Racial Differences in Employment Outcome After Traumatic Brain Injury at 1, 2, and 5 Years Postinjury

Article Outline

• Abstract
• Methods
  • Participants
  • Measures
    • Dependent variable
    • Independent variables
  • Statistical Analyses
• Results
  • Description of the Sample
  • Included Versus Excluded
  • Differences in Covariates Between Blacks and Whites (Step 1)
  • Effects of Demographic and Injury Characteristics on Employment Status and Changes in Employment Status Over Time (Steps 2 and 3)
  • Adjusted Model (Steps 4 and 5)
    • Comparison in employment status between blacks and whites at each follow-up year (aim 1)
    • Changes in employment status over time for race (aim 2)
    • Comparison of changes in employment status between races (aim 3)
    • Effects of demographic and injury characteristics on employment status postinjury and changes in employment status over time
• Discussion
  • Study Limitations
• Conclusions
• References
• Copyright

Abstract 

Gary KW, Arango-Lasprilla JC, Ketchum JM, Kreutzer JS, Copolillo A, Novack TA, Jha A. Racial differences in employment outcome after traumatic brain injury at 1, 2, and 5 years postinjury.

Objectives

To examine racial differences in competitive employment outcomes at 1, 2, and 5 years after traumatic brain injury (TBI) and to determine whether changes in not competitive employment rates over time differ between blacks and whites with TBI after adjusting for demographic and injury characteristics.

Design

Retrospective cohort study.

Setting

Sixteen TBI Model System Centers.

Participants

Blacks (n=615) and whites (n=1407) with moderate to severe TBI.

Interventions

Not applicable.

Main Outcome Measure

Employment status dichotomized as competitively employed versus not competitively employed.

Results

After adjusting for demographic and injury characteristics, repeated-measures logistic regression indicated that (1) the odds of not being competitively employed were significantly greater for blacks than whites regardless of the follow-up year (all P<.001); (2) the odds of not being competitively employed declined significantly over time for each race (P≤.004); and (3) changes over time in the odds of not being competitively employed versus being competitively employed were not different between blacks and whites (P=.070). In addition, age, discharge FIM and Disability Rating Scale, length of stay in acute and rehabilitation, preinjury employment, sex, education, marital status, and cause of injury were significant predictors of employment status postinjury.

Conclusions

Short- and long-term employment is not favorable for people with TBI regardless of race; however, blacks fare worse in employment outcomes compared with whites. Rehabilitation professionals should work to improve return to work for all persons with TBI, with special emphasis on addressing specific needs of blacks.

Key Words: Brain injuries, Employment, Rehabilitation

List of Abbreviations: CI, confidence interval, DRS, Disability Rating Scale, GCS, Glasgow Coma Scale, GED, general education development, GLMM, generalized linear mixed model, LOC, loss of consciousness, LOS, length of stay, OR, odds ratio, PTA, posttraumatic amnesia, TBI, traumatic brain injury, TBIMS, Traumatic Brain Injury Model Systems

IN THE UNITED STATES, TBI is one of the leading causes of disability among adults. More than 1.1% of the U.S. population (∼3.17 million residents) is estimated to have long-term or lifelong disability from TBI.¹ Annually, 1.4 million people sustain a TBI; of these, 79% are released after receiving treatment in emergency departments, 17% are hospitalized for further rehabilitation, and 4% die. In addition, 58% of survivors are of working age (15–74y).² Thus, a large proportion of adults who sustain a TBI will be released from the hospital and return to their communities in hopes of resuming employment.

Unfortunately, the physical, cognitive, and emotional consequences of TBI may present a substantial obstacle to recovery and community reintegration after injury. Decreased locomotion,³ memory and new learning problems,⁴ and depression⁵ are just a few deficits associated with TBI that can negatively affect function. These symptoms result in difficulty with independent living, educational and vocational pursuits, and creating social networks.⁶, ⁷, ⁸ One of the most recognized challenges after TBI is obtaining and maintaining employment. Studies have shown that employment is intricately linked to several aspects of a person's life that are of great importance, such as financial viability, social integration, and quality of life.⁹, ¹⁰, ¹¹

Several preinjury characteristics, such as age, education level, and prior employment, have been identified as strong predictors of employment outcomes after TBI.¹² For example, being older than 40 years at the time of injury is negatively correlated with return to work post-TBI.¹³ Gollaher et al¹⁴ noted that education is one of the most important predictors of employment postinjury, and those with greater than a high school education were more likely to experience vocational reentry after TBI. Additionally, those who were employed preinjury, compared with those unemployed preinjury, were 3 to 5 times more likely to return to work 1 to 5 years postinjury.¹³ These and many other demographic and injury characteristics (eg, sex, marital status, functional status) have been studied extensively in TBI employment outcomes research. However, despite statistics showing that minorities are disproportionately affected by TBI,¹⁵, ¹⁶, ¹⁷ fewer studies have documented the role of race/ethnicity in post-TBI employment outcomes.

Early studies of employment outcomes after TBI tended to have small sample sizes and conducted subanalyses to examine employment outcomes among black, Hispanic, and members of other minority groups.¹⁸, ¹⁹ Recently, especially over the past 6 years, TBI investigators have demonstrated an increased interest in the influence of race and ethnicity on functional and psychosocial outcomes, including employment. Most of these studies, however, were still limited by small samples, short follow-ups, ambiguous operationalizations of employment, and/or a focus on minority status as a whole, instead of examining specific subgroups (eg, blacks, Hispanics, Asians, Native Americans).

A review of the literature shows a few studies that were conducted with the objective of examining the effect of race/ethnicity on employment after TBI in samples of more than 100 people. Vanderploeg et al²⁰ found a significant race-by-region-by-LOC interaction on employment in a study of 4322 Vietnam-era army veterans (758 minorities, 3563 white). Analysis of simple effects appeared to indicate that minorities with a history of LOC had significantly lower rates of full-time employment than whites with LOC in the Midwest (50% vs 79%), Northwest (12% vs 72%), and South (48% vs 86%), but not in the West (86% vs 86%). Arango-Lasprilla et al²¹ examined the influence of race and ethnicity on competitive employment 1 year post-TBI in a nationwide sample of 5259 persons receiving care at a TBIMS center (1791 minorities vs 3468 white). After controlling for demographic and injury characteristics found to affect employment outcome, minorities were over 2 times more likely than their white counterparts to be unemployed versus competitively employed.

Two additional studies examined the influence of ethnicity on employment after TBI over time, or job stability.²², ²³ Kreutzer et al²² studied 186 persons from the TBIMS database (34% minorities, 66% nonminorities) and found that minorities were significantly less likely than nonminorities to be stably employed (19% vs 43%) and more likely to be unemployed (50% vs 31%) at 1, 2, and 3 or 4 years after TBI. Arango-Lasprilla et al²³ conducted a similar study of 627 persons from the TBIMS database (205 minorities, 422 white) using the same definition of job stability. After controlling for demographic and injury characteristics, the odds of being unemployed in comparison with being stably employed were nearly 3 times greater for minorities than whites. Furthermore, minorities' odds of being unstably employed versus stably employed and being unemployed versus unstably employed were over 2 times greater compared with whites.

Even though blacks have higher rates of TBI than whites and other minorities,¹⁵, ¹⁶, ¹⁷ only one study has examined the employment outcomes specifically for blacks after TBI rather than combining minorities into one category. Sherer et al²⁴ studied productivity in 1083 persons who had sustained a TBI (32% black, 13% other minorities, 55% white). Productivity was defined as those who were competitively employed at least part-time, full- or part-time students, or full-time homemakers. After controlling for preinjury productivity, education, cause of injury, sex, and age, blacks were over 2 times more likely than whites to be nonproductive, or noncompetitively employed, 1 year after injury.

Currently, there is a lack of knowledge of longer-term employment outcomes of blacks. Furthermore, given the heterogeneous makeup of minority groups and small sample sizes of previous longitudinal studies, it is not known whether results are generalized to blacks in particular. The TBIMS database provides an opportunity to examine the influence of race on employment outcomes in a nationally representative sample of persons who have sustained a TBI and been treated at one of the TBIMS centers across the country. To address some of the weaknesses of previous studies and further our knowledge of longer-term employment outcomes among blacks who have sustained a TBI, the present study has 3 primary aims: (1) to compare employment outcomes between blacks and whites at 1, 2, and 5 years postinjury; (2) to examine changes in employment over time within each racial group; and (3) to compare the changes in employment over time between blacks and whites. Final results will be independent of demographic and injury characteristics (covariates) that differ between blacks and whites, or affect employment or change in employment over time.

Back to Article Outline

Methods 

Participants 

The U.S. Department of Education, National Institute on Disability and Rehabilitation Research funds the TBIMS. The 16 level-1 trauma centers initiate care in the emergency department, followed by acute neurotrauma management and interdisciplinary inpatient rehabilitation, followed by long-term outpatient services.²⁵ In addition, the funded centers collect data for prospective, longitudinal, multicenter studies that examine numerous aspects of recovery and outcomes after TBI via a centralized database. Eligibility criteria to be included in the database are PTA greater than 24 hours, trauma-related intracranial neuroimaging abnormalities, LOC exceeding 30 minutes (unless due to sedation or intoxication), or GCS in the emergency department of less than 13 (unless due to intubation, sedation, or intoxication). GCS is a neurologic scale that assesses level of consciousness after brain injury. Scores between 3 and 8 indicate severe injuries; 9 to 12, moderate injuries; and 13 to 15, mild injuries.²⁶ Enrollment is limited to those 16 years or older at the time of injury.

A sample of 2395 participants was initially selected from the TBIMS database based on the following criteria: (1) 18 to 60 years of age at injury, (2) classified as black or white, and (c) injury occurred between January 1, 1989 and December 31, 2001. The end date of December 31, 2001 was chosen because 5 years of follow-up would have been due by December 31, 2006 and available in the database.

In order to assess employment status 1 to 5 years postinjury, the participants used in this analysis must have had at least 1 year of follow-up information available for analysis. A total of 373 (15.6%) of the 2395 subjects could not be included in this analysis because information was missing from all 3 time points. This left 2022 subjects for analysis.

Measures 

The measures and modeling types for the independent and dependent variables used in this study are summarized below.

The dependent variable used for this analysis was a categorical variable indicating whether the subject was competitively employed (only subjects categorized as engaging in paid employment full-time or part-time) or not competitively employed (unemployed, full-time student, part-time student, special education, homemaker, volunteer work, retired, and others). This measure was collected from participants at 1, 2, and 5 years postinjury.

A definition of each independent variable that was included in the model as covariates is provided in table 1. The primary independent variable was a categorical variable indicating whether the participants self-report that they were black or white. Other demographic variables available for analyses included age at injury, sex, preinjury employment status, preinjury level of education, and preinjury marital status. With the exception of age, all other demographic variables were categorical and, for the purpose of this analysis, were dichotomized. Preinjury employment status was dichotomized into competitively employed and not competitively employed, in the same manner as the dependent variable. Education data taken from the database were dichotomized into less than high school (grade 8 or less or grades 9–11) or high school or more (GED, GED/high school, high school, trade school, high school diploma, some college, Associate's degree, Bachelor's degree, Master's degree, Doctoral level degree). Marital status was dichotomized as married or not married (single, divorced, separated, widowed). All demographic information was obtained during interviews based on self-report of the injured person.

