Supported Employment for Persons With Traumatic Brain Injury : A Preliminary Investigation of Long-Term Follow-Up Costs and Program Efficiency

Text-only Preview

Supported Employment for Persons With Traumatic Brain
Injury: A Preliminary Investigation of Long-Term Follow-Up
Costs and Program Ef?ciency

Paul Wehman, PhD, John Kregel, EdD, Lori Keyser-Marcus, MA, Pam Sherron-Targett, MEd,
Leanne Campbell, BS, Michael West, PhD, David X. Cifu, MD

ABSTRACT. Wehman P, Kregel J, Keyser-Marcus L,
Key Words: Brain injuries; Employment, supported; Reha-
Sherron-Targett P, Campbell L, West M, Cifu DX. Supported
employment for persons with traumatic brain injury: a
© 2003 by the American Congress of Rehabilitation Medi-
preliminary investigation of long-term follow-up costs and
cine and the American Academy of Physical Medicine and
program ef?ciency. Arch Phys Med Rehabil 2003;84:192-6.
Objective: To investigate the long-term follow-up costs of
supported employment as well as the wage and employment
TRAUMATIC BRAIN INJURY (TBI) can result in a vari-
characteristics for individuals with moderate to severe trau-
ety of problems, including cognitive de?cits, impaired
matic brain injury (TBI) who participated in supported employ-
psychosocial functioning, and physical or sensory disabilities.
ment services over a 14-year time period.
The impact of these sequelae on employment is often detri-
Design: Longitudinal design with prospectively collected
mental. Research has shown that individuals with TBI often
experience dif?culty securing and/or returning to competitive
Setting: A university-based supported employment program
employment postinjury and maintaining employment for ex-
that uses the individual placement model of supported employ-
tended periods of time.1-6 Estimates of the employment rate for
persons with TBI range from around 20% to 50% depending on
Participants: Fifty-nine individuals with moderate to severe
the severity of the injury, the prior work experience of the
TBI who were consecutively referred for supported employ-
individual, and demographic characteristics (eg, age, educa-
ment services. The sample was restricted to individuals who
tion, socioeconomic status).2,3,7-9 Prior investigations10-12 have
were placed into a least 1 supported employment position
noted several negative psychosocial consequences of unem-
during the study period.
ployment for individuals with TBI, including depression, de-
Interventions: Not applicable.
creased social functioning, and various physical ailments. Ad-
Main Outcome Measures: Data were collected on clients
ditionally, ?nancial dif?culties commonly exist; lost wages
placed into at least 1 competitive supported employment posi-
present increased economic burdens to families of individuals
tion from 1985 to 1999. Analyses were performed to examine
with TBI, as well as to society as a whole when dependence on
the costs of supported employment, employment characteris-
public assistance results.
tics (eg, wages, length of employment), and bene?t-cost ratios
Supported employment services are often used to assist
of supported employment for individuals with TBI.
individuals with disabilities to secure and maintain employ-
Results: The average length of employment for the current
ment.13-15 In recent years, the supported employment model has
sample was 42.58 months. Average gross earnings were
been applied successfully to individuals with TBI, greatly
$26,129.74 for individuals during their entire duration of em-
enhancing employability and employment rates postinjury.16-18
ployment. Billing charges accrued for employment services
Supported employment is de?ned as “paid work, occurring in
averaged $10,349.37. Individuals with TBI earned an average
integrated settings, with the provision of ongoing support ser-
of $17,515 more than the costs associated with their supported
vices.”19 Essential elements of supported employment services
have been pay for real work, integration in the workplace with
Conclusions: Our investigation provides additional support
nondisabled coworkers, long-term ongoing supportive services
for the conclusion that supported employment is cost effective
to facilitate job retention, placement of individuals with severe
for individuals with disabilities, including individuals with
handicapping conditions, and interagency cooperation and
TBI, and that the costs of supported employment decrease over
funding of these services.20 The individual placement model of
supported employment uses a vocational specialist (or job
coach) to assist the person with locating employment, and
coordinating or providing on-the-job supports such as new
employee skills training and/or identifying accommodations.
Also, long-term follow-up services are offered throughout the
From the Department of Physical Medicine and Rehabilitation, Virginia Common-
person’s employment. During this time, additional on-the-job
wealth University/Medical College of Virginia (Wehman, Cifu) and School of
assistance is available if needed, and, as indicated, select case
Education, Virginia Commonwealth University (Kregel, Keyser-Marcus, Sherron-
Targett, Campbell, West), Richmond, VA.
management services related to off-the-job-site issues that if
Supported in part by the National Institute on Disability and Rehabilitation Re-
left unattended would impact job retention, are provided.
search, US Department of Education (grant nos. H133A980026, H133B980036).
Supported employment continues to be a program in which
No commercial party having a direct ?nancial interest in the results of the research
there is an interest in ef?ciency of outcomes. Cost ef?ciency
supporting this article has or will confer a bene?t upon the authors or upon any
organization with which the authors are associated.
provides a comparison of the monetary bene?ts of a program
Reprint requests to Paul Wehman, PhD, Rehabilitation Research and Training
versus costs. Cost-effectiveness measures the long-term results
Center, Virginia Commonwealth University, Box 842011, Richmond, VA 23284,
and achievements of services. If the value of a program’s
e-mail: [email protected]
bene?ts exceeds the costs, the program is considered to be a
good investment, or “cost-ef?cient.”21 Previous research has
Arch Phys Med Rehabil Vol 84, February 2003

shown that supported employment is cost ef?cient (as well as
Table 2: Lengths of Employment and Client Earnings
cost effective when compared with sheltered workshops) for
Employment Outcomes
individuals with all types of disabilities, including individuals
with severe disabilities.21-28 Further, the cost ef?ciency of
Mean months employed
supported employment programs increases with each year of
Mean monthly earnings ($)
program operation, with the bene?ts of supported employment
Mean total earnings ($)
exceeding the costs by the fourth year of operation.21
Cumulative months employed
Information regarding the long-term costs of supported em-
Cumulative earnings ($)
ployment services as well as the long-term bene?ts to consum-
ers is crucial in determining the ef?cacy of supported employ-
ment in individuals with TBI. To date, research has not
examined the long-term costs of providing supported employ-
Glasgow Coma Scale score of less than 13 on admission to the
ment to individuals with TBI. The goal of our investigation was
hospital. In addition, each client had to present clear indications
to examine the costs associated with, as well as the wages
of the need for ongoing vocational intervention to return to
earned from supported employment, and the length of employ-
preinjury employment or to obtain and maintain new employ-
ment for individuals with TBI who participated in a supported
ment. These indications could come from the individual’s
employment over a 14-year time period. More speci?cally, the
postinjury employment history or from reports from his/her
questions we attempted to answer were (1) What is the average
family, physician, or vocational rehabilitation counselor. Indi-
cost of supported employment services for individuals with
viduals were not screened out of the program on the basis of the
TBI? (2) What is the average length of employment for indi-
nature or severity of cognitive, behavioral, or other impair-
viduals with TBI who have received supported employment
ments. However, all individuals who received services through
services? and (3) How do bene?ts (ie, participant income)
supported employment were asked to abstain from using alco-
compare with costs of supported employment during a 14-year
hol and other nonprescription drugs.
time period (from 1985 to 1999)?
Data regarding client wages, length of employment, hours
worked per week, and demographic information were collected
prospectively by the employment specialist assigned to each
The current sample consisted of 59 individuals with TBI
client. Additionally, data regarding charges for supported em-
who were consecutively referred to the supported employment
ployment services were provided by billing records, accessed
program at Virginia Commonwealth University’s Rehabilita-
through the Virginia Commonwealth University/Medical Col-
tion Research and Training Center, between August 1985 to
lege of Virginia Associated Physicians’ billing of?ce.
