Models of brain injury vocational rehabilitation: The evidence for resource facilitation from efficacy to effectiveness

1 Return to work (RTW) after acquired brain injury (ABI)

Return to work after brain injury has always been regarded as a critical outcome metric in research, yet established continuums of rehabilitation services typically do not extend through vocational placement and follow-up. When vocational services do exist, they are often a) not brain injury specialized or b) integrated into an existing continuum of services, which typically results in people not receiving vocational rehabilitation services, or if they do, it is many years post-injury. In addition to the brain injury specific barriers to RTW, these system barriers compromise vocational outcome.

A variety of studies have demonstrated that approximately 30– 40% of people with ABI ultimately return to work. A systematic review (Van Velzen, Van Bennekom, Edekaar, Sluiter, & Frings-Dresen, 2009) found that 40.8% of those with traumatic brain injury (TBI) and 39.3% with non-TBI returned to work. Inclusion criteria in this review included a) the subjects were working prior to their injury, b) were aged 18– 65 years, and c) return to work was an outcome measure. Their review was based on a sample of 49 studies that met all of the inclusion criteria. It should be noted however that this study did not separate outcomes based on severity of injury.

A more recent study was completed using population estimates from the Traumatic Brain Injury Model Systems National Database (TBIMS-NDB) (Cuthbert et al., 2015). In this study, the investigators culled 3121 subjects from the database which were weighted to obtain population estimates to match the US TBI rehabilitation population based on both the Uniform Data System for Medical Rehabilitation (UDS, 2012) and the American Medical Rehabilitation Providers Association, eRehab (American Medical Rehabilitation Providers Association, 2012). Their sample was based on inclusion criteria that included less than 60 at the age of injury, not retired at injury, and alive two years post-injury. These investigators found that 39.6% were employed, which included paid legal or illegal work, with or without accommodations. Of these 39.6% that were able to RTW, 65% of them were employedfull time.

A number of studies have found that severity of brain injury has a significant impact on return to work. For example, in 2002, Groswasser et al. found that 84% of subjects with mild TBI were able to RTW, and in another study of mild TBI, 78% were able to RTW (Hanlon et al., 1999). Dikmen and colleagues (Dikmen, Temkin, Machamer, Holubkov, Fraser, & Winn 1994) found that 37 percent of subjects with severe traumatic brain injury (TBI) had returned to work, 64 percent with moderate TBI, and 83 percent of mild TBI returned to work at two years post-injury.

Even these findings may however over-estimate rates of RTW for moderate and severe brain injury. Certainly not all people who need inpatient rehabilitation receive these services, and consequently, are not included into the TBIMS database. Individuals who are seen in their State Vocational Rehabilitation agency are typically many years post-injury, did not have access to specialized brain injury rehabilitation services, have developed multiple co-morbidities, have lost vocational skills and networks, in addition to presenting with persisting cognitive and neurobehavioral disabilities (often among others), making return to work a certain challenge. For example, Schopp, Johnstone, Unger, & Goldman (2003) found that only 18 percent of State Vocational Rehabilitation clients with TBI were successfully placed.

2 Barriers to RTW after acquired brain injury

Vocational rehabilitation of people with brain injury represents a significant challenge from all perspectives. Barriers to RTW span across individual and family variables, as well as social, environmental, and system/organizational domains. It is for these reasons that an effective approach to RTW following brain injury will not be a “medical” or a “vocational” model, but rather one that is capable of responding to the unique and individualized constellation of barriers, and the interactions that each case presents.

Critical individual barriers include neurobehavioral impairments (e.g., disinhibition, impulsivity, decreased initiation) or cognitive impairments (e.g., impaired goal setting or task monitoring, memory, attention) that are ubiquitous following ABI and can significantly affect job performance and adjustment in the workplace (Dikmen et al., 1994). Medical consequences of brain injury (e.g., posttraumatic seizures) or co-morbidities (e.g., depression or substance abuse) also represent barriers to RTW or work stability. Level of behavioral adjustment post-injury affect family adjustment (Kreutzer, Marwitz, & Kepler, 1992), which in turn influence recovery and adaptation. Further, changes in social roles typically result in increased emotional and economic burden for family members, making it even more difficult to sustain family advocacy and emotional engagement, as well as support for the family member with the brain injury (Kolakowsky-Hayner & Kishore, 1999; Kreutzer, Gervasio, & Camplair, 1994).

Social barriers include lack of reimbursement for services and the absence of brain injury expertise among the host of providers involved in the vocational rehabilitation of people with brain injury. Multiple providers (e.g., employment services and cognitive rehabilitation) do not typically collaborate, compromising the potential effectiveness of each of their interventions. Awareness of State agency resources is typically lacking, and access to these resources can be cumbersome and overwhelming. Further, States do not typically have an organized State plan for brain injury, resulting in service silos within and between State agencies with resulting service gaps.