Table 1. Description of Demographic and Injury Characteristics (Covariates)

Covariates	Description
Demographic
Age at injury	Classification of age in years at time of injury
Sex	Classification of sex preinjury categorized as male or female
Preinjury employment	Status of preinjury employment categorized as competitive employment or not competitive employment
Preinjury education	Level of education preinjury categorized as eighth grade or less, grades 9 through 11, GED/high school, trade school, high school diploma, some college, Associate's degree, Bachelor's degree, Master's degree, Doctoral degree
Preinjury marital status	Status of long-term union/partnership categorized as single, divorced, separated, and widowed
Race	Self-report race categorized as black or white
Injury
Cause of injury	Cause of injury categorized as violent or nonviolent
PTA	Duration of being disoriented in days
GCS at admission	Combination of eye opening, verbal, and motor response when emerging from coma ranging from 3 (lowest) through 15 (highest)
FIM at admission	Measure of independence based on what a subject does at admission to rehabilitation ranging from total assist (1) to independent (7)
FIM at discharge	Measure of independence based on what a subject does at discharge from rehabilitation ranging from total assist (1) to independent (7)
DRS at admission	Measure taken at admission to rehabilitation that assesses changes in function over course of recovery from coma to community discharge ranging from 0 (no disability) to 29 (vegetative state).
DRS at discharge	Measure taken at discharge from rehabilitation that assesses changes in function over course of recovery from coma to community discharge ranging from 0 (no disability) to 29 (vegetative state).
Acute LOS	LOS in acute hospitalization measured in days
Rehabilitation LOS	LOS in inpatient rehabilitation measured in days

Measures of injury characteristics obtained from the database based on medical records include cause of injury, PTA, GCS at admission, FIM at admission, FIM at discharge, DRS at admission, DRS at discharge, LOS acute, and LOS rehabilitation. Cause of injury was the only categorical injury characteristic and was dichotomized as violent (gunshot wound, assault with blunt instrument, other violence) or not violent (vehicular, sports related, fall, pedestrian accident). All other injury-related characteristics were continuous. PTA and LOS were measured in days. GCS is a discrete, continuous variable with a range from 3 (lowest) to 15 (highest). The FIM and DRS are ordinal and psychometrically sound assessment scales. The FIM measures level of independence for 18 items categorized into activities of daily living, bladder and bowel care, locomotion, transfers, and cognition with scores ranging from 1 (total assistance) to 7 (complete independence) for a total range of 18 (lowest) to 126 (highest).²⁷, ²⁸ Reliability for the FIM was reported as .86 to .97.²⁹ The DRS consists of 8 items on a 30-point scale, where lower numbers denote higher levels of function, and the scale measures patients' abilities from a coma state to activities in the home or community.³⁰ The DRS demonstrates a good interrater agreement of .98.³¹, ³²

Statistical Analyses 

All statistical analyses were conducted using SAS version 9.2.^a A total of 373 participants would have been included in the analysis if their employment status had not been missing at all 3 follow-up years. These participants were compared with the 2022 included participants with respect to demographic and injury characteristics using chi-square analyses for categorical variables and t tests for continuous variables.

To test the hypotheses of interest, a repeated-measures logistic regression model was fit to model the unemployment rates over time using the framework of a GLMM.³³ The mixed model included effects for follow-up year, race, and the 2-way interaction between race and year. To correctly understand the effects of race on competitive employment and changes in competitive employment over time, the model should adjust for important demographic and injury characteristics (covariates) that were significantly different among the races (because the design was not randomized), as well as those that may affect competitive employment and changes in competitive employment over time.

Because there were a variety of potential covariates and interaction effects that could be added to this model, model-building strategies similar to those used for logistic regression as outlined by Hosmer and Lemeshow³⁴ were used. The steps are briefly outlined as follows: (1) t tests and chi-square tests were conducted to identify differences between blacks and whites with respect to the covariates. Any covariates demonstrating differences between the races (P<.05) were considered for the adjusted model. (2) The relationship between each covariate and unemployment rates over time was modeled using GLMMs. The model included the main effects for the follow-up year and the covariate of interest. Any covariate with P less than .25 was considered for the adjusted model. (3) The relationship between each covariate and the change in unemployment rates over time was modeled using GLMMs. The model included the main effects for the follow-up year and the covariate of interest, as well as the follow-up year by covariate 2-way interaction effect. Any interaction effect with P less than .10 was considered for the adjusted model. (4) The adjusted model was initially fit with effects for race, follow-up year, the race by follow-up year interaction effect, and all potential covariates and interaction effects obtained from the first 3 steps. (5) Any covariate (if it was not included in a significant interaction) or covariate by follow-up year interaction effect that no longer contributed to the fit of the model (P<.05) was removed in a backwards selection manner.

Back to Article Outline

Results 

Description of the Sample 

The demographic and injury characteristics of the sample of 2022 subjects are summarized without regard to race in the third column of table 2. Most of the subjects were white (69.9%), male (75.5%), employed preinjury (72.0%), had at least a high school education (72.1%), and were not married preinjury (71.1%). There was a large degree of missing data for PTA (n=564), admission GCS (n=399), and admission (n=162) and discharge (n=200) FIM. Thus, to avoid biasing the final conclusions, these variables were not included in the adjusted analysis.

Table 2. Demographic and Injury Characteristics of Sample

Characteristics	Black (n=615)	White (n=1407)	Overall (N=2022)	Comparison
	Mean ± SD	Mean ± SD	Mean ± SD	T (df)	P
Age at injury (y) (N=2022)	34.90±10.82	33.68±11.86	34.05±11.56	2.19(2020)	.029
PTA (d) (n=1458)	27.27±21.30	29.87±27.09	29.14±25.61	1.75(1456)	.081
Admission GCS (n=1623)	9.23±4.12	8.27±4.16	8.61±4.17	4.49(1621)	<.001
Admission FIM (n=1860)	55.80±25.72	55.28±26.67	55.44±26.38	.39(1858)	.694
Discharge FIM (n=1822)	95.75±21.18	97.08±23.18	96.68±22.60	1.16(1820)	.246
Admission DRS (n=1979)	12.80±5.35	12.42±5.74	12.54±5.62	1.39(1977)	.169
Discharge DRS (n=1992)	6.20±3.28	5.87±3.93	5.97±3.75	1.79(1990)	.074
LOS acute (d) (N=2022)	20.22±14.68	22.22±17.39	21.61±16.63	2.49(2020)	.013
LOS rehabilitation (d) (N=2022)	28.83±19.81	31.84±31.13	30.93±27.43	2.27(2020)	.023