August 1999. All participants were placed into at least 1
Data Analysis
supported, competitive position, by using the individual place-
ment model of supported employment. To be eligible for
Descriptive statistics were calculated for length of employ-
services, individuals had to meet the following criteria: be of
ment, costs associated with supported employment services,
working age (18 – 64y) and have sustained a moderate to severe
and wages earned. Subgroup comparisons were also performed
TBI, as indicated by length of coma greater than 24 hours or a
to examine the in?uence of length of employment (?2y and
?2y) on wages and supported employment costs. For the
purpose of our study, calculations were based on individual
clients, not job placements. Therefore, employment and billing
Table 1: Client Characteristics (N?59)
data were combined and averaged for individuals who were
placed into more than 1 job over the study duration.
Mean age ? SD (y)
Descriptive statistics were computed on demographic char-
Gender (%)
acteristics of the sample and are presented in table 1. The
average age of participants was 32.6?8.4 years; 81.4% were
men, and 74.6% were white.
Race (%)
The majority of the sample (91.7%) had sustained a severe
TBI, as indicated by length of coma of more than 1 day. With
African American
regard to cause of injury, 67% of the sample were injured in
vehicular crashes, 8% had acquired their injuries because of
Education (%)
abuse or assault, 10% were injured in falls, 11% sustained their
Less than high school
injuries from gunshot wounds, and mechanism of injury was
High school graduate or equivalent
classi?ed as “other” for 4% of the sample. The majority of the
Trade or technical school
sample (71.4%) were employed full-time at the time of injury.
Some college
College graduate
Employment Characteristics
Unknown (n?3)
Results of the employment calculations revealed that the
Preinjury work status (%)
average hourly wage for individuals in the current sample was
Full-time employment
$5.25 (range $3.35–$11.99/h). Individuals worked anywhere
Part-time employment
from 8 to 44 hours a week, with an average of 30 hours a week.
Student working
The average length of employment for the sample was 42.58
Student not working
months (range,
Unemployed not student
?1–141mo). Key employment outcomes are
described in table 2. Over a 14-year period, the average length
Unknown (n?3)
of employment was 42.58 months. Average monthly earnings
Abbreviation: SD, standard deviation.
were $633.63, which is substantially higher than the reported
Arch Phys Med Rehabil Vol 84, February 2003

Table 4: Costs of Supported Employment Services
Program Costs
Mean per person program costs ($)
Mean monthly program costs (per person) ($)
Cumulative program costs ($)
group comparisons. Average earnings and supported employ-
ment costs for individuals who were employed for 0 to 24
months (brief employment group) were compared with average
costs and earnings of individuals who were employed for 25
months or more (extended employment group). Results of the
computations revealed similar earnings levels between groups,
Fig 1. Employment retention over time.
with individuals in the brief employment group earning an
average of $659.93 per month, and individuals in the extended
earnings for other supported employment groups, and consid-
employment group earning an average of $567.26 per month of
erably higher than the Substantial Gainful Activity (SGA)
employment (t ?1.31, P?.19). However, a t test performed to
threshold, which was in effect during the study time period.
compare billing costs between groups found signi?cantly
Data provided by the Rehabilitation Service Administration
higher monthly employment service costs attributed to the brief
reported mean weekly earnings at closure for individuals with
employment group (t ?4.04, P?.01). Average cost per month
disabilities (not only TBI) who used supported employment
of employment services was $1304.09 for individuals in the
services in ?scal year (FY) 1998 as $142.93 (Rehabilitation
brief employment group, whereas average monthly cost of
Services Administration, unpublished raw data for FY 1997
employment services for individuals in the extended employ-
and FY 1995).
ment group was $156.05.
Average gross earnings were $26,129.74, with a range of
$571.43 to $168,290.57 for individuals during their entire
duration of employment (across placements, when applicable).
The results of our investigation provide additional support
Ranges for length of employment and gross monthly earnings
for the conclusion that supported employment is cost effective
are displayed in ?gure 1 and table 3, respectively. In ?gure 1
for individuals with disabilities, including individuals with
the length of employment contains a skewed distribution. Over
TBI, and that the costs of supported employment decrease over
half of the sample worked less than 2 years (24mo). However,
time. In an earlier study, Wehman et al29 found that the mean
approximately 25% of participants worked for 7 years (84mo)
annual expenditure of supported employment for individuals
or longer. In fact, nearly 10% of the sample was employed for
with TBI was $10,198 for the ?rst year. Conversely, the aver-
12 years (144mo) or longer. With regard to monthly earnings,
age annual costs of supported employment services in our
nearly 82% of the sample were earning over $400 per month
current longitudinal investigation were $8614, suggesting a
(median earnings, $602/mo), and 63% of all participants earned
savings over time. Findings from our prospective investigation
more than the SGA threshold in effect during the time of the
also appear to support prior research suggesting that the earn-
ings reported by individuals with TBI in supported employ-
ment far exceed the costs associated with supported employ-
Supported Employment Costs
ment services,30 as participant income was an average of
Billing charges accrued for employment services averaged
$17,515 greater than the cost of supported employment ser-
$10,349.37 (range, $646.96 –$69,635.99). Average costs per
vices received over the 14-year time period.
person for employment services were also calculated and are
Results of the subgroup comparisons suggest that, although
displayed in table 4; mean program cost was $8614. It is
wages did not differ with regard to length of employment, the
important to note that this ?gure includes all of the costs of
costs associated with supported employment services were
supported employment services for individuals over the study
signi?cantly lower for individuals who maintain employment
period, resulting in a mean monthly program cost of $202.
for 2 or more years. There are several potential explanations for
these ?ndings. One explanation may be that costs of supported
Subgroup Cost-Earnings Comparisons
employment decrease over time, as suggested by Kregel et al.21
The skewed nature of employment duration (see ?g 1)
An alternative explanation for the differences related to costs of
prompted further examination of the data in the form of sub-
supported employment between groups could be that the sub-
group of individuals who were able to maintain employment
for an extended duration (
Table 3: Monthly Gross Earnings for Individuals in Supported
?2y) were higher functioning than
those who were employed only brie?y, and thus did not require
as much on-site support, and/or extent of services. A recent
Range of Monthly Earnings
study by Keyser-Marcus et al1 showed strong associations
$ Range
between key demographic and functional variables and return
to work after TBI. Similarly, demographic and/or functional
characteristics may play a substantial role in long-term job
retention as well. Unfortunately, our subgroup analyses were
limited in measuring such differences between subgroups. Fu-
ture investigations should focus on demographic and functional
variables related to employment retention of individuals with
1001 or more
Arch Phys Med Rehabil Vol 84, February 2003

It is important to note that although the ?ndings of our
investigation are encouraging, these results are only prelimi-
1. Keyser-Marcus L, Bricout J, Wehman P, et al. Acute predictors of
nary and are not representative of the range of costs that may
return to employment after traumatic brain injury: a longitudinal
be involved in supported employment services. For example,
follow-up. Arch Phys Med Rehabil 2002;83:635-41.
we did not include the costs of assistive devices or other
2. Asikainen L, Kaste M, Sarna S. Predicting late outcome for
accommodations that might be necessary for job placement
patients with traumatic brain injury referred to a rehabilitation
and/or retention. Additionally, the costs incurred by services
programme: a study of 508 Finnish patients 5 years or more after
injury. Brain Inj 1998;12:95-107.
