Model System:

TBI

Reference Type:

Journal

Accession No.:

J71812

Journal:


Archives of Physical Medicine and Rehabilitation

Year, Volume, Issue, Page(s):

, 96, 6, 994-999

Publication Website:

Abstract:

Study developed and evaluated the external validity of models to predict long-term survival in adults with traumatic brain injury (TBI). Survival/mortality data were obtained from 2 cohorts of long-term survivors of TBI: (1) the Traumatic Brain Injury Model Systems (TBIMS) cohort comprised 7,365 patients who were admitted to a TBIMS facility and were assessed at 1 year after injury, and (2) the California Department of Developmental Services (CDDS) cohort comprised 5,116 individuals who sustained a TBI and received long-term services from the CDDS. Analysis revealed that older age, male sex, and severity of disability in walking and feeding were significant predictors of increased long-term mortality rates. The CDDS model predicted 623 deaths for subjects in the TBIMS cohort, with an observed-to-expected ratio of .94. The TBIMS model predicted a total of 525 deaths for subjects in the CDDS cohort, with an observed-to-expected ratio of 1.08. Regression calibration statistics were satisfactory, and both models ranked survival times well from shortest to longest. Despite underlying differences in the CDDS and TBIMS, statistical models for mortality rates derived from each cohort accurately predicted survival in the other. The external validity of these models provides strong support for their use in practical prognostic work.

Author(s):


Brooks, Jordan C., Shavelle, Robert M., Strauss, David J., Hammond, Flora M., Harrison-Felix, Cynthia L.

Participating Centers: