Functional Outcomes From Inpatient Rehabilitation After Traumatic Brain Injury: How Do Hispanics Fare?

Methods 

Data were extracted from the National Institute on Disability and Rehabilitation Research (NIDRR) funded Traumatic Brain Injury Model Systems (TBIMS) national database. The protocol for the proposed study was reviewed and deemed to have met ethical standards according to the institutional review board at Kessler Medical Rehabilitation Research and Education Corp.

For the purposes of the present study, TBI was defined as damage to brain tissue caused by an external mechanical force. Such damage is normally suspected if the person loses consciousness or experiences posttraumatic amnesia (PTA) as a result of brain trauma or if the neurologic findings confer with physical or mental examination of the patient and can be attributed to TBI.25 Most participants had a moderate to severe TBI, which was determined by meeting 1 of the following criteria: PTA for longer than 24 hours, intracranial neuroimaging abnormalities caused by trauma, loss of consciousness for more than 30 minutes (excluding unconsciousness as a result of drug or alcohol use), or an emergency department–documented Glasgow Coma Scale (GCS) score of less than 13. Patients who were chemically paralyzed, in a chemically induced coma, or intubated at the time of GCS scoring were labeled as such in the database and excluded from the sample.25

Additional inclusion criteria included being age 16 or older at the time of the injury and arriving at a TBIMS acute care hospital within 24 hours postinjury. All participants received both acute hospital care and comprehensive rehabilitation in a designated inpatient TBIMS rehabilitation program between 1989 and 2003. Comprehensive rehabilitation must have occurred in a hospital, rehabilitation unit, rehabilitation hospital, hospital-based skilled nursing facility, skilled nursing facility, or long-term acute care hospital in which (1) medical and rehabilitation care are regularly supervised by a TBIMS-affiliated physician; (2) the patient receives nursing care around the clock; and (3) at the very minimum, physical therapy, occupational therapy, speech therapy, rehabilitation psychology and/or neuropsychology, and family support and/or education are all available as needed to enhance functional gains. The patient (or the patient’s legal guardian or family member if appropriate) provided informed consent so that relevant health information from medical records or from laboratory tests could be included in the database.26

The final sample consisted of 3056 participants (2745 whites, 311 Hispanics) with primarily moderate to severe TBI who received acute care and inpatient rehabilitation in a TBI TBIMS during a 14-year period (1989−2003).

Results 

Admission 

As seen in table 1, Hispanic TBI survivors were more likely to be men (P≤.001), less educated (P≤.001), earned less money (P≤.05), and were younger (P≤.001) than whites. There were no significant differences in marital and employment status between groups. Table 2 shows that Hispanics were twice as likely to suffer a violence-related TBI compared with whites (P≤.001). Hispanics had significantly lower GCS scores at admission (P≤.01). No statistically significant differences were found between whites and Hispanics on length of PTA, duration of rehabilitation services, or the total LOS. As indicated in table 3, there were no significant differences between Hispanics and whites on the DRS or on the FIM instrument at admission.

Table 1. Demographic Characteristics

	Variables	Hispanics	Whites	P
	Mean age ± SD (y)	33.2±14.4 (n=311)	38.0±17.9 (n=2745)	<.001
	Sex (%)	(n=311)	(n=2745)	<.001
	Male	80.7	71.5
	Female	19.3	28.5
	Marital status (%)	(n=309)	(n=2741)	.526
	Married	32.0	33.9
	Unmarried	68.0	66.1
	Employment status (%)	(n=108)	(n=1066)	.689
	Student	5.6	8.8
	Competitively employed	64.8	64.4
	Unemployed	12.0	9.6
	Retired	13.0	13.8
	Other	4.6	3.5
	Education (%)	(n=286)	(n=2605)	<.001
	No college education	82.5	63.6
	Any college education	17.5	36.4
	Annual earning (%)	(n=53)	(n=590)	<.05
	≤$49,999	96.2	80.7
	$50,000–$99,999	3.8	15.6
	≥$100,000	0.0	3.7

Abbreviation: SD, standard deviation.

