RSA services and employment outcome in consumers with traumatic brain injury.

It is estimated that 1.5 million Americans sustain a traumatic brain injury (TBI), annually, and that an estimated 5.3 million .individuals live with permanent disabilities as a result of their injuries (Centers for Disease Control, 2004). TBI occurs most frequently in three populations: young children, young adults, and older adults (Centers for Disease Control, 2004). The majority of individuals with TBI that vocational rehabilitation counselors see are in their early adulthood, less than 30 years of age (Ashley, Persel, Clark, & Krych, 1997). Return to work following TBI is critical because individuals typically have the majority of their work life remaining. The inability to gain and maintain employment not only affects individuals' self-esteem and self-identity, but also has larger social and economic consequences. Failure to return to work can result in poverty and long-term dependence on public assistance (Abrams, Barker, Haffrey, & Nelson, 1993; Dikman et al., 1994).

Disability sequelae following TBI often involve physical, behavioral, and cognitive problems that either alone, or in combination, cause a substantial impediment to employment (Centers for Disease Control, 2004; Golden, Smith, & Golden, 1993; Van Baalen et al., 2003). Employment rates within this population vary by study; however, the majority of researchers indicate high rates of unemployment that become even higher over time (Cattelani, Tanzi, Lombardi, & Mazzucchi, 2002; Isaki & Turkstra, 2000; Kosciulek, 1994; Kruetzer et al., 2003). Those who do return to work rarely do so at the same pre-injury level and pay scale (Gollaher, et al., 1998; Uomoto, 2000). In a recent analysis, Johnstone, Mount, and Schopp (2003) found that 31% of their participants with TBI were employed at one-year post injury with a 51% reduction in earnings and an estimated 275% increase in the amount of public assistance received. Given the large population, high cost, and poor employment outcome associated with consumers with TBI, numerous researchers have identified variables related to or predicting employment outcomes. Several researchers have indicated the need for more accurate prediction studies to improve consumer services and employment outcome (Leahy & Lam, 1998; Ponsford, Olver, Curran, & Ng, 1995, Schonbrun & Kampfe, 2004).

Using Rehabilitation Services Administration (RSA) data for a Southeastern state, Gamble and Moore (2003) examined the relationship of six services received by consumers with TBI (i.e., assessment, college, counseling/guidance, job placement, restoration, and work adjustment) to employment outcome. Their findings indicated that college, counseling/guidance, and job placement were significantly related to a competitive employment outcome for consumers with TBI. Of these three services, job placement most significantly predicted successful employment. This is consistent with findings of previous literature, that regardless of disability type, job placement services are frequently a significant predictor of employment outcome (Bolton et al., 2000). Gamble and Moore (2003) indicated that they did not find any similar studies that examined the relationship between specific services provided by RSA and employment outcome in consumers with TBI. They recommended that future research replicate their efforts to determine the consistency of their findings throughout the RSA system.

The purpose of the current research was to evaluate the relationship between specific services and employment outcome of clients with TBI using national data available through the RSA database for fiscal year 2002. The research questions were: 1) What specific services are related to competitive employment outcome in consumers with TBI? and 2) What is the relationship between the provision of specific services and consumers' earnings at case closure.

Methodology

Participants

The data analyzed in the current study were drawn from and represent a sample of RSA consumers whose cases were closed and recorded in the RSA 911 database for fiscal year 2002. The National RSA 911 database sample of consumers with TBI listed as the primary disability consisted of 14,311 consumers whose cases were closed. Similar to the general population of individuals with TBI, men comprised the majority of consumers (i.e., 64.9 %) and women, the minority (i.e., 35.1%). The primary disability was defined as "the individual's primary physical or mental impairment that causes or results in a substantial impediment to employment" (p. 10, United States Department of Education, 2000).