	Count	%	Count	%	Count	%	χ2 (df)	P
Sex (N=2022)
Men	493	80.16	1033	73.42	1526	75.47	10.51(1)	<.001
Women	122	19.84	374	26.58	496	24.53
Preinjury employment (n=1999)
Competitively employed	371	61.12	1069	76.80	1440	72.04	51.56(1)	<.001
Not competitively employed	236	38.88	323	23.20	559	27.96
Preinjury level of education (n=1955)
Less than high school	222	37.00	323	23.84	545	27.88	35.83(1)	<.001
High school or more	378	63.00	1032	76.16	1410	72.12
Preinjury marital status (n=2020)
Married	123	20.03	460	32.72	583	28.86	33.49(1)	<.001
Not married	491	79.97	946	67.28	1437	71.14
Cause of injury (n=2005)
Not violent	376	61.24	1259	90.51	1635	81.55	242.56(1)	<.001
Violent	238	38.76	132	9.49	370	18.45

Included Versus Excluded 

The 373 participants who were excluded because of missing employment information at all 3 follow-up visits were compared with the 2022 participants used for analyses. There was no evidence that the included and excluded groups differed significantly with respect to age at injury, sex, preinjury marital status, PTA, discharge FIM, discharge DRS, acute LOS, or rehabilitation LOS (all P≥.133). Included and excluded subjects were significantly different with respect to race (P=.045), employment at admission (P=.002), preinjury level of education (P<.001), cause of injury (P=.038), admission GCS (P=.036), admission FIM (P=.001), and admission DRS (P=.004). Specifically, the group of excluded subjects had a lower percentage of blacks (35.7% vs 30.4%), a lower percentage preinjury of competitive employment (63.9% vs 72.0%), a greater percentage preinjury with less than a high school level of education (39.0% vs 27.9%), a greater percentage of violent injuries (23.1% vs 18.4%), a higher mean admission GCS (9.2 vs 8.6), a higher mean admission FIM (60.6 vs 55.4), and a lower mean admission DRS (11.6 vs 12.5) than the group of included subjects.

Differences in Covariates Between Blacks and Whites (Step 1) 

The demographic and injury characteristics of the sample of 2022 subjects are summarized by race in the first 2 columns of table 2 and compared in the last column of table 2. No significant differences were identified between blacks and whites with respect to PTA, admission FIM, discharge FIM, admission DRS, and discharge DRS. Blacks and whites were significantly different with respect to age at injury, sex, preinjury employment status, preinjury level of education, preinjury marital status, admission GCS, acute LOS, rehabilitation LOS, and cause of injury. Specifically, blacks were older (34.9 vs 33.7y) and had a greater percentage of males (80.2% vs 73.4%), a greater percentage preinjury of not competitive employment (38.9% vs 23.2%), a greater percentage with less than a high school education (37.0% vs 23.8%), a greater percentage who were not married (80.0% vs 67.3%), had higher admission GCS scores (9.2 vs 8.3), had a shorter acute LOS (20.2 vs 22.2d), a shorter rehabilitation LOS (28.8 vs 31.8d), and a greater percentage of violent injuries (38.8% vs 9.5%) as compared with whites.

Effects of Demographic and Injury Characteristics on Employment Status and Changes in Employment Status Over Time (Steps 2 and 3) 

The distribution of employment status for each follow-up year is summarized in table 3 overall and separately for blacks and whites. These raw proportions do not reflect adjustment for relevant demographic and injury characteristics. Race, age at injury, sex, preinjury employment status, preinjury level of education, preinjury marital status, PTA, admission GCS, admission and discharge FIM, admission and discharge DRS, discharge DRS, acute and rehabilitation LOS, and cause of injury were at least marginal univariate predictors (ie, all P<.25) of employment status postinjury. Sex, preinjury level of education, admission FIM, and cause of injury were not significant univariate predictors of changes in employment status over time (ie, all P≥.10), while race, age at injury, preinjury employment status, preinjury marital status, PTA, admission GCS, discharge FIM (P=.094), admission and discharge DRS, and acute and rehabilitation LOS were at least marginal predictors (ie, all P<.10).

Table 3. Observed Employment Status 1, 2, and 5 Years Postinjury by Race

Employment Status	1-Year Follow-Up		2-Year Follow-Up		5-Year Follow-Up
	Count	%	Count	%	Count	%
Black
Competitively employed	82	17.12	94	21.51	89	21.81
Not competitively employed	397	82.88	343	78.49	319	78.19
White
Competitively employed	461	37.79	432	41.66	457	51.12
Not competitively employed	759	62.21	605	58.34	437	48.88
Overall
Competitively employed	543	31.96	526	35.69	546	41.94
Not competitively employed	1156	68.04	948	64.31	756	58.06

Adjusted Model (Steps 4 and 5) 

A repeated-measures logistic regression model was fit with effects for follow-up year, race, and the 2-way interaction between race and year. In order to correctly understand the effects of race on competitive employment and changes in competitive employment over time, the model also adjusted for the demographic and injury characteristics that were significantly different between blacks and whites, as well as those that may have affected postinjury competitive employment and changes in postinjury competitive employment over time. As indicated earlier, there was a high degree of missing data for PTA, admission GCS, admission FIM, and discharge FIM. To avoid biasing the sample, these variables were not considered in the adjusted model. We feel that admission and discharge DRS, cause of injury, and LOS for acute care and rehabilitation will adequately control for the injury characteristics in the adjusted model.

Two-sample t tests and chi-square tests indicated that all of the demographic and injury characteristics are significantly different between the races (P≤.029) except for admission and discharge DRS. Univariate repeated-measures logistic regression models indicated that all of the demographic and injury characteristics significantly affected postinjury employment status (P≤.244). The univariate repeated-measures models further showed that age at injury, preinjury employment status, preinjury marital status, admission DRS, discharge DRS, acute LOS, and rehabilitation LOS significantly affected changes in postinjury employment status over time (P≤.051). Thus, additional effects for each of the demographic and injury characteristics were initially included in the model, as well as interactions between year and each of the covariates: age at injury, preinjury employment status, preinjury marital status, admission DRS, discharge DRS, acute LOS, and rehabilitation LOS.