Complicating the vocational rehabilitation process even more is the fact that, at least in moderate to severe TBI, most people will have chronic disability that is variable in type and severity over time (Corrigan & Hammond, 2013). In fact, the severity of TBI disability has been found to vary more often that it remains static in long-term follow-up studies. Further, the individual and social-environmental variables dynamically interact with each other, where, for example, a loss of a significant relationship can result in depression that, in turn, results in further impairment of memory, which then results in impaired vocational performance. On the other hand, successful cognitive rehabilitation may result in new compensatory strategies, that when integrated into the work setting, with improvement in self-efficacy, and then a better response to psychotherapy for depression. Again, brain injury vocational rehabilitation is neither medical nor vocational; it is of necessity an individualized process of eliminating or mitigating brain injury-specific and other barriers with a holistic scope.

3 Models of vocational rehabilitation for acquired brain injury

A systematic review of the literature on different types of vocational rehabilitation for people with ABI (Fadyl & McPherson, 2009) demonstrated that there have been essentially three different approaches that included a supported employment (SE) model, a “Program-based vocational rehabilitation model” often referred to as a comprehensive day treatment program (CDT) model, and a “vocational case coordination (VCC)” model. The development of SE for people with brain injury was a significant contribution to improving vocational outcome (Wehman et al., 1989; Wehman, Bricout, & Targett, 2000; Wehman et al., 2003). Wehman and colleagues modified the individual placement model for people with brain injury and there was a clear emphasis on individualized training at the work site through a job coach rather than through pre-placement training and intervention. In contrast, the CDT program provided individual and group treatment for cognitive, neurobehavioral, and psychological difficulties driven by a neuropsychological approach (Ben-Yishay, Silver, Piasetsky, & Rattock, 1987; Prigatano et al., 1994). These programs were typically provided services four or five days a week for approximately six months and included family education and training, as well as vocational trials, placement and follow-up. The first two studies on “vocational case coordination,” the third approach, were completed by Malec and colleagues (Malec, Buffington, Moessner, & Degiorgio, 2000; Malec & Moessner, 2006). This approach features an individualized approach to promoting access to vocationally-related needs and ensuring integration of services to impact on vocational skills and workplace adaptation. The overall structure of these different models from Fadyl and McPherson (2009) are provided in Fig. 1.

OPEN IN VIEWER

Fig.1

Overview of the structure of three different types of vocational rehabilitation programs for brain injury (reproduced from Fadyl and McPherson, 2009 with permission).

These investigators also rated the quality of the research for each of these models to determine the relative strength of experimental support for their effectiveness. Their findings in this respect are summarized in Table 1.

The research by Malec and colleagues (2000, 2005) on VCC and the systematic review by Fadyl and McPherson (2009) provided the scientific background for the development of the RF model developed at the Rehabilitation Hospital of Indiana (RHI).

4 Objectives

The overall objective of this study was to examine the effectiveness of the RHI RF program for a clinical prospective cohort of participants referred to this program from the Indiana Vocational Rehabilitation Services. Based on our previous research and previous research, we had two hypotheses that included 1) the participants in the RF prospective clinical cohort would demonstrate a significantly better rate of RTW or school, and 2) group assignment (RF versus controls) would predict outcome with baseline level of disability as a covariate.

5 Methods

6 Results

Some very noteworthy demographic differences between the two groups were present at baseline: years post injury, age at injury, and injury type (see Table 2). As previously noted, the control group was derived from a previous RCT where subjects were recruited while in acute inpatient rehabilitation or shortly thereafter, and were a little more than two months post-injury. In marked contrast, the participants in the RF group in the present study were referred from Indiana Vocational Rehabilitation Services and were found to be, on average, over 9 years post-injury. Additionally, the age at injury for the RF group was found to be significantly younger (29.6 years) as compared to the control group which was on average 40 years old. This difference is likely attributable to significant differences between the two groups in terms of diagnosis, where there were more participants with stroke in the control group and more moderate to severe TBI subjects in the RF group.

6.1 Hypothesis 1: Success rate by group

Of the 210 participants in the treatment group, 69% (n = 145) were successful for return to competitive work. Six of these 145 successful outcomes were for participants that had a goal of return to school. Of the portion of the sample for which work hours and type of employment information was available, it was found that the average hours worked per week was 24.88 (sd = 10.38) and 36% of the successfully closed cases were full-time (30+ hours per week). The occupations to which they were placed are provided in Fig. 2. Almost half of the successful participants returned to either administrative support or laborer positions, but the other half were distributed across all other types of occupations. Of the 33 participants in the control group, 48% (n = 16) successfully returned to paid employment. This difference was significant, indicating a greater proportion of success in the treatment group than the control at the end of the treatment duration [X² ₍₁₎  = 5.39, p = 0.018] (see Fig. 3).