Instrument

RSA requires each state to submit data annually for each case closure in each fiscal year. Data are collected in a specific format and each record entry must be unique. Those with duplicate records are systematically rejected. Each record represents an individual whose case was closed during that year, regardless of the reason for closure. Each record is 213 characters in length and each set of characters represents one of 43 values or elements (United States Department of Education, 2000). Variables include the state where the consumer was served, consumer identification number, date of application and closure, referral source, gender, race, primary and secondary disabilities, information regarding the consumer's status at the time of both application and closure (i.e., employment status, hourly and weekly earnings, type and amount of public support), and type of and reasons for closure. Data are also collected on service information including services provided and cost of purchased services.

These data are collected in a specific format coded according to the Reporting Manual for the Case Service Report (United States Department of Education, 2000). Data are collected for each of 43 elements and each element is assigned a specific numeric code. For example, the record position of Primary Disability occupies position 43-46 of the 213-character record. The first two digits (i.e., position 43 and 44) represent categories of sensory/communication impairments, physical impairments, or mental impairments. The last two digits (i.e., positions 45 and 46) represent the codes for cause and sources of the impairments (i.e., a code of 37 occupying space 45 and 46 of a given record indicates the individual's primary impairment was caused by a traumatic brain injury).

There were six specific categories of employment status at case closure. The majority of consumers were closed in competitive employment (i.e., 82.1%). A smaller percentage of closed cases fell into extended employment, self-employment, homemaker, or unpaid family worker (i.e., combined total of 6.0%). The remainder of consumers in employment at closure (i.e., 11.9%) were closed in employment with support in an integrated setting. For purposes of this analysis, as in Gamble and Moore's (2003) analysis, only the category, status of competitive employment at closure (both successfully and unsuccessfully closed) was used because the other categories represented only a small proportion of cases and were excluded from this study. This resulted in a total of 6,460 cases. Of these, 4,498 consumers had successfully obtained competitive employment through RSA services and 1,962 had obtained attempted to obtain employment, but were unsuccessful. For clarity, this analysis has referred to the two groups as the employed group and the not-employed group. The employed group consists of the RSA consumers with TBI who obtained a competitive job following services (i.e., status 26). The unemployed group consists of the RSA consumers with TBI who attempted to obtain a competitive job following services, but were not successful and remained unemployed (i.e., a status 28).

Analysis

This research was an ex post facto correlational study (Shadish, Cook, & Campbell, 2002). The goal was to observe the size and direction of the relationship between the service variables and employment outcome. The following statistical techniques were utilized to analyze the RSA database in addressing the two research questions. These were the same techniques used by Gamble and Moore (2003) and included descriptive, chi-square analysis, logistic regression and ANOVA. SPSS was used in the analysis. A value ofp < .05 was selected as an accepted level of significance.

Hypothesis 1: The provision of specific services will be equivalent in the employed and unemployed groups.

Ho: e = e

Ha: e > e

This hypothesis was addressed through two separate analyses using a chi-square test and a logistical regression, respectively. The Chi-square test of independence analysis addressed the relationship between specific services provided and employment status. Each specific service constituted a categorical variable (i.e., service received or not received), as did each of the two groups of consumers (i.e., employed or not employed). The logistic regression analysis addressed the ability of specific services to predict employment outcome. Logistic regression is frequently used in prediction studies when outcome measures are categorical (i.e., employed or unemployed) (Willis, Hendershot, & Fabian, 2005). The goal was to assess the relative association of each independent variable to the criterion variable (Greenspan, Wrigley, Kresnow, Branche-Dosey, & Fine, 1996). In replicating Gamble and Moore's (2003) research, only the five most frequently provided services were analyzed.

Hypothesis 2: Consumers provided specific services will have weekly earnings equivalent to consumers not provided specific services.

Ho: e = e

Ha: e > e

To analyze the relationship between each service and consumers' weekly earnings, a factorial ANOVA was used. A factorial ANOVA is used to evaluate the relationship between several independent variables and a continuous variable (Huck & Cormier, 1996). In replicating Gamble and Moore's (2003) research, only the five most frequently provided services were analyzed.