Table 2. Injury and Rehabilitation Characteristics

	Variables	Hispanics	Whites	P
	Injury characteristics
	GCS	7.9±4.1(n=246)	8.8±4.4(n=2117)	<.01
	PTA	27.3±19.8(n=215)	28.0±23.8(n=1882)	.370
	Cause of the injury (%)	(n=310)	(n=2736)	<.001
	Vehicular	49.0	62.9
	Acts of violence	20.0	8.3
	Sports	0.3	1.9
	Falls	29.0	25.1
	Pedestrian	1.6	1.9
	Rehabilitation characteristics
	LOS in acute care	23.0±17.0(n=311)	21.0±16.5(n=2745)	.022
	LOS in rehabilitation	28.7±21.8(n=303)	29.9±27.8(n=2671)	.397
	Total LOS	51.6±31.1(n=303)	50.9±37.4(n=2671)	.082

NOTE. Values are mean ± SD or as otherwise indicated.

Table 3. Functional Outcomes

	Functional Outcomes	Hispanics	Whites	P
	DRS
	Admission	12.7±5.3(n=299)	12.3±5.8(n=2672)	.112
	Discharge	6.1±4.0(n=300)	6.0±4.0(n=2680)	.588
	1-year follow-up	3.5±3.6(n=182)	2.4±3.4(n=2089)	<.001
	FIM
	Admission	54.0±24.8(n=206)	54.5±26.3(n=1799)	.867
	Discharge	95.1±22.5(n=215)	96.2±23.4(n=1903)	.171
	1-year follow-up	113.2±17.1(n=114)	116.4±18.5(n=1197)	<.001
	CIQ	(n=112)	(n=1265)
	1-year follow-up	13.8±5.4	16.1±5.7	<.001
	GOS-E	(n=121)	(n=1628)
	1-year follow-up	5.0±1.7	5.7±1.7	<.001

NOTE. Values are mean ± SD.

One-Year Follow-Up 

Fig 1, Fig 2, Fig 3, Fig 4 and table 4 show that Hispanics were significantly more likely to score in the worse quartile (1 quartile) for the DRS, FIM, GOS-E, and CIQ compared with whites at 1-year follow-up (Cochran-Armitage trend tests, P<.001) (for each variable, see table 4). For example, significantly more Hispanics (34%) had higher DRS scores (1 quartile) compared with whites (18%). For the FIM, 41% of Hispanics were in the worse quartile compared with 24% of whites. Similar results were observed for the GOS-E (48% vs 27%) and CIQ (37% vs 26%).

View Large Image Download to PowerPoint

Fig 1. DRS score at 1 year postinjury. ■, Hispanics; □, whites.

View Large Image Download to PowerPoint

Fig 2. FIM score at 1 year postinjury. ■, Hispanics; □, whites.

View Large Image Download to PowerPoint

Fig 3. CIQ score at 1 year postinjury. ■, Hispanics; □, whites.

View Large Image Download to PowerPoint

Fig 4. The GOS-E score at 1 year postinjury. ■, Hispanics; □, whites.

Table 4. Distribution of Outcomes for Hispanics and Whites

	Outcome Measure	0.0	0.1−1.5	1.6−4.0	≥4.0	P⁎
	DRS (1-y follow-up)					<.001
	Hispanic	15%	21%	30%	34%
		27	39	55	61
	Whites	30%	28%	24%	18%
		628	577	498	386

		18−116	117−122	123−125	≥126
	FIM (1-y follow-up)					<.001
	Hispanic	41%	21%	20%	18%
		47	24	23	20
	Whites	24%	25%	27%	24%
		288	298	321	290

		1–4	5–6	7–8	P⁎
	GOS-E (1-y follow-up)				<.001
	Hispanic	48%	30%	22%
		58	36	27
	Whites	27%	35%	37%
		445	577	606

		0−11	12−15	16–19	≥20
	CIQ (1-y follow-up)					<.001
	Hispanic	37%	25%	21%	18%
		41	28	23	20
	Whites	26%	24%	24%	26%
		273	272	338	382

⁎ Cochran-Armitage trend test.

