Vocational rehabilitation and older adults: patterns in participation and outcome.

The goal of returning adults with disabilities to their highest level of productive employment is a national priority (Xie, Dain, Becker, & Drake, 1997). Despite continuing efforts to identify effective models of rehabilitation, the rate of sustained competitive employment by participants in vocational rehabilitation continues to be modest (Noble, Honberg, Hall, & Flynn, 1997). Even among the most successful models, outcomes in terms of the percentage of participants with psychiatric disabilities who are working steadily in competitive jobs peak around 40% (Drake, McHugo, Becker, Anthony, & Clark, 1996).

One related issue which has not been well studied, is whether consumers are participating in vocational rehabilitation with a uniform goal of seeking competitive employment. In a recent study of 228 participants in a Veterans Health Administration (VHA) vocational rehabilitation program primarily serving adults involved in psychiatric and/or substance abuse treatment, we reported that only 52% of participants were entering vocational rehabilitation in order to obtain a competitive job and only 5% reported that obtaining a competitive job was their primary reason for participating (Drebing et al., in submission). Other goals included participating to supplement ongoing psychiatric or substance abuse treatment, participating in order to work in the program due to its greater support and adaptations to their disabilities relative to competitive jobs, and participating in order to stay active and make additional income.

In this study of VHA participants, one of the subgroups least likely to be seeking a competitive job was composed of adults over the age of 55. This is not surprising, as the social, economic, and psychological role of work typically changes between middle and late adulthood (Levinson, 1986), suggesting that participation in vocational rehabilitation may have a different meaning and role for older vs. younger adults. For example, the clinical problems underlying participation are likely to differ as a function of age, given the increased incidence of medical problems which impinge upon the ability to work in older adults (Mutchler, Burr, Massagli, & Pienta, 1999). The incentives for participation are likely to differ as well, given the availability of Social Security and pension retirement benefits for adults of retirement age. Given evidence of age-related differences in the meaning of work, the incentives for work, and the clinical problems effecting work, it seems reasonable to hypothesize that participation and outcome of rehabilitation efforts in the area of work would also differ as a function of age. Unfortunately, there is very little published empirical data regarding vocational rehabilitation participation to document if and how older participants differ from their younger counterparts or how many older adults participate.

In one of the few available data sources, records from vocational rehabilitation programs administered by the VHA suggested that adults aged 55 and older represent 9% of all vocational rehabilitation participants, or about 1,800 participants annually (Seibyl, Rosenheck, Corwel, & Medak, 2000). The VHA data also indicated that the number and proportion of older participants has been steadily growing over the past six years (Seibyl et al., 2000). This may be related to findings from several studies of work among older adults that suggested work has become increasingly attractive to older adults (Maloney & Paul, 1990; Taylor & Leitman, 1989). The increasing life expectancy has created a need for greater financial resources to support older adults over a longer `retirement' period. At the same time, there has been growing recognition of significant non-monetary incentives for older adults to continue in paid employment (Kaye, Alexaner, & Kauffman, 1999). Several studies provided convincing evidence that paid work by older adults is associated with a range of personal advantages, including greater social contact and social support (Aquino, Russell, Cutrona, & Altmaier, 1996; Mor-Bank, Scharlach, Birba, & Sokolov, 1992), greater physical and mental activity (2001), better morale, and a greater sense of meaning (Jacobs, 1978) and life satisfaction (Soumerai & Avorn, 1983).

The current study seeks to explore whether older participants in vocational rehabilitation are significantly different from their younger counterparts. Given the lack of published studies addressing age and vocational rehabilitation, we are basing our hypotheses on broader findings related to age-related changes in the areas of work, health, and finances. We seek to better understand how age is related to the reasons older adults participate, how they participate, and the outcome of that participation, by testing the following hypotheses:

1. At the time of entry into a VHA vocational rehabilitation program, participants over the age of 55 will have different clinical conditions related to their work problems relative to younger participants.

2. Older participants will face different work-incentives in terms of income contingent upon work. Specifically, they will have (a) significantly lower rates of receiving income from competitive employment prior to admission and (b) higher rates of non-wage income--primarily Social Security retirement income--than participants under the age of 55.

3. The stated goals of participants over the age of 55 will (a) reflect less interest in participating in vocational rehabilitation as a means of obtaining competitive employment and (b) more interest in participation as an end in itself--work within the program as means of staying active and earning additional income.

4. Because of these differences in goals (i.e. lower interest in transitioning to competitive employment, higher interest in participating to stay active and earn income), older participants will tend to (a) stay in the program longer and (b) make more money.

5. Participants over the age of 55 will be less likely to be discharged to competitive employment.

Method

Archival data for a sample of adults entering the Compensated Work Therapy Program (CWT) of the VHA were used for this study. Administrative data for 27,799 participants in CWT from 1993-99 were analyzed. These data represent the administrative variables collected on participants in VHA CWT programs by the Northeast Program Evaluation Center (NEPEC) of the VITA. Currently there are over 100 CWT programs in 42 states and the District of Columbia. Additional data related to participant goals were available for a subsample of this dataset, composed of 256 consecutive admissions to the CWT program at the Bedford VA Medical Center. These data represent administrative data collected only at that site consisting of participant responses to survey questions regarding participant goals for participating in CWT.

The Program

Compensated Work Therapy is the primary vocational rehabilitation program within the VHA. CWT, which is fairly consistent across sites in its format, uses a mixed model of vocational rehabilitation, including sheltered workshops, transitional employment, and competitive job placement (Drebing, Rosenheck, & Penk, 2001). The goal of CWT is to help maximize participants' level of vocational functioning, preparing some for a successful return to competitive employment and providing longer-term work placements to others (Losardo, 1999), The limited experimental evaluation data available on CWT suggests that participation is associated with reduced drug and alcohol use, fewer episodes of homelessness, and fewer episodes of incarceration, but no clear change in psychiatric symptoms (Kashner et al., in submission).

The Bedford CWT program is the largest CWT program in the VHA system, serving 200-300 veterans each day. Relative to other CWT programs, the population served at Bedford has a higher percentage of veterans diagnosed with co-morbid psychiatric and substance use disorders (SUD) (Seibyl et al., 2000). This likely reflects the presence of a large domiciliary and a large transitional residence program to serve homeless mentally ill. The Bedford CWT program offers a diversified placement model consisting of three placement components: a relatively small sheltered workshop, transitional employment placements, and job placement. Except for the workshop, these jobs exist in integrated, real-work settings; program participants are doing the same work as non-participants in the same surroundings (Losardo, 1999). There are no restrictions upon who is admitted to CWT related to applicants' age or goals for participation.

Data Collection

The large dataset used for the current study represent data routinely collected by the NEPEC, which monitors vocational rehabilitation programs in the VHA (Seibyl et al., 2000). At each site, administrative data were collected by the Vocational Rehabilitation Specialists (VRS), the primary professional staff, based on a clinical interview at intake and at discharge, patient clinical records, and financial records of the vocational rehabilitation services program (Seibyl et al., 2000).

Variables included in the analyses were determined by the hypotheses and included age at the time of entry into vocational rehabilitation, other demographic variables, diagnosis, measures of work functioning prior to admission, whether participants received non-wage-related income, participation variables, and rehabilitation outcome. Diagnostic data from the NEPEC database are drawn from the medical record and clinical interview, and included (a) nine of the most common psychiatric diagnoses or clusters of diagnoses (see Table 1) as well as a general residual category of "other psychiatric disorder" (b) drug and alcohol abuse or dependence, and (c) "disabling" medical conditions. Measures of vocational functioning prior to admission included whether the veteran was employed, not working due to disability (unable to work) or unemployed in the absence of disability in the three years prior to admission. Categories of non-wage related income included Social Security Retirement Income, Social Security Disability Income--including Social Security Disability Income and Supplemental Income, and VA Disability Pension Income.

Measures of work functioning during participation in the program included whether the participant worked in a sheltered workshop or transitional employment setting. Participants in sheltered workshops typically receive a high degree of staff supervision, work exclusively with other CWT participants, and face relatively low work demands (Losardo, 1999). Participants in transitional employment settings typically work in integrated community and hospital settings in positions with relatively low supervision and high work demands. Assignment to sheltered workshop or transitional employment settings typically reflects a combination of the participants' preferences and the staff's evaluation of participants' need for supervision and ability to meet work demands (Losardo, 1999).

Treatment participation and compliance variables included duration of participation (defined as the number of days from the date of intake to the day of discharge); intensity of participation (defined as the mean number of hours of participation per week); mean hourly wage during participation; and total earnings during participation.

The outcome of treatment was measured in terms of placement in competitive employment at the time of discharge, which is the most commonly used outcome variable for vocational rehabilitation in the VHA as elsewhere. Competitive employment was defined as full-time or part-time employment in an on-going position which could have been filled by any applicant, i.e. not a position contracted for a vocational rehabilitation candidate alone.

Additional data related to participant goals for participation were available for a subsample of the larger dataset, composed of 256 consecutive admissions at one CWT site. The data for this subsample included participant responses to a brief survey regarding their goals for participating in vocational rehabilitation, completed at the time of admission. The questionnaire was developed by quality assurance staff within the CWT program (Drebing et al., in submission) (see appendix). To develop the questions, two focus groups were conducted with CWT participants to collect initial data regarding the range of potential client goals. Additional information was collected from individual interviews with veterans entering the program. Using this information, two multiple choice items were composed, one asking respondents to identify their primary goal for participating in CWT at that time, and the second asking them to identify all goals relevant to their participation. Both questions used the same multiple choice responses consisting of the thirteen goals most commonly identified by participants in the focus groups. Two additional responses were included, one for clients who wished to state that their goals were unclear, and a residual category of "other" for which they were asked to describe the residual goal.

Sample

Table 1 summarizes demographic and background variables for the sample and subsample. The full sample was made up primarily of Caucasian and African-American males, most of whom had twelve or more years of formal education, and were divorced or separated. Eighty-four percent had a SUD, 50% had a psychiatric disorder, and 37% were dually diagnosed.

When the full sample was compared with the subsample from Bedford, the subsample was more likely to be Caucasian, receiving disability income, to be diagnosed with a psychiatric disorder, and to be diagnosed with a medical disorder. The racial and ethnic difference reflects the relative racial makeup of Massachusetts, where the portions of the population that are self-identified as African-American or Hispanic/Latino are less than half of the corresponding national portions (United States Census Bureau, 2000). The differences with respect to diagnosis may reflect the fact that the Bedford program works closely with a homeless program and has a larger portion of dually diagnosed, homeless adults with multiple problems. The subsample was less likely to be employed in the three (3) years prior to admission and less likely to be employed at discharge.

Analyses

The analyses were conducted in three steps. First, to identify any potential confounding covariates of age, the sample was divided into two age groups, those younger than 55 and those 55 and older. These two groups were compared with respect to basic demographic data, diagnostic data, and work variables. Secondly, for each hypothesis, dependant variables were plotted as a function of age. Thirdly, we examined the relationship between age and the dependant variables, using a series of logistic regression and multiple regression analyses. Because of demographic differences found between the two age groups, we included education, gender, and race as co-variates. As these analyses focused on possible differences related to older age, and because of the potential confounding effect of age changes noted in years 20-40 (see figures 1-3), we truncated the age variable to 40-85 for the regression analyses.

[FIGURES 1-3 OMITTED]

Results

Statistically significant differences were found between the two age groups for a number of the demographic and background variables, including education, race, and diagnosis (see table 1). The 55+ group was less educated, and more likely to be Caucasian. There were differences with respect to disability compensation, with the 55+ group having a higher rate of receiving disability compensation. The older group also was more likely to be diagnosed with a disabling medical condition and less likely to be diagnosed with a psychiatric or substance use disorder. Looking at specific diagnoses, the older group was less likely to be diagnosed with PTSD, alcohol abuse or dependence, or drug abuse or dependence.

Hypothesis 1

Strong support is found for hypothesis one. As can be seen in figure 1, the rate of disabling medical conditions climbs steadily after the age of 25, rising from a rate of 30% at age 25 to a rate of 82% after age 74. The incidence of SUD's has a curvilinear relationship to age, with the rate climbing between ages 25 to 35, plateauing for 5-10 years, and declining at an increasing rate. The incidence of psychiatric disorders follows a less predictable course, varying between 45 and 65% over the age range. Results of the regression analyses indicate that after co-varying demographic variables, age was a significant predictor of having a disabling medical condition (Least Likelihood = 26664, * = .04, Odds Ratio = 6.0, 95% Confidence Interval = 4.8-7.4), a SUD (Least Likelihood = 16901, * = -.05, Odds Ratio = 0.11, 95% Confidence Interval = 0.08-0.14), and a psychiatric disorder (Least Likelihood = 25365, * = -.01, Odds Ratio = 0.73, 95% Confidence Interval = 0.59-0.91).

Hypothesis 2

Strong support is also found for both parts of hypothesis two. As can be seen in figure 2, the rate of receiving income from employment in the three years prior to admission declines steadily after the age of 25, dropping from a rate of 87% at age 25 to a rate of 17% after age 70. Not surprisingly, the rate of receiving retirement income climbs dramatically from age 55 to 65, moving from 0% to 64%, and increasing modestly after 65. Unexpectedly, the rate for receiving disability income also rises sharply after age 55, before dropping again after age 74.

Results of the regression analyses indicate that after co-varying demographic variables, age is a significant predictor of working prior to admission (Least Likelihood = 24283, * = -.06, Odds Ratio = .13, 95% Confidence Interval = 0.11-0.17). It is also a significant predictor of whether someone is receiving retirement (Least Likelihood = 1871, * = .30, Odds Ratio = 523783, 95% Confidence Interval = 221214-1240197), or disability income (Least Likelihood = 23328, * = .03, Odds Ratio = 2.5, 95% Confidence Interval = 2.0-3.1).

Hypothesis 3

Support is found for part A of hypothesis three. The percentage of participants who state that they are entering CWT as a means of obtaining competitive employment, declines steadily with age, from a high of 62% for 30-34 year olds, to 4% for participants over the age of 74 (see figure 3) (Least Likelihood = 269, = -.04, Odds Ratio = 0.21, 95% Confidence Interval = 0.04-0.94). Support is not found for part B of hypothesis three. Age was not a significant predictor of endorsing goals of participating in CWT in order to stay busy or to earn money

Hypothesis 4

Support is found for part A of hypothesis four. Length of stay steadily increases with age, rising from a mean of 85 days for participants younger than 25, to 198 days for participants aged 70-74 (see Table 2) (B = 1.46, SEB = 0.20, * = .06, p < .001). Surprisingly, part B is not supported. Total earnings during CWT participation declines after age 55, dropping from a high of $2,472 at age 55 to $1,303 after age 74 (B = 12.99, SEB = 4.21, = .02, p < .01). Given the apparent paradox of increased length of stay and decreased total earnings, we examined pay rates, mean hours worked per week, and the rate of participation in sheltered workshops, where participants are usually paid at piece rate. Both the number of hours worked per week and the average wage per hour decline significantly after the age of 55. Participants between the age of 25 and 55 work a mean of 25 hours per week in CWT. This drops steadily to a mean of 13 hours per week for participants over the age of 74 (B = 1.12, SEB = .03, * = -.03, p < .001). Wages drop steadily from a mean of $4.65/hour at age 59 to $2.94/hour by age 70 (B = -.02, SEB = .003, * = -.04, p < .001). The drop in wages is accompanied by the steady increase with age of the percentage of participants working in sheltered workshops. For the 30-55 group, an average of 33% of participants are working in sheltered workshops, which increases steadily with age until 85% of participants 75 and older are working in the workshops (Least Likelihood = 25146, * = .02, Odds Ratio = 2.6, 95% Confidence Interval = 2.1-3.3).

Hypothesis 5

Strong support is found for hypothesis five. As can be seen in figure 13, the rate of being competitively employed at discharge declines steadily after the age of 34, dropping from a rate of 35% at age 25 to a rate of 5% after age 70 (East Likelihood = 23463, * = -.03, Odds Ratio = .32, 95% Confidence Interval = 0.25-0.41).