Results

Hypothesis One--Service Variables

A significant difference in employment outcome was found between the employed and not employed groups in four out of five of the most commonly provided service variables analyzed using chi-square tests. Therefore, the null hypothesis was rejected and the alternative hypothesis partially accepted. Table 1 provides the percentages of consumers who received the five most frequently provided specific services overall, the percentage of consumers who received services in the employed and unemployed groups, the chi-square statistics, and the p values for each chi-square.

Consumers who received any of four specific services (i.e., assessment, job placement, job search, and diagnosis and treatment) were more likely to be competitively employed than those who did not receive those services. However, there was no such relationship between the specific service of counseling/guidance and employment outcome.

The five most frequently provided RSA services were further evaluated using logistic regression to identify the abilities of each service variable to predict employment outcome. The results are summarized in Table 2. When the confidence interval includes the value of 1, then that variable is not considered a useful predictor, therefore, the variables of assessment and counseling/guidance do not predict employment outcome. As shown in Table 1, job placement, job search, and diagnosis and treatment are considered statistically significant as predictor variables. The further away from 1 the odds ratio is indicates the stronger the relationship. Therefore, the ability of these services to predict employment outcome is strongest between job placement and least strong between diagnosis and treatment.

Hypothesis Two--Weekly Earnings

Table 3 provides the results of the ANOVAs that analyzed the relationship between the service variables and consumers' weekly earnings at case closure. To give a frame of reference, TBI consumers whose cases were closed in competitive employment had average weekly earnings of $314.35. Consumers whose cases were not closed in employment had average weekly earnings of $34.57. In these analysis, when assessment services were provided by RSA funds, consumers' weekly earnings at case closure were $271.65 weekly, compared to consumers' weekly earnings of $253.20 when this service was not provided (p = .070). Consumers who were provided counseling/guidance service had weekly earnings of $263.17; whereas consumers who were not provided counseling/guidance had weekly earnings of $267.12 (p = .499). Consumers who were provided job placement services, had weekly earnings of $247.40. When this service was not provided, weekly earnings averaged $275.89 (p = .001). When job search services were provided, consumers averaged weekly earnings of $267.03 compared to $274.09 when this service was not provided (p = .001). When diagnosis and treatment were provided by RSA funds, weekly earnings averaged $289.30; whereas when this service was not provided, consumers' mean weekly earnings were $249.30 (p = .001).

Discussion

The most frequently provided specific services to consumers with TBI were assessment, counseling/guidance, job placement, job search, and diagnosis and treatment. Assessment was not a predictor of employment outcome; however, consumers who received assessment services were more likely to achieve competitive employment than those who did not receive these services (p = .001). Job placement predicted employment outcome. Consumers who received this service were more likely to have obtained competitive employment than those who did not receive these services (p = .001). Job search and diagnosis/treatment services predicted employment outcome and consumers who received these services were more likely to achieve competitive employment than their counterparts (p = .001). There was no significant relationship between counseling/guidance and employment, nor did the provision of counseling/guidance predict employment outcome.

There was no significant relationship between assessment services or counseling/guidance services and weekly wages of consumers at case closure. The service variables of job placement and job search were significantly related to weekly income; however, this was a negative relationship. Consumers had lower weekly earnings (i.e., -$28.49 and -$7.06 respectively) when these services were provided than when they were not provided. There was a positive relationship between consumers' weekly earnings and diagnosis and treatment services. Those who received diagnosis and treatment services had higher weekly earnings (i.e., +$40).

As noted earlier, this study represents a national replication and expansion of Gamble and Moore's (2003) research, who used a state RSA database for consumers with TBI. The current analysis produced both similar and different findings from Gamble and Moore's (2003) research regarding specific services provided. In the national data set, the frequencies of specific services provided to consumers with TBI were different from the frequencies of specific service provision to consumers in the state data set analyzed by Gamble and Moore. Only three service variables (i.e., assessment, job placement, and counseling/guidance) were the same as those examined by Gamble and Moore. The findings regarding assessment were similar to that of Gamble and Moore in that both results indicated that the provision of assessment services was not a predictor of employment outcome. Both analyses found that job placement significantly predicted competitive employment at case closure. Contrasting with Gamble and Moore (2003), counseling/guidance did not have a significant relationship to employment outcome or the ability to predict employment outcome, nor did this service variable have a significant relationship to weekly earnings of consumers with TBI. The analysis of specific services and consumers' wages were done differently in the current analysis than that of Gamble and Moore and cannot be directly compared.

Limitations

There are three types of validity in experimental and quasi-experimental studies used in social science research (Shadish, Cook, & Campbell, 2002). This analysis has two threats to statistical conclusion validity (i.e., unreliability of treatment implementation and heterogeneity of units) (Shadish, 2002). It is not possible to know from the data the intensity, duration, or quality of the specific services that were provided. For the purposes of this analysis, services were classified and analyzed as a) not provided or b) provided by RSA funds. Therefore, statistical significance may be either under or over reported depending upon the quality and quantity of services that a consumer may have actually received. It is likely that services were not provided in a standardized manner to all clients and effects may be underestimated when less than full participation in services was provided to RSA consumers.

This analysis also has two threats to internal validity (i.e., maturation and attrition). Maturation is a concern in that consumers may have returned to employment with or without specific RSA services as a natural progression in their recovery. Attrition is an important factor and limits generalization of this analysis. Consumers' cases may be closed for a number of reasons. In this population, a large number of consumers with TBI were not closed in a status that reflected a competitive employment outcome. As an example, although employment outcome was the largest reason for case closure, the next three largest reasons were a) inability to locate or contact, b) refusal of service, and c) failure to cooperate. It is not unreasonable that these consumer behaviors may be related to emotional, behavioral or cognitive sequelae following TBI. It is important to note that TBI, by its nature, is heterogeneous in its results and sequelae. While these results may be applicable for group and practice implications, they may not be appropriate to a particular individual.

Implications

Assessment was positively related to employment outcome, but not related to consumer income. One possible explanation may be that consumers receiving this service were more impaired than those who did not receive this service. Alternatively, it is possible that consumers who required this service had less healthcare services prior to seeking employment than consumers who did not require this service.

The specific service variable of counseling/guidance did not show a relationship to employment outcome. One explanation may be that consumers who needed counseling/guidance services may represent individuals with greater levels of impairment or need at the time the specific services were provided. It is important to note that this variable refers to services, separate and apart, from the counseling relationship that exists between the consumer and counselor throughout the vocational rehabilitation process.

Job placement and job search were positively associated with employment outcome, however, they were negatively associated with consumers' weekly earnings at case closure. One explanation may be that consumers with greater impairment or lower levels of education may have benefited from these services. Job search is defined as "support and assist a consumer in searching for an appropriate job. May include help in resume preparation, identifying appropriate job opportunities, developing interviewing skills and making contacts" (United States Department of Education, 2000 p. 24). Job placement assistance is defined as "a referral to a specific job resulting in an interview, whether or not the individual obtained the job." It may then be assumed that consumers needing these services may represent more severely impaired or less educated consumers who would face difficulties in obtaining competitive employment without these services. Similarly, consumers who obtained competitive employment without these services may have been higher functioning or educated, and obtained higher paying jobs on their own.

Diagnosis and treatment was associated positively with employment outcome and weekly wages. It also predicted employment outcome; however, this service was provided just over 25% of the time. It is not possible to know from the data exactly what was diagnosed and treated; however, it is a positive indicator of employment outcome when specific medical/psychological services were addressed during the vocational rehabilitation process.