Similar conclusions can be drawn for the DRS, FIM, and CIQ from the adjusted proportional odds logistic regression models (table 5). The cumulative odds for being in worse quartiles is significantly higher for Hispanics compared with whites for the DRS, FIM, and CIQ, the adjusted or ordinal multiple logistic regression models. For example, in the model with the DRS as the 4-category ordinal outcome variable after adjusting for DRS score at admission, age, PTA, education, and GCS score, Hispanics had 2.8 times higher odds than whites for being in the worst quartile (1 quartile), worst 2 quartiles (1 or 2 quartiles), or worst 3 quartiles (1 or 2 or 4 quartiles) versus being in the remaining quartile(s), respectively. Similar results were obtained for the FIM and CIQ in their adjusted models.

Table 5. Proportional Odds Models for 1-Year Follow-Up

	Complete Case Analyses (Listwise Deletion)					Multiple Imputation Analyses
	Variable	Reference	Adjusted OR	95% CI	P⁎	Variable	Reference	Adjusted OR	95% CI	P†
	DRS (n=1196)					DRS (n=3056)
	DRS admission	1 unit	1.03	1.02−1.04	<.001	DRS admission	1 unit	1.02	1.01−1.02	<.001
	Age	1 unit	1.03	1.02−1.04	<.001	Age	1 unit	1.03	1.02−1.03	<.001
	PTA	1 unit	1.02	1.01−1.02	<.001	PTA	1 unit	1.02	1.01−1.02	.004
	College	Yes	0.54	0.43−0.67	<.001	College	Yes	0.62	0.52−0.74	<.001
		No	1				No	1
	GCS	Mild	0.89	0.67−1.17	.407	GCS	Mild	0.65	0.49−0.85	.003
		Moderate	1.08	0.80−1.45	.635		Moderate	0.74	0.58−0.94	.017
		Severe	1				Severe	1
	Hispanic	Yes	2.79	1.90−4.09	<.001	Hispanic	Yes	2.21	1.66−2.94	<.001
		No	1				No	1
	FIM (n=557)					FIM (n=3056)
	FIM admission	1 unit	0.98	0.97−0.99	<.001	FIM admission	1 unit	0.98	0.97−0.99	.002
	Age	1 unit	1.02	1.01−1.03	<.001	Age	1 unit	1.02	1.01−1.03	.001
	PTA	1 unit	1.02	1.01−1.02	.001	PTA	1 unit	1.01	1.00−1.02	.131
	College	Yes	0.56	0.40−0.79	.001	College	Yes	0.73	0.58−0.91	.007
		No	1				No	1
	GCS	Mild	1.47	0.95−2.29	.085	GCS	Mild	0.96	0.62−1.49	.842
		Moderate	1.39	0.90−2.15	.138		Moderate	0.94	0.64−1.37	.713
		Severe	1				Severe	1
	Hispanic	Yes	2.98	1.68−5.29	<.001	Hispanic	Yes	1.70	1.23−2.35	.002
		No	1				No	1
	CIQ (n=703)					CIQ (n=3056)
	Age	1 unit	1.04	1.03−1.05	<.001	Age	1 unit	1.04	1.03−1.05	<.001
	PTA	1 unit	1.02	1.01−1.02	<.001	PTA	1 unit	1.02	1.01−1.02	.002
	College	Yes	0.40	0.30−0.54	<.001	College	Yes	0.46	0.37−0.56	<.001
		No	1				No	1
	GCS	Mild	0.83	0.58−1.19	.301	GCS	Mild	0.75	0.57−1.05	.043
		Moderate	0.77	0.53−1.12	.174		Moderate	0.78	0.61−1.02	.063
		Severe	1				Severe	1
	Hispanic	Yes	2.92	1.78−4.79	<.001	Hispanic	Yes	1.61	1.06−2.44	.030
		No	1				No	1

Abbreviations: CI, confidence interval; OR, odds ratio.

⁎ Based on Wald test. † Based on t distribution.

Accepting the idea that imputations produced reasonable replacements for the missing values, the analyses based on multiple imputations revealed a shift toward the null but did not eradicate the complete case analysis (see table 5). Hispanics still fared worse compared with whites with significantly higher odds of worse functional outcomes and community integration at 1 year postdischarge.

There were no differences between the pre- and post-1995 cohorts of Hispanics on the the following demographics (covariates) used in each model: college (P=.729), GCS score (P=.808), age (P=.976), and PTA (P=.469). Also, there was no difference in the proportion enrolled in TBIMS between Hispanics and whites pre- and post-1995 (P=.139), and finally the dummy variable for cohort was not significant in each of the proportional odds models either as a main effect or as an interaction effect with race and ethnicity.

Discussion 

This is the largest, multicenter, longitudinal study of the impact of ethnicity on functional outcome after TBI in the literature. The purpose of the present study was to examine the relation between Hispanic ethnicity and rehabilitation outcome in TBI survivors on admission to and discharge from inpatient rehabilitation as well as 1 year postinjury. This study extends prior TBIMS research by focusing specifically on Hispanics, controlling for injury severity, and assessing outcomes at 1 year postinjury. Results showed that there were no significant differences between whites and Hispanic TBI survivors in functional outcome at both admission and discharge. In contrast, at 1 year postinjury, Hispanics were now approximately 2.5 times more likely than whites to have worse outcomes across all functional outcome measures. Even controlling for differences between Hispanics and whites at admission to rehabilitation, significant differences in all categories of functioning at 1 year postinjury were observed.

The findings of this large-scale, multicenter study confirm some of the demographic, injury, and discharge functional outcome results previously reported in the literature. For example, a smaller TBIMS database study comparing a minority (black, n=216; Hispanic, n=42) group of TBI survivors to a white group (n=310) of TBI survivors used data from the 4 initial TBIMS.36 The finding of significantly lower levels of education among minority TBI survivors at admission was reported in both studies. In addition, both studies found that the minority and white TBI survivor groups were similar with respect to their cognitive and physical functioning at admission and discharge. Most important, both the present study and Rosenthal et al36 found that the minority group had significantly lower levels of community integration at 1 year postdischarge. Another small study by Hart et al30 compared 55 white and 39 black TBI survivors and found that despite the similar demographic characteristics, at the 1-year follow-up, blacks scored significantly lower on the social integration subscale of the CIQ than whites. Additionally, Hart30 reported that black TBI survivors also significantly decreased their income from preinjury to 1-year follow-up, whereas their white counterparts did not. Thus, although differences in educational level may initially appear to account for differences in community integration at 1 year postdischarge, loss of income may be a more important contributor to lower community integration scores.

Some findings of the present study are not consistent with previously reported findings. For example, Burnett et al37 examined data from a single TBIMS center over a 10-year period and described the demographic, injury, and rehabilitation discharge outcome characteristics of a minority group composed of African-American (n=82) and Hispanic (n=5) TBI survivors. Burnett reported no differences in injury characteristics at admission or functional outcomes on discharge when this group was compared with a white group. Unfortunately, the statistical analyses and corresponding results are not reported in the study and no information is provided regarding the white TBI survivor comparison group. The present much larger scale study did find significant differences in outcome between Hispanics and whites and by virtue of the sample size alone, more likely reflects true differences between groups than the descriptive analysis of a relatively small sample of minority TBI survivors as reported by Burnett.37 Initial sociodemographic factors (eg, income and educational levels) found in the present study may have differentially affected rehabilitation characteristics and functional outcomes at 1-year follow-up, even though admission data and discharge outcomes were not affected. It is likely that while patients are in rehabilitation, they receive similar services, regardless of preinjury socioeconomic levels. However, when patients are discharged, sociodemographic variables may assume a more important role (eg, in access to transportation and outpatient rehabilitation services, level of income, language). Alternatively, TBI has been shown to affect socioeconomic levels after discharge (eg, loss of employment), and it is not clear if preinjury sociodemographic variables are as influential as the environment to which one returns to19 in determining rehabilitation outcomes in diverse individuals. Because very little is known about the factors that influence rehabilitation outcomes in minorities, future studies should examine the role of pre- and postinjury sociodemographic variables, independent of and in combination with ethnicity and other factors.

In the present study, the Hispanic group was more than twice as likely as the white group to have sustained a TBI as a result of violence. This is not surprising given the findings of other TBIMS studies that have looked at risk factors for brain injury in minorities and show that violent injuries are especially common (eg, Hanks et al38). This may be a consequence of a higher likelihood of living in potentially dangerous residential environments and/or being exposed to occupational hazards than their white counterparts.19 Differences in TBI etiology among Hispanics and whites may have contributed to a significantly lower initial GCS at admission for the Hispanic group.

Although similarities in functional status at admission and discharge may initially seem surprising given the differences in sociodemographic and injury characteristics reported between groups, an understanding of the inpatient rehabilitation selection process clarifies these seemingly disparate findings. For persons to “qualify” for admission to a TBIMS inpatient brain injury rehabilitation unit, they must first be able to show specific social support, physical, neurologic, and medical characteristics. Namely, they must be able to have reached a certain physical and functional threshold to meet admission criteria. Although this process is necessary for rehabilitation units to meet their mandate of effectively and efficiently returning individuals to the community, it also reduces the physical and functional variability observed after TBI. Likewise, even though initial demographic variables differ between groups, it is likely that both Hispanics and whites in the current sample had similar levels (even if they were of different make up) of social supports in place to enable a discharge to “home.” Thus, the inpatient rehabilitation requirements for admission to a TBIMS program may artificially “normalize” much of the variation seen acutely after TBI. Unfortunately, this homogenization limits the generalizability of this research for the population of all patients after moderate to severe TBI. At the same time, the conclusion from the current results on Hispanics must be somewhat conservative, and even larger differences may be observed if the admission criteria were in some way affecting Hispanics more than whites.

Importantly, the present study identified significant differences in functional and community integration outcomes at the 1-year postinjury follow-up between TBI survivors of Hispanic and white ethnicities. The Hispanic TBI survivors showed lower cognitive functioning on the DRS, lower cognitive and physical functioning on the GOS-E and FIM, and lower community integration on the CIQ. Given the similarities between the 2 groups in these parameters at rehabilitation discharge, it is unlikely that these new differences were the result of the initial brain injury. Rather, the differences are likely to have been caused by factors after discharge such as differences in socioeconomic status, ongoing social supports, late medical complications (both those related to the TBI and non-TBI issues), access to medical and rehabilitation aftercare, societal barriers (eg, unequal access to vocational opportunities), and cultural differences (eg, acculturation level, English language proficiency, religious issues, family roles, mistrust in majority group care providers, low expectations of recovery). Future research must necessarily focus on identifying the post–inpatient rehabilitation factors that are the most important contributors to successful functional and community integration outcomes in the Hispanic TBI survivor subpopulation.

The family plays an important role in the rehabilitation process once the TBI survivor is discharged. Hispanics often believe that the best place for the injured person is in the home to be cared for by a member of the family.39 Given the fact that many of the caregivers of Hispanic TBI survivors are dealing with multiple challenges that were most likely in place before their loved one’s injury (eg, fewer financial resources, less access to resources) and that these challenges may have been exacerbated by numerous postinjury stressors, the interplay between caregiver functioning and TBI survivor outcome deserves further study. Interventions geared toward assisting the caregiver (eg, providing information in the person’s native language and offering support groups [possibly even psychotherapy] to discuss problems and how to solve them) could then be developed and implemented to improve functional and community integration outcomes.

Conclusions 

In this large, multicenter, longitudinal study, Hispanic ethnicity was associated with poorer functional outcomes at 1 year postinjury relative to whites, despite a lack of significant group differences at discharge. Future studies should be aimed toward understanding variables that contribute to these long-term group differences as well as examining the potential impact on everyday quality of life such as return to work, productivity, attitudes, and beliefs about disability and rehabilitation as well as access to health care.