Predicting Functional Outcome with Disability Rating Scale Scores Following Primary Rehabilitation

Philip Schatz, Ph.D., Saint Joseph's University, Psychology Department
Frank Hillary, M.S., Drexel University University, Psychology Department
Douglas L. Chute, Ph.D., Drexel University University, Psychology Department

Abstract

Considerable research effort has been devoted to identify predictive factors of rehabilitation progress and outcome following traumatic brain injury (TBI). Factors identified as affecting outcome include severity of coma and disability, using measures such as the Glasgow Coma Scale and Glasgow Outcome Scale. Recent research by Schatz (1995) identified the Disability Rating Scale (DRS) at discharge from primary rehabilitation as having significant predictive ability for level of independence up to 9 years post injury. The present study was performed to determine which sub-scales of the Disability Rating Scale contributed to that prediction of rehabilitation outcome.

Medical records were analyzed for 153 moderate and severe TBI patients enrolled in the Pennsylvania Head Injury Program. Inclusion in the study was contingent upon having a measurable outcome rating at two or more years post-injury and a measurable DRS rating at discharge from primary rehabilitation. Sub-scales of the DRS include: Eye Opening; Best Verbal and Motor Response; Cognitive Ability to Feed, Groom, and Toilet; Dependence on Others; and Employability. Our dependent measure was Functional Independence Level (FIL), a ten-point measure of rehabilitation outcome which includes need for assistance with communication, self-care routines, basic living skills, and overall level of supervision. Post-hoc interrater reliability was significant for both the DRS (.90) and the FIL (.80) ratings. Stepwise multiple regression analysis identified the optimal combination of the "Grooming" and "Toileting" subscales of the DRS as having the sole significant predictive value, resulting in a multiple correlation of .69 and accounting for 47% of the total variance in FIL scores. The multiple regression equation was as follows:

FIL = Grooming (-8.78) + Toileting (-.41) + 8.03.

Results show that much of the variance associated with functional outcome following TBI can be predicted by ratings of a patient's ability to perform self-care routines at the time of their discharge from primary rehabilitation.


Methods

Subjects: Subjects were 300 individuals considered to be domiciled in the state of Pennsylvania at the time of their head injury who enrolled in the Pennsylvania Head Injury Program. Inclusion in the study was contingent upon:
  • documented evidence of a traumatic brain injury
  • measureable outcome rating at two or more years post-injury
  • measurable Disability Rating Scale rating at discharge from primary rehabilitation.

The above criteria resulted in the exclusion of 147 subjects, leaving a sample size of 153.

Procedures: Retrospective inspection of Pennsylvania Head Injury Program files was performed, coding Disability Rating Scale scores from acute care and rehabilitation hospital discharge summaries. Functional Independence Level scores were coded from post-acute rehabilitation hospital discharge summaries and/or Pennsylvania Head Injury Program case management reports. Where specific Disability Rating Scale or Functional Independence Level scores were not recorded, specific information within these documents was used to extrapolate Disability Rating Scale and Functional Independence Level scores. Sub-scales of 153 subjects' Disability Rating Scale scores were analyzed to determine their predictive ability of Functional Independence Level scores.


Results:

Stepwise Multiple Regression Results

IV's: Eye Opening, Verbalization, Motor Response, Feeding, Toileting, Grooming, Dependence, Psychosocial Adaptability

DV:Functional Independence Level

Variables in Equation: Grooming, Toileting

Multiple R .69249

R Square .47954

Adjusted R Square .47260

Standard Error 1.45778

 

Analysis of Variance

DF
Sum of Squares
Mean Square

Regression

2
293.70343
146.85171

Residual

150
318.76716
2.12511

F = 69.10297Signif F = .0000

Variables in the Equation (95% Conf)

Variable
B
SE B
Intrvl
B
Beta

GROOM

-.877338
.205965
-1.284305
-.470372
-.464313

TOILET

-.405368
.173446
-.748081
-.062655
-.254754

(Constant)

8.033035
.175326
7.686608
8.379463

Variable
T
Sig T

GROOM

-4.260
.0000

TOILET

-2.337
.0208

(Constant)

45.818
.0000


Conclusions:

Nearly half of the variability in long-term TBI patient outcomes were explained by ratings of functional ability at the time of discharge from primary rehabilitation.

Ability to groom and toilet one's self at the time of discharge emerged as the lone significant predictor of level of independence.

Injury and coma severity, length of loss of consciousness, hospital length of stay, and other measures of functional ability were not significant predictors of functional outcome.