provided by medical, psychiatric, and other rehabilitation pro-
3. Curl RM, Fraser RT, Cook RG, Clemmons D. Traumatic brain
fessionals (eg, physical therapists, speech therapists) were not
injury vocational rehabilitation: preliminary ?ndings for the co-
considered in our investigation. Finally, individuals in our
worker as trainer project. J Head Trauma Rehabil 1996;11(1):75-
sample participated in supported employment services that had
been established before their enrollment in services. Conse-
4. McMordie WR, Barker SL. The ?nancial trauma of head injury.
quently, start-up costs for supported employment were not
Brain Inj 1988;2:357-64.
re?ected in the current analyses. Therefore, emerging sup-
5. Kreutzer J, Morton MV. Traumatic brain injury: supported em-
ported employment programs should anticipate higher costs in
ployment and compensatory strategies for enhancing vocational
the early stages of program implementation than are reported
outcomes. In: Wehman P, Moon S, editors. Vocational rehabili-
tation and supported employment. Baltimore: PH Brookes; 1988.
p 291-311.
Although examination of monetary costs and bene?ts of
6. Brooks N, McKinlay W, Symington C, Beattie A, Campsie L.
supported employment for individuals with TBI is important
Return to work within the ?rst seven years after head injury. Brain
with regard to allocation of funds and resources, it is impera-
Inj 1987;1:5-19.
tive that rehabilitation professionals acknowledge the nonmon-
7. Burleigh SA, Farber RS, Gillard M. Community integration and
etary bene?ts of employment (eg, increased community inte-
life satisfaction after traumatic brain injury: long-term ?ndings.
gration, improved quality of life, increased self-esteem,
Am J Occup Ther 1998;52:45-52.
increased levels of worker integration) for this population.
8. O’Connell MJ. Prediction of work following traumatic brain in-
Many individuals with TBI often express feelings of social
jury: intellectual, memory, and demographic variables. Rehabil
Psychol 2000;45:212-7.
isolation, worthlessness, and decreased self-esteem as a conse-
9. Wall JR, Niemczura JG, Rosenthal M. Community-based training
quence of TBI.10-12 Research investigating the cost effective-
and employment: an effective program for persons with traumatic
ness of supported employment with individuals with mental
brain injury. NeuroRehabilitation 1998;10:39-50.
illness has shown that supported employment had a positive
10. Bell KR, Sandel ME. Brain injury rehabilitation. 4. Postacute
impact on clients’ mental health and may have even substituted
rehabilitation and community integration. Arch Phys Med Rehabil
for traditional mental health services.23,31 Future research
1998;79(Suppl 1):S21-5.
should examine the impact of these intangible bene?ts of
11. Crisp R. Return to work after traumatic brain injury. J Rehabil
supported employment on individuals with TBI. Further, when
computing bene?t-cost ratios, information about intangible and
12. Lubusko AA, Moore AD, Stambrook M, Gill DD. Cognitive
beliefs following severe traumatic brain injury: association with
nonmonetary bene?ts, fringe bene?ts (eg, health insurance,
postinjury employment status: Brain Inj 1994;8:65-70.
leave time), and costs including tax liability (federal, state,
13. Anthony WA, Brown MA, Rogers ES, Derringer S. A supported
local) and public assistance (Social Security Income, Social
living/supported employment program for reducing the number of
Security Disability Insurance) should be considered. Research
people in institutions. Psychiatr Rehabil J 1999;23(1):5-23.
reveals a national growth in supported employment participa-
14. Bond GR, Dietzen LL, Vogler K, Katuin CH, McGrew JH, Miller
tion from FY 1986 through FY 1995. Participation rose from
LD. Towards a framework for evaluating cost and bene?ts of
approximately 10,000 persons in FY 1986 to over 139,000 in
psychiatric rehabilitation: three case examples. J Vocational Re-
FY 1995. This re?ects a 16% annual growth rate from partic-
habil 1995;5:75-88.
ipant totals reported in FY 1993.32 These data indicate that
15. Danley KS, Rogers ES, MacDonald-Wilson K, Anthony W. Sup-
ported employment for adults with psychiatric disability: result of
vocational rehabilitation agencies and other funding sources
an innovative demonstration project. Rehabil Psychol 1994;39:
are turning to supported employment with increasing fre-
quency. However, emphasis is on serving individuals with
16. Wehman P, Kreutzer J, West M, et al. Employment outcomes for
mental retardation and mental illness labels. Focus also re-
persons following traumatic brain injury: pre-injury, post-injury,
mains on serving people with mild and moderate levels of
and supported employment. Brain Inj 1989;3:397-412.
disability rather than those persons with the most severe dis-
17. Wehman P, Kreutzer J, West M, et al. Return to work for persons
with traumatic brain injury: a supported employment approach.