Discussion

These results provide strong evidence that older participants in vocational rehabilitation are different from their younger counterparts in a variety of important ways. With respect to the clinical problems participants face, their goals for participation, their incentives and disincentives for work, how they participate in the program, and what happens at discharge, significant effects for age are found. Taken together, the results suggest that older adults are much less likely to use vocational rehabilitation as a means of obtaining competitive employment. Instead, they are likely to use it as an opportunity to obtain the benefits of paid work, while not having to do so in competitive jobs which they view as not well adapted to their medical needs.

One of the factors which makes this shift in motivation possible is the difference in income contingencies. Adults over the age of 55 are much more likely to have retirement or disability income which would serve to reduce their financial need to work. Futhermore, disability income, and until the year 2000 Social Security retirement income, typically has disincentives for earned income which actively discourage participants from competitive employment. The balance of incentives and disincentives for competitive employment would be expected to shift dramatically for most people between the ages of 55 and 65 and likely results in older adults approaching vocational rehabilitation with a different set of goals.

From one perspective, these data indicate that many older adults are using vocational rehabilitation services inappropriately. The data suggest that most older adults participate for the benefits of participation and not as a means of achieving competitive employment. To the degree to which vocational rehabilitation is conceptualized as a means for helping someone transition to competitive employment, participating for the benefits of working within the program alone may not be considered valid. It is important to note, however, that these data do suggest that older adults are seeking to accomplish clinical objectives. Eighty-four percent of participants over the age of 55 endorsed at least one clinically related goal, such as working in a setting appropriate to medical or psychiatric needs, maintaining sobriety, maintaining participation in treatment programs. While they may be uninterested in competitive employment, these goals do fit under most common definitions of rehabilitation (Martin, 1999; Wright, 1980), as well as within the official mandate of the CWT program-`to help veterans meet their highest level of functioning' (Losardo, 1999).

Historically, many VA and non-VA vocational rehabilitation programs have been designed to provide specialized work settings adapted for clinical needs without an immediate focus on competitive employment. Until the last 10-15 years, extended periods of work in sheltered workshops was a common form of vocational rehabilitation (Lassiter & Lassiter, 1999). To the degree that these settings have failed to help participants reach their highest level of vocational functioning, and have been a means of marginalizing adults with disabilities, they have been appropriately criticized (Noble et al., 1997). However, the current data suggest that older adults may be one subgroup of participants who are not seeking to transition to competitive employment, and whose goals may be appropriately served by working within a specialized setting.

These data suggest a number of reasons to consider the development of separate programs or subprograms for adults who want vocational rehabilitation services but are not seeking competitive employment. First, as programs increasingly focus on competitive employment as the primary goal, they are less able to help participants pursuing goals other than competitive employment. For example, older adults who are not interested in competitive employment may not participate if working in a competitive job is the only type of work activity. Second, adults who are not seeking competitive jobs are not likely to use resources provided to support the transition to competitive employment. Therefore, including these adults in programs focusing on competitive employment will make these programs less efficient. Third, to the degree that programs are increasingly evaluated by the percentage of participants who obtain and maintain competitive employment, a disincentive will be created for programs to serve adults pursuing other goals. For example, older participants will be more likely to reduce program success rates and may therefore be formally or informally discouraged from participation. Creating a separate track or specialized program for adults pursuing other goals will allow program staff to focus on the unique constellation of needs and objectives of different participant subgroups, and thus improve the degree to which older adults, among others, are served. Program evaluation efforts can reflect these differences by using outcome indices which are more appropriate for the range of participants, including older adults.

Several limitations of the study should be acknowledged. The generalizability of these results to the larger population of vocational rehabilitation participants needs to be carefully considered. Given the use of a veteran population entering a VHA rehabilitation program, likely sample biases include a largely male population, receiving clinical treatment at a VA, and receiving compensation from the VA. Given the differences between the Bedford sample and the larger sample, generalizability must again be carefully considered for the analyses using this subsample. Because there are no restrictions on admission to CWT based on age, clinical condition, or desired outcome, the sample may be more diverse :in terms of those variables relative to participants entering programs with more specific admission criteria or targeted outcomes. Due to the archival nature of the data, the range of variables is limited. For example, it is difficult to interpret the finding that older participants are more likely to participate in a sheltered workshop without additional data. Additional data may help clarify whether their increased participation in workshops is due to their own preference or whether they are relegated to these settings for other reasons, such as staff perceptions of employability. The goals questionnaire used in the Bedford sample was developed using input from CWT participants and may reflect options that are more common among the 90% of participants below the age of 55. Goals which may be important for older adults may not be fully represented, limiting these analyses. Finally, the number of older adults in the Bedford subsample was small (n = 34), suggesting that results for hypothesis 3 should be replicated in a larger sample.