Recommendations for Future Research

Further research is warranted regarding the various forms of purchased counseling services (i.e. marital/family counseling, adaptation to disability/personal counseling, substance abuse, or cognitive/behavioral counseling) and their relationships to employment outcome in order to both provide better services to consumers and reduce RSA costs of services. Further research may address the effectiveness of these services in improving consumer outcome. In the current analysis, counseling services were provided to over half of consumers, but did not show a positive relationship to employment outcome. Under the presumption (and heavy literature support) that counseling is a beneficial service, future researchers may address why this service did not show a relationship to consumers in the RSA program. Further researchers may also wish to explore the relationship between counseling services and non-employment related outcomes (i.e., quality of life, medication dependence, etc)

One purpose of this analysis was to further develop a prediction model for RSA consumers with TBI. If additional information could be collected through the existing RSA 911 Database system, this would be helpful in furthering such a prediction model in future research. For example, information such as a consumers' injury severity classification and the date of their injury would have produced more specific information regarding this variable and employment outcome. Additionally, because assessment was a frequently provided service (i.e., 60%), additional information regarding the type of assessment (i.e., vocational evaluation, functional capacity evaluation, work hardening, or neuropsychological examination) would have produced multiple layers of information for future analysis.

Summation 

In summary, this analysis produced mixed findings regarding differences between the employed and unemployed group regarding specific services and employment outcome. There was a significant relationship between the specific services of assessment, job placement, job search, and diagnosis/treatment and employment outcome; however there was no such relationship between counseling/guidance, and employment outcome. In evaluating the relationship between weekly earnings and specific services, consumers who were provided diagnosis/treatment services had higher average weekly earnings than did consumers who did not receive this service. Conversely, consumers who were provided job placement and job search services had significantly lower earnings than did consumers not provided these services. There was no relationship between counseling/guidance services and consumers' weekly earnings.

References

Abrams, D., Barker, L. T., Haffrey, W., & Nelson, H. (1993). The economics of return to work for survivors of traumatic brain injury: Vocational services are worth the investment. Journal of Head Trauma Rehabilitation, 8(4), 59-76.

Ashley, M. J., Persel, C. S., Clark, M. C., & Krych, D. K. (1997). Long-term follow-up of post-acute traumatic brain injury rehabilitation: A statistical analysis to test for stability and predictability of outcome. Brain Injury, 11(9), 677-690.

Bolton, B. F., Bellini, J. L., & Brookings, J. B. (2000). Predicting client employment outcomes from personal history, functional limitations, and rehabilitation services. Rehabilitation Counseling Bulletin, 44(1), 10-21.

Cattelani, R., Tanzi, F., & Mazzucchi, A. (2002). Competitive reemployment after severe traumatic brain injury: Clinical, cognitive and behavioral predictive variables, Brain Injury, 16(1), 207-218.

Centers for Disease Control. Traumatic Brain Injury in the United States: A report to Congress. Downloaded from http://www.cdc.gov/doc/do/id.

Dikman, S. S., Temkin, N. R., Machamer, J. E., Holubkov, A. L., Fraser, R. L., & Winn, H. R. (1994). Employment following traumatic head injuries. Archives of Neurology (51), 177-186.

Gamble, D., & Moore, C. L. (2003). The relationship between VR services and employment outcomes of individuals with traumatic brain injury, Journal of Rehabilitation, 69(3), 31-38.

Golden, T. P., Smith, S. A., & Golden, J. H. (1993). A review of current strategies for the enhancement of vocational outcomes following brain injury, Journal of Rehabilitation, 55-60.

Gollaher, K., High, W., Sherer, M., Bergloff, P., Boake, C., Young, M., & Ivanhoe, C. (1998). Prediction of employment outcome one to three years following traumatic brain injury, Brain Injury, 12(4), 255-263.

Greenspan, A. I., Wrigley, J. M., Kresnow, M., Branche-Dorsey, & Fine, P. R. (1996). Factors influencing failure to return to work due to traumatic brain injury. Brain Injury, 10(3), 207-218.

Huck, S. W. & Cormier, W. H. (1996). Reading statistics and research (2nd edition). Harper Collins. New York, NY.

Isaki, E., & Turkstra, L. (2000). Communication abilities and work re-entry following traumatic brain injury. Brain Injury 10(3), 207-218.

Johnstone, B., Mount, D., & Schopp, L. H. (2003). Financial and vocational outcomes 1 year after traumatic brain injury. Archives of Physical Medicine and Rehabilitation, (84), 238-241.

Kosciulek, J. (1994). Conceptions of head injury: Implications for vocational rehabilitation. Journal of Applied Rehabilitation Counseling, 25(3), 61-63.

Kruetzer, J. S., Harwitz, J. H., Walker, W., Sander, A., Sherer, M., & Bogner, J. (2003). Moderating factors in return to work and job stability after traumatic brain injury. Journal of Head Trauma Rehabilitation, 18(2), 128-138.

Leahy, B. J., & Lam, C. S. (1998). Neuropsychological testing and functional outcome for individuals with traumatic brain injury. Brain Injury, 12(12), 1025-1035.

Ponsford, J. L., Olver, J. H., Curran, C., & Ng, K. (1995). Prediction of employment status 2 Years after traumatic brain injury. Brain Injury (9),11-20.

Rehabilitation Services Administration, Fact Sheet. Downloaded 10/9/2004.from http://www.ed.gov/about/offices/list/osers/ rsa/faq.html.

Schonbrun, S., & Kampfe, C. (2004). Variables relating to employment outcomes in consumers with traumatic brain injury. Journal of Applied Rehabilitation Counseling, 35(3), 14-20.

Shadish, W. L., Cook, T. D., & Campbell, D.T. (2002). Experimental and QuasiExperimental Designs for Generalized Causal Inference. Houghton Mifflin Company: Boston.

Uomoto, J. M. (2000). Application of the Neuropsychological Evaluation in Vocational Planning after Traumatic Brain Injury Rehabilitation: Practical vocational, neuropsychological & psychotherapy interventions. Boca Raton, FL: CRC Press, Inc.

United States Department of Education (2000, March 16). Policy directive: RSA-PD-00-06. Retrieved March 1, 2005, from http://www.ed.gov/policy/speced/guid/rsa/pdf.

Van Baalen, B., Odding, E., Maas, A. I. R., Ribbers, G. M., Bergen, M. P., & Stam, H. J. (2003). Traumatic brain injury: Classification of initial severity and determination of functional outcome. Disability and Rehabilitation, 25(1), 9-18.

Willis, S. P., Hendershot, G. E., & Fabian, E. S. (2005). Help-seeking behavior among people with disabilities: Results from a national survey. Journal of Rehabilitation, 71(1), 42-48.

Staci L. Schonbrun

Labor Market Consulting Service 

Amos P. Sales

University of Arizona 

Charlene M. Kampfe

University of Arizona

Staci L. Schonbrun, Labor Market Consulting Services, 1661 N. Swan Road, Sute 136, Tucson; AZ 85712.

Table 1

Most Frequently Provided VR Services to Consumers with TBI

Services              % Service   Employed     Not       Chi-     p
                      provided               employed   square

Assessment              60.0%      61.4%      56.7%     16.86    .001
Counseling/guidance     54.2%      54.3%      54.2%      4.26    .234
Job placement           28.6%      37.3%       8.7%     633.35   .001
Job search              27.4%      34.3%      11.5%     414.16   .001
Diagnosis & treat       25.9%      27.4%      22.5%     33.74    .001

Table 2

Logistic Regression Analysis of Most Frequently Provided VR Services

Services                B      Odds ratio   df    p

Assessment            .235     1.265        1    .209
Counseling/guidance   -.124     .883        1    .720
Job placement         -1.859    .156        1    .000
Job search            -.763     .466        1    .033
Diagnosis & treat     -.633     .531        1    .000

Services              95%Confidence Level

Assessment                .877-1.827
Counseling/guidance       .448-1.743
Job placement              .067-.365
Job search                 .232-.938
Diagnosis & treat          .396-.712

Table 3

Specific Service Variables and Weekly Earnings at Case Closure

Services                Provided   Not provided     F       p

Assessment              $271.65      $253.20       2.35   .070
Counseling/guidance     $263.17      $267.12        .79   .499
Job placement           $247.40      $275.89       5.49   .001
Job search              $267.03      $274.09       5.58   .001
Diagnosis & treatment   $289.30      $249.30      12.68   .001