Secondary Logo

Journal Logo

Institutional members access full text with Ovid®

The Accuracy of Artificial Neural Networks in Predicting Long-term Outcome After Traumatic Brain Injury

Segal, Mary E. PhD; Goodman, Philip H. MD, MS; Goldstein, Richard PhD; Hauck, Walter PhD; Whyte, John MD, PhD; Graham, John W. PhD; Polansky, Marcia ScD; Hammond, Flora M. MD

Section Editor(s): Kothari, Sunil MD

Journal of Head Trauma Rehabilitation: July-August 2006 - Volume 21 - Issue 4 - p 298–314

Objective This study compared the accuracy of artificial neural networks to multiple regression and classification and regression trees in predicting outcomes of 1644 patients in the Traumatic Brain Injury Model Systems database 1 year after injury.

Methods Data from rehabilitation admission were used to predict discharge scores on the Functional Independence Measure, the Disability Rating Scale, and the Community Integration Questionnaire

Results Artificial neural networks did not demonstrate greater accuracy in predicting outcomes than did the more widely used method of multiple regression. Both of these methods outperformed classification and regression trees

Conclusion Because of the sophisticated form of multiple regression with splines that was used, firm conclusions are limited about the relative accuracy of artificial neural networks compared to more widely used forms of multiple regression.

Research Center for Health Care Decision- making, Inc, Wyndmoor, Pa (Dr Segal); the University of Nevada, Reno (Dr Goodman); Harvard Medical School, Boston, Mass (Dr Goldstein); Thomas Jefferson University, Philadelphia, Pa (Dr Hauck); Moss Rehabilitation Research Institute, MossRehab Hospital, Philadelphia, Pa (Dr Whyte); Pennsylvania State University, University Park, Pa (Dr Graham); Drexel School of Public Health, Philadelphia, Pa (Dr Polansky); and Carolinas Rehabilitation, Charlotte, NC (Dr Hammond).

Corresponding author: Mary E. Segal, PhD, Research Center for Health Care Decision-making, Inc, 8200 Flourtown Ave, Suite 1-C, Wyndmoor, PA 19038 (e-mail:

This work was partially supported by the US Department of Education, grant #H133A70033 from the National Institute on Disability and Rehabilitation Research to the Moss Rehabilitation Research Institute, and by the TBI Model System National Data Center funded by the National Institute on Disability and Rehabilitation Research. We thank Jennifer Bogner, PhD, Tessa Hart, PhD, Mark Johnston, PhD, Jeffrey Kreutzer, PhD, and Thomas Novack, PhD, for many useful suggestions in reviewing manuscript drafts. Kenneth Wood and Slava Gavurin at the TBI Model Systems National Data Center provided the data set and descriptions of its coding and history, and Rebecca Sheppard at the Moss Rehabilitation Research Institute assisted with data management.

© 2006 Lippincott Williams & Wilkins, Inc.