The interaction effects for year by acute LOS (P=.339), preinjury marital status (P=.102), rehabilitation LOS (P=.077), discharge DRS (P=.079), and preinjury employment status (P=.087) did not significantly contribute to the fit of the adjusted model and were removed using backwards selection (step 4). The final model then included main effects for year, race, age at injury, sex, preinjury employment status, preinjury education level, preinjury marital status, admission DRS, discharge DRS, acute LOS, and rehabilitation LOS, as well as interaction effects for year by race and year by age at injury. The effect tests for the final adjusted model are summarized in table 4.

Table 4. Effect Tests for Adjusted Model

Abbreviations: DDF, denominator degrees of freedom; NDF, numerator degrees of freedom.

Blacks had significantly greater odds of not being competitively employed versus being competitively employed as compared with whites at 1 year (OR=2.61; 95% CI, 1.93–3.53), 2 years (OR=2.10; 95% CI, 1.56–2.83), and 5 years (OR=3.15; 95% CI, 2.30–4.30) postinjury. These 3 differences remain significant after adjusting for multiple comparisons using a Bonferroni correction (α=.05/3=.017) (all P<.001).

For blacks, the odds of not being competitively employed versus being competitively employed were 2.35 times greater at 1 year postinjury than at 2 years postinjury (95% CI, 1.43–3.88), 2.46 times greater at 2 years postinjury than at 5 years postinjury (95% CI, 1.46–4.14), and 5.78 times greater at 1 year postinjury than at 5 years postinjury (95% CI, 3.43–9.76). For whites, the odds of not being competitively employed versus being competitively employed were 1.89 times greater at 1 year postinjury than at 2 years postinjury (95% CI, 1.23–2.92), 3.69 times greater at 2 years postinjury than at 5 years postinjury (95% CI, 2.30–5.92), and 6.98 times greater at 1 year postinjury than at 5 years postinjury (95% CI, 4.36–11.19). These 6 differences remain significant after adjusting for multiple comparisons using a Bonferroni correction (α=.05/6=.008) (all P≤.004).

There was no evidence of a significant interaction effect between races and year (F_2,2289=2.66, P=.070). Thus, the odds of not being competitively employed versus being competitively employed for blacks as compared with whites did not change significantly over time. Similarly, the changes over time in the odds of not being competitively employed versus being competitively employed were not different between blacks and whites.

The final adjusted model further indicated that age at injury, sex, preinjury employment status, preinjury level of education, preinjury marital status, admission DRS, discharge DRS, acute LOS, rehabilitation LOS, and cause of injury also had significant effects on postinjury employment status. Furthermore, age at injury had an effect on changes in postinjury employment status over time.

The odds of not being competitively employed postinjury versus being competitively employed were significantly higher for females than for males (OR=1.26; 95% CI, 1.01–1.57), for those who were not competitively employed preinjury than for those who were competitively employed (OR=3.39; 95% CI, 2.65–4.32), for those with less than a high school education preinjury than for those with at least a high school education (OR=2.34; 95% CI, 1.86–2.94), for those who were not married preinjury than for those who were married (OR=1.39; 95% CI, 1.11–1.74), and for those with violent injuries than for those with nonviolent injuries (OR=1.45; 95% CI, 1.09–1.94).

With respect to continuous variables, comparisons were made between the 25th and 75th percentiles of the sample. The odds of not being competitively employed postinjury versus being competitively employed were significantly higher for those with higher (17) admission DRS scores than lower (8) admission DRS scores (OR=1.43; 95% CI, 1.15–1.77), for those with higher (7) discharge DRS scores than lower (4) discharge DRS scores (OR=1.51, 95% CI, 1.33–1.72), for those with longer (28d) acute LOS than shorter (10d) acute LOS (OR=1.80; 95% CI, 1.55–2.09), and for those with longer (38d) rehabilitation LOS than shorter (14d) rehabilitation LOS (OR=1.27; 95% CI, 1.11–1.46).

Finally, with respect to age at injury, there was a significant interaction between year and age. Thus, the odds of not being competitively employed versus being competitively employed for older participants as compared with younger participants changed significantly over time. In general, the odds of not being competitively employed versus being competitively employed increased with age. Specifically, the odds of not being competitively employed versus being competitively employed were greater for older participants (43y) than for younger participants (24y) at 1 year (OR=1.60; 95% CI, 1.29–1.97), 2 years (OR=2.03; 95% CI, 1.63–2.53), and 5 years (OR=3.40; 95% CI, 2.67–4.32) postinjury. These 3 differences remain significant after adjusting for multiple comparisons using a Bonferroni correction (α=.05/3=.017) (all P<.001). The final adjusted model is summarized in table 5.

Table 5. Adjusted ORs for Not Competitively Employed Versus Competitively Employed at Follow-Up

Back to Article Outline

Discussion 

The first aim of the present study was to examine racial differences in employment outcomes at 1, 2, and 5 years post-TBI. It was hypothesized that as compared with whites, blacks with TBI were more likely to not be competitively employed at 1, 2, and 5 years postinjury. Our findings confirm the hypothesis that blacks had significantly greater odds of not being competitively employed as compared with whites at all follow-up years. These findings coincide with previous studies that have shown blacks are more likely to have poorer employment outcomes at 1 to 4 years post-TBI compared with whites.²¹, ²², ²³, ²⁴ It is important to know that the unemployment rate for blacks in the uninjured U.S. population is nearly 2 times greater compared with whites (12.6% vs 6.9%).³⁵ However, results of the present study illustrate that even after controlling for demographic (eg, preinjury employment status) and injury characteristics (eg, DRS at admission), the odds of not being competitively employed for blacks are 2.91 times greater at year 1, 2.10 at year 2, and 3.15 at year 5 postinjury.

As in the present study, most of the studies that have found racial differences in employment outcomes used the TBIMS database. Even though the results of the present study are similar to others using TBIMS data, there are some methodological differences. Most importantly, the length of years postinjury employment data that were examined extended beyond previous research. All of the previous studies investigated racial differences in employment outcomes between 1 and 4 years postinjury.²¹, ²², ²³, ²⁴ To our knowledge, the present investigation is the first of its kind to examine racial differences in employment outcomes at 1, 2, and 5 years post-TBI. This study introduced new knowledge, which suggests blacks are more likely than whites to be unemployed even up to 5 years postinjury.

