: The current predictor model for TBI survivors who require extended inpatient rehabilitation may allow for enhanced rehabilitation team planning, improved patient and family education, and better use of health care resources. Cross-validation of this model with other TBI populations is recommended.

Language: en