Arch Phys Med Rehabil 1990;71:1047-52.
18. Wehman P, West M, Kregel J, Sherron P, Kreutzer J. Return to
work for persons with severe traumatic brain injury. A data-based
Supported employment participation has expanded very little
approach to program development. J Head Trauma Rehabil 1995;
for persons with TBI. In FY 1995, only 1.7% of supported
employment participants had a primary disability classi?cation
19. Rehabilitation Act, Pub L No. 99-506, 100 Stat 1845 (1986).
of TBI.32 Although it is encouraging to see some progress,
20. Shafer MS, Wehman P, Kregel J, West M. National supported
many more funds now directed to segregated or work readiness
employment initiative: a preliminary analysis. Am J Ment Retard
types of programs need to be redirected to competitive em-
ployment efforts. Furthermore, individuals with TBI often do
21. Kregel J, Wehman P, Revell G, Hill J, Cimera R. Supported
not have access to any type of long-term funds, and this must
employment bene?t-cost analysis. J Vocational Rehabil 2000;14:
be included in the future long-range funding plans of states.
22. Rogers SE. Cost-bene?t studies in vocational services. Psychiatr
Finally, supported employment programs are effective when
Rehabil J 1997;20(3):25-33.
provided by well-trained staff dedicated to understanding the
23. Rogers SE, Sciarappa K, MacDonald-Wilson K, Danley K. A
needs of the person served as well as the business. Many
bene?t-cost analysis of a supported employment model for per-
programs are not adequately prepared to serve persons with
sons with psychiatric disabilities. Eval Program Plann 1995;18:
TBI at this time.
Arch Phys Med Rehabil Vol 84, February 2003

24. Thompson L, Powers G, Houchard B. The wage effects of sup-
for persons with mental retardation. Res Dev Disabil 1987;8:71-
ported employment. J Assoc Persons Severe Handicaps 1992;
29. Wehman P, Kregel J, West M, Cifu D. Return to work for patients
25. Shearn J, Beyer S, Felce D. The cost-effectiveness of supported
with traumatic brain injury: analysis of costs. Am J Phys Med
employment for people with severe intellectual disabilities and
Rehabil 1994;73:280-2.
high support needs: a pilot study. J Appl Res Intellect Disabil
30. West M, Wehman P, Kregel J, Kreutzer J, Sherron P, Zasler N.
Costs of operating a supported work program for traumatically
26. McCaughrin WB, Ellis WK, Rusch FR, Heal LW. Cost-effective-
brain-injured individuals. Arch Phys Med Rehabil 1991;72:127-
ness of supported employment. Ment Retard 1993;31:41-8.
27. Cimera R. Are individuals with severe mental retardation and
31. Clark RE, Bush PW, Becker DR, Drake RE. A cost-effectiveness
multiple disabilities cost-effective to serve via supported employ-
comparison of supported employment and rehabilitative day treat-
ment programs? Ment Retard 1998;36:280-92.
ment. Admin Policy Ment Health 1996;24:63-77.
28. Hill ML, Banks PD, Handrich RR, Wehman PH, Hill JW, Schafer
32. Wehman P, Grant R, Kregel J. Supported employment: a decade
MS. Bene?t-cost analysis of supported competitive employment
of rapid growth and impact. Am Rehabil 1998;Spring:31-43.
Arch Phys Med Rehabil Vol 84, February 2003