OPEN IN VIEWER

Fig.2

Occupational categories for successfully closed cases.

OPEN IN VIEWER

Fig.3

Employment Rates for RF and Control Groups.

OPEN IN VIEWER

Table 3

Summary of logistic regression analysis predicting employment

Variable	B	SE	Wald Statistic	p
Treatment Group	– 0.824	0.388	4.52	0.033
Disability	– 0.056	0.019	8.51	0.004

6.2 Hypothesis 2: Treatment group predicts outcome with baseline level of disability as a covariate

The first model examined the relationship between baseline level of disability and employment. A logistic regression model was statistically significant, χ²(1) = 9.92, p = 0.002, but it should be noted that this model only explained 5.5% of the variance in productive activity and correctly classified 67.9% of the cases. Level of disability at baseline provided an exponential slope of 0.94 indicating that with every one point increase in MPAI-4 T score (higher scores indicating a greater level of disability), the odds of successful employment declines by nearly one percent (Wald = 9.06, p = .003) (See Fig. 4).

OPEN IN VIEWER

Fig.4

Outcome by Initial Level of Disability.

Adding treatment group to the model and leaving level of disability as a covariate increased the models significance and improved the odds of successful employment for those in the treatment group. This model was also statistically significant, χ²(2) = 14.41, p = 0.001 and explained 8% of the variance in productive activity and correctly classified 67.1% of the cases. When controlling for baseline level of disability, treatment group significantly predicted employment outcome (Wald = 4.52, p = 0.033) and participants in the treatment group were 2.3 times more likely to return to work than controls.

In conclusion, participants receiving RF had a higher employment rate than participants in past control groups. In fact, participants in the treatment group were over two times more likely to return to work than the control group when controlling for level of disability. As demonstrated by previous research, successful vocational outcome varied as a function of level of disability as measured by the MPAI-4, where 74% of participants with mild disability, 65% of participants with moderate disability, and 48% of participants with severe disability were successful with obtaining employment through RF. Further, the participants in RF were almost ten years post-injury as compared to the control group, which was slightly over two months post-injury. Additionally, the RTW rate of 69% for the RF cohort is considerably better than what the literature would suggest for people of ABI of approximately 40% for all levels of disability. The findings from the present study are also consistent with the RTW rates we found in our two previous RCT’s. The present findings provide some initial support for the clinical effectiveness of RF for a sample referred by the Indiana Vocational Rehabilitation Services.

However, the present study has several limitations. Specifically, the sampling method in this study is not traditional, and although our control group was randomized into the control condition, our treatment group was not, therefore potentially adding bias to our study. In addition, some significant differences between the groups at baseline could confound some of the outcome variables, specifically time since injury, age at injury, as well as diagnosis. However, this sample is more representative of a clinical population seen in a State Vocational Rehabilitation agency, and likely better represents an otherwise ignored sector of the population previously underrepresented in clinical trials. Therefore, these findings support the transition from the established RF efficacy into clinical effectiveness.

The present study took place over different economic epochs, which could serve to influence RTW. The recession occurred from 2007 through 2011. Control group data was collected between 2008 and 2013, while data from the treatment group was collected between 2009 and 2012. Therefore, a larger proportion of data collection took place during the recession for the treatment group than the control groups. Therefore, it would seem unlikely that the changes in the economy served as a positive bias for RTW for the treatment group.

Although this study successfully demonstrates the effectiveness of RF, additional established predictors of employment success should be taken into account. For example, Cuthbert and colleagues found that age, race, gender, marital status, rehabilitation payment source, education, pre-injury work status, length of stay, and the disability rating scale (DRS) were significant predictors of employment two years post injury (Cuthbert et al., 2015). Due to our sampling methods, we were unable to use well-known predictors of outcome in our statistical analyses. This exclusion of key predictors likely led to the small effect size of our prediction model. Also, future studies in RF need to consider relationship between pre-injury rates of employment and post-RF employment and study change in occupational type from pre-injury to post-injury. Last, we were not able to evaluate durability or stability of employment for more than 90 days, and given that the disability associated with these injuries was likely chronic, and for some, the level of disability might get worse. This would thereforesuggest a need for long-term surveillance to proactively monitor vocational adjustment and stability.

Conflict of interest

The authors declare that they have no conflicts of interest.