The second important distinction is the focus on blacks compared with other minorities. This study's primary focus was on employment outcomes specifically for blacks after TBI. Arango-Lasprilla,²¹, ²³ Kreutzer, and colleagues²² reported worse employment outcomes for minorities, but their minority samples consisted of blacks, Hispanics, Native Americans, and Asians. On the other hand, Sherer et al²⁴ recognized the necessity of distinguishing blacks from other minority subgroups when examining broader components of employment outcomes after TBI. Although it is extremely important to explore clinical outcomes for all minorities that experience TBI, our study was limited to blacks because they are more affected by TBI compared with other minorities,¹⁵, ¹⁶, ¹⁷ the sample sizes in the other individual minority groups were not large enough for powerful statistical comparisons, and there was a large degree of missing data for the other minority groups at year 5 postinjury.

Third, different definitions are used for employment. In the study by Sherer,²⁴ employment was included in the definition of productivity, which consisted of those competitively employed, students, and homemakers. As in the present study, Kreutzer²² and Arango-Lasprilla²³ focused on competitive employment; however, it was included in their operationalization of job stability. They categorized job stability as stable (competitive employment at all follow-up periods), unstable (mixture of competitive employment and unemployment over follow-up visits), and unemployed (unemployed all follow-up periods). As in a recent study by Arango-Lasprilla,²¹ the outcome variable used here was also categorized into those who were competitively employed versus others who were not considered to be competitively employed, such as the unemployed, students, homemakers, and volunteers. The disadvantage of doing this is that some may argue that students and homemakers should be in the category of competitive employment. After all, students may be working toward entering the workforce, and homemakers may be just as productive around the house as those competitively employed outside the home. However, there are distinct factors that justify excluding them from competitive employment. Those that have difficulty returning to competitive employment can have different concerns than those returning to school or homemaking. For instance, loss of financial income may be a greater concern for those who were competitively employed preinjury than for those who were homemakers and in school. Also, job loss can be typical for persons after TBI, whereas students have opportunities to reenroll in school, and homemakers do not have to worry about financially related job loss.

The fourth distinction between the present study and previous research is the difference in sample sizes. Previous studies that investigated employment outcomes for minorities with TBI had total sample sizes ranging from 1083 to 5259 subjects at 1 year post-TBI,²¹, ²⁴ and those that extended their follow-up to 4 years postinjury had sample sizes from 186 to 633 subjects.²¹, ²² The present study used a substantially large sample of 2022 subjects secured from the TBIMS for employment outcomes up to 5 years postinjury.

Another unique characteristic of the present study compared with previous work is the difference within the longitudinal sample used when the subjects were followed up beyond 1 year. The inclusion criteria in Kreutzer²² and Arango-Lasprilla²³ required that participants have nonmissing employment information at admission, and have nonmissing employment information available for follow-up years 1, 2, and 3 or 4 (in order to define job stability). Although this sample used a longitudinal sample from the TBIMS as well, the present study's inclusion criteria for follow-up were different. The present study required that participants have a report of employment data for at least 1 of the 3 follow-up intervals. Here approximately 16% of the sample was missing employment information for year 1, 27% for year 2, and 36% for year 5. The benefit of the GLMM is that subjects need not be excluded if they are missing some data, as long as at least 1 year of follow-up information is known. The present study had approximately 16% of the sample missing all 3 follow-up years of employment data and could not be included. An analysis was provided comparing the excluded sample with the included sample and did identify statistical differences with respect to race, preinjury employment and education levels, cause of injury, and admission GCS, FIM, and DRS. These differences do bias the results somewhat, and this should be considered when interpreting these results. Nonetheless, the benefit of the longitudinal nature of the design with a relatively small percentage of missing data allowed for keen examination of the relationship between race and employment and the effect of other predictors and any change over time.

The second aim of this study was to examine changes in competitive employment over time within each race goup. Within each race, there is a gradual improvement in the probability of gaining competitive employment status from years 1 to 2 and 5. This concurs with previous longitudinal studies that have found that for the general TBI population, employment rates after injury steadily increase over time.³⁶, ³⁷, ³⁸

The third aim of this study was to determine whether the changes over time in the odds of no competitive employment versus competitive employment differ significantly between blacks and whites, after adjusting for demographic and injury characteristics. No significant difference between the groups was detected. Our results shed a positive light on employment outcomes after TBI over time for blacks and whites by showing similar gains in competitive employment for both groups. However, there is still a caveat in regards to improvement for blacks. To move towards diminishing racial disparities in employment outcomes after TBI, blacks' competitive employment rates are going to have to improve at a much faster pace than whites to essentially reduce the current gap that appears to persist in long-term employment outcomes.

For the analyses of demographic and injury characteristics found to affect employment post-TBI, age, total discharge FIM, discharge DRS, LOS acute, LOS rehabilitation, preinjury employment, sex, education, marital status, and cause of injury were significant predictors of employment status postinjury. These results provide additional support for past studies that have suggested the same demographic and injury characteristics mentioned above are significant predictors of employment post-TBI.⁷, ⁸, ¹², ¹⁴, ¹⁹ Of these, the effect of age significantly increased over time. Moreover, the likelihood of returning to work post-TBI over time (1–5y postinjury) is less for persons who are older compared with those who are younger.

The results of this study suggest that special attention needs to be given to blacks, and potentially other members of minority groups, in terms of vocational intervention. This attention is warranted over the long-term, not just in the first 6 months after injury. It is possible that because of cultural factors or other unknown factors, implementation of existing vocational programs may not be as effective for blacks experiencing TBI. As an initial step, it is essential that awareness of vocational rehabilitation services increase among minority members. Sykes-Horn et al³⁹ found that blacks were significantly less likely to know about vocational rehabilitation services after TBI than whites. The reason for this difference was not established in the study, but it implies that existing methods of accessing services after TBI may not be as effective for blacks. The variability in vocational rehabilitation services across sites, as noted by Hart et al,⁴⁰ may also undermine attention to special groups such as blacks. Federal monies often fund vocational rehabilitation programs, but the development of programs is left to individual states, which understandably means that primary efforts are focused on those who derive the quickest benefit as a means of justifying the expenditures. More difficult cases, which may encompass blacks, may not garner the same attention.

Study Limitations 

The present study has some limitations, and the results should be interpreted with caution. All the participants of the present study received acute and inpatient rehabilitation, which may limit generalizability of these results when considering persons who do not receive such services. Blacks with TBI were the primary focus of this study. Therefore, these findings are not generalizable to other minority subgroups such as Hispanics, Native Americans, and Asians. In addition, other factors related to employment were not extensively measured or could not be controlled, such as annual income, postdischarge medical care, insurance limitations, medication usage, and concomitant medical disorders. Lastly, subjects who were missing employment data for all of the follow-up years could not be included in this analysis. This may have biased the sample to some degree because loss to follow-up has been associated with a variety of demographic and injury characteristics.

There is a growing concern about racial and ethnic disparities in health care, and this study points to the need for continued racial and ethnic disparity research on employment outcomes (eg, employment outcomes beyond 5y, job stability beyond 4y, and employment outcomes in other racial groups), as well as other aspects of health and rehabilitation care delivery. Future mixed methods research may help identify both environmental and personal factors that also affect employment outcomes after TBI, such as access to transportation or motivation to seek employment. These efforts could potentially identify modifiable risk factors for lack of competitive employment in the TBI population and suggest specific interventions to improve employment outcomes.

Back to Article Outline

Conclusions 

The present investigation is the first of its kind to examine racial differences in employment outcomes at up to 5 years post-TBI. This study adds to the growing TBI racial disparities literature, showing that blacks are less likely to be employed at 1, 2, and 5 years post-TBI, even after controlling for demographic or injury characteristics. Clearly, many persons with TBI encounter challenges with return to employment after injury.⁴¹ This study demonstrates that compared with whites, the risk of not obtaining competitive employment 1 year post-TBI is higher for blacks even after accounting for demographic and injury characteristics, and these inequities continue long-term. Although there is evidence that competitive employment rates progress over time for both races, the existing gap suggests that more strategies need to be implemented to reduce disparate employment outcomes. Researchers in the area of TBI must continue to explore and document the impact race and ethnicity has on employment and other postinjury outcomes. Then they must move beyond the first step of identifying disparities towards addressing these differences through effective, culturally based interventions.

Supplier

Back to Article Outline

References 

1. Zaloshnja E, Miller T, Langlois JA, Selassie AW. Prevalence of long-term disability from traumatic brain injury in the civilian population of the United States, 2005. J Head Trauma Rehabil. 2008;23:394–400
   • View In Article
   • CrossRef
2. Langlois JA, Rutland-Brown W, Thomas KE. Traumatic brain injury in the United States: emergency department visits, hospitalizations, and deaths. Atlanta: Centers for Disease Control and Prevention, National Center for Injury Prevention and Control; 2006;
   • View In Article
3. Vallee M, McFadyen BJ, Swaine B, Doyon J, Cantin JF, Dumas D. Effects of environmental demands on locomotion after traumatic brain injury. Arch Phys Med Rehabil. 2006;87:806–813
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (400 KB)
   • CrossRef
4. Chu BC, Millis S, Arango-Lasprilla JC, Hanks R, Novack T, Hart T. Measuring recovery in new learning and memory following traumatic brain injury: a mixed-effects modeling approach. J Clin Exp Neuropsychol. 2007;29:617–625
   • View In Article
   • CrossRef
5. Draper K, Ponsford J, Schonberger M. Psychosocial and emotional outcomes 10 years following traumatic brain injury. J Head Trauma Rehabil. 2007;22:278–287
   • View In Article
   • CrossRef
6. Fleminger S, Ponsford J. Long term outcome after traumatic brain injury. BMJ. 2005;331:1419–1420
   • View In Article
   • CrossRef
7. Hawley CA, Ward AB, Magnay AR, Mychalkiw W. Return to school after brain injury. Arch Dis Child. 2004;89:136–142
   • View In Article
   • CrossRef
8. Cifu DX, Keyser-Marcus L, Lopez E, et al. Acute predictors of successful return to work 1 year after traumatic brain injury: a multicenter analysis. Arch Phys Med Rehabil. 1997;78:125–131
   • View In Article
   • Abstract
   • Full-Text PDF (993 KB)
   • CrossRef
9. Finset A, Dyrnes S, Krogstad JM, Berstad J. Self-reported social networks and interpersonal support 2 years after severe traumatic brain injury. Brain Inj. 1995;9:141
   • View In Article
   • MEDLINE
   • CrossRef
10. Johnstone B, Mount D, Schopp LH. Financial and vocational outcomes 1 year after traumatic brain injury. Arch Phys Med Rehabil. 2003;84:238–241
    • View In Article
    • Abstract
    • Full-Text PDF (114 KB)
    • CrossRef
11. O'Neill J, Hibbard MR, Brown M, et al. The effect of employment on quality of life and community integration after traumatic brain injury. J Head Trauma Rehabil. 1998;13:68–79
    • View In Article
    • MEDLINE
    • CrossRef
12. Shames J, Treger I, Ring H, Giaquinto S. Return to work following traumatic brain injury: trends and challenges. Disabil Rehabil. 2007;29:1387–1395
    • View In Article
    • CrossRef
13. Keyser-Marcus LA, Bricout JC, Wehman P, et al. Acute predictors of return to employment after traumatic brain injury: a longitudinal follow-up. Arch Phys Med Rehabil. 2002;83:635–641
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (62 KB)
    • CrossRef
14. Gollaher K, High W, Sherer M, et al. Prediction of employment outcome one to three years following traumatic brain injury (TBI). Brain Inj. 1998;12:255–263
    • View In Article
    • MEDLINE
    • CrossRef
15. Cooper KD, Tabaddor K, Hauser WA, Shulman K, Feiner C, Factor PR. The epidemiology of head injury in the Bronx. Neuroepidemiology. 1983;2:70–88
    • View In Article
    • CrossRef
16. Jager TE, Weiss HB, Coben JH, Pepe PE. Traumatic brain injuries evaluated in U.S. emergency departments, 1992-1994. Acad Emerg Med. 2000;7:134–140
    • View In Article
    • MEDLINE
    • CrossRef
17. Whitman S, Coonley-Hoganson R, Desai BT. Comparative head trauma experiences in two socioeconomically different Chicago-area communities: a population study. Am J Epidemiol. 1984;119:570–580
    • View In Article
    • MEDLINE
18. Rosenthal M, Dijkers M, Harrison-Felix C, et al. Impact of minority status on functional outcome and community integration after traumatic brain injury. J Head Trauma Rehabil. 1996;11:69–79
    • View In Article
    • CrossRef
19. Greenspan AI, Wrigley JM, Kresnow M, Branche-Dorsey CM, Fine PR. Factors influencing failure to return to work due to traumatic brain injury. Brain Inj. 1996;10:207–218
    • View In Article
    • MEDLINE
    • CrossRef
20. Vanderploeg RD, Curtiss G, Duchnick JJ, Luis CA. Demographic, medical, and psychiatric factors in work and marital status after mild head injury. J Head Trauma Rehabil. 2003;18:148–163
    • View In Article
    • MEDLINE
    • CrossRef
21. Arango-Lasprilla JC, Ketchum JM, Williams K, et al. Racial differences in employment outcomes after traumatic brain injury. Arch Phys Med Rehabil. 2008;89:988–995
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (121 KB)
    • CrossRef
22. Kreutzer JS, Marwitz JH, Walker W, et al. Moderating factors in return to work and job stability after traumatic brain injury. J Head Trauma Rehabil. 2003;18:128–138
    • View In Article
    • MEDLINE
    • CrossRef
23. Arango-Lasprilla JC, Ketchum JM, Gary KW, et al. The influence of minority status on job stability after traumatic brain injury. PM&R. 2009;1:41–49
    • View In Article
24. Sherer M, Nick TG, Sander AM, et al. Race and productivity outcome after traumatic brain injury: influence of confounding factors. J Head Trauma Rehabil. 2003;18:408–424
    • View In Article
    • MEDLINE
    • CrossRef
25. Harrison-Felix C, Newton CN, Hall KM, Kreutzer JS. Descriptive findings from the Traumatic Brain Injury Model Systems National Database. J Head Trauma Rehabil. 1996;11:1–14
    • View In Article
    • CrossRef
26. Teasdale G, Jennett B. Assessment of coma and impaired consciousness: a practical scale. Lancet. 1974;2:81–84
    • View In Article
    • CrossRef
27. State University of New York at Buffalo. Guide for the Uniform Data Set for Medical Rehabilitation (including the FIM instrument), Version 5.1. Buffalo: State University of New York at Buffalo; 1997;
    • View In Article
28. Granger CV, Hamilton BB, Keith RA, Zielesny M, Sherwin FS. Advances in functional assessment for medical rehabilitation. Top Geriatric Rehabil. 1986;1:59–74
    • View In Article
29. Stineman MG, Shea JA, Jette A, et al. The Functional Independence Measure: tests of scaling assumptions, structure, and reliability across 20 diverse impairment categories. Arch Phys Med Rehabil. 1996;77:1101–1108
    • View In Article
    • Abstract
    • Full-Text PDF (948 KB)
    • CrossRef
30. Rappaport M, Hall KM, Hopkins K, Belleza T, Cope DN. Disability rating scale for severe head trauma: coma to community. Arch Phys Med Rehabil. 1982;63:118–123
    • View In Article
    • MEDLINE
31. Zafonte RD, Hammond FM, Mann NR, Wood DL, Black KL, Millis SR. Relationship between Glasgow coma scale and functional outcome. Am J Phys Med Rehabil. 1996;75:364–369
    • View In Article
    • MEDLINE
    • CrossRef
32. Gouvier WD, Blanton PD, LaPorte KK, Nepomuceno C. Reliability and validity of the Disability Rating Scale and the Levels of Cognitive Functioning Scale in monitoring recovery from severe head injury. Arch Phys Med Rehabil. 1987;68:94–97
    • View In Article
    • MEDLINE
33. Molenberghs G, Geert Verbeke G. Models for discrete longitudinal data. New York: Springer; 2006;
    • View In Article
34. Hosmer DW, Lemeshow S. Applied logistic regression. 2nd ed.. New York: John Wiley & Sons; 2000;
    • View In Article
35. U.S. Bureau of Labor Statistics. Employment situation summary. Washington (DC): Bureau of Labor Statistics; 2009;
    • View In Article
36. Franulic A, Carbonell CG, Pinto P, Sepulveda I. Psychosocial adjustment and employment outcome 2, 5 and 10 years after TBI. Brain Inj. 2004;18:119–129
    • View In Article
    • MEDLINE
    • CrossRef
37. Dikmen SS, Temkin NR, Machamer JE, Holubkov AL, Fraser RT, Winn HR. Employment following traumatic head injuries. Arch Neurol. 1994;51:177–186
    • View In Article
    • MEDLINE
38. Johnson R. How do people get back to work after severe head injury? (A 10 year follow-up study). Neuropsychol Rehabil. 1998;8:61–79
    • View In Article
    • CrossRef
39. Sykes-Horn W, Wrigley M, Wallace D, Yoels W. Factors associated with awareness of vocational rehabilitation services after traumatic brain injury. Arch Phys Med Rehabil. 1997;78:1327–1330
    • View In Article
    • Abstract
    • Full-Text PDF (510 KB)
    • CrossRef
40. Hart T, Dijkers M, Fraser R, et al. Vocational services for traumatic brain injury: treatment definition and diversity within model systems of care. J Head Trauma Rehabil. 2006;21:467–482
    • View In Article
    • MEDLINE
    • CrossRef
41. Yasuda S, Wehman P, Targett P, Cifu D, West M. Return to work for persons with traumatic brain injury. Am J Phys Med Rehabil. 2002;80:852–864
    • View In Article
    • MEDLINE
    • CrossRef

• a SAS Institute, 100 SAS Campus Dr, Cary, NC 27513.