Further study is needed to understand how and when older adults participate in vocational rehabilitation and to more clearly understand their feelings about competitive employment. Studies which could clarify the circumstances in which competitive employment would be attractive for older participants would be valuable. Additional study is needed to clarify whether the finding regarding lower pay and participation in the workshop is related to participant preferences or barriers created by the program, possibly as a function of staff perceptions of "employability". Studies documenting the broader benefits of competitive work in older adults should be replicated with older participants working within rehabilitation programs. Finally, studies examining the potential interest of older non-participants in using vocational rehabilitation services could clarify whether they represent an under-served population. Together, these studies will help clinicians and administrators insure that a potentially large and growing group of consumers receive services that address their specific needs.

Appendix

Goals for Participation in Vocational Rehabilitation Questionnaire

What do you hope to gain from participation in CWT?
  --In the left-hand column, check all the answers that apply
  --In the right-hand column, check only the one the answer that is
    most important to you.

Goal                                           All Goals   Primary Goal

Learn job skills/get on the job training          []            []
Have a temporary job until I find other work      []            []
Keep busy/have something to do during
  the day                                         []            []
Earn money/pay off debts/build up savings         []            []
Help get or keep housing                          []            []
Work at a site which lets me meet my family
  responsibilities                                []            []
Help me to get a competitive job                  []            []
Be able to work while keeping medical
  appointments or staying in treatment            []            []
Work in a setting that helps maintain
  sobriety                                        []            []
Build confidence in my ability to work            []            []
Work in a setting that helps me cope w/
  emotional/psychiatric problems                  []            []
Work in a setting appropriate for my medical
  or physical limitations                         []            []
Work in a less stressful setting                  []            []
Not sure                                          []            []
Other (please explain)                            []            []
Table 1: Demographics and Diagnosis

                                        National Sample

Variable                                     Sample

n                                            27799
Years of Education                           12.6 (1.6)
Gender--Male                                 96%
Race
    Caucasian                                50%
    African-American                         43%
    Hispanic                                  4%
    Other                                     3%
Marital Status
    Married                                   9%
    Separated/Divorced                       59%
    Widowed                                   3%
    Never Married                            29%
Receiving Non-wage Income
    Retirement Income                         4%
    SS or VA Disability Income               28%
Diagnosis
  Any Psychiatric Disorder                   50%
    Schizophrenia                             6%
    Other Psychotic Disorders                 2%
    Bipolar                                   7%
    Affective Disorder (not Bipolar)         17%
    PTSD                                     15%
    Anxiety Disorder (not PTSD)               9%
Any SUD                                      84%
    Alcohol Abuse/Dependence                 73%
    Drug Abuse Dependence                    55%
  Any Disabling Medical Disorder             45%
Employed, 3 Years Prior to Admission         69%
Competitively Employed at Discharge          30%

                                            National Sample

                                               Age Group

Variable                                  20-54         55+

n                                         25297 (91%)   2502 (9%)
Years of Education                        13.0 (1.8)    12.6 (1.5)
Gender--Male                              96%           98%
Race
    Caucasian                             48%           72%
    African-American                      45%           22%
    Hispanic                               4%            2%
    Other                                  3%            4%
Marital Status
    Married                                9%            9%
    Separated/Divorced                    58%           65%
    Widowed                                2%            9%
    Never Married                         30%           14%
Receiving Non-wage Income
    Retirement Income                      3%           13%
    SS or VA Disability Income            27%           33%
Diagnosis
  Any Psychiatric Disorder                50%           44%
    Schizophrenia                          6%            6%
    Other Psychotic Disorders              2%            2%
    Bipolar                                8%            7%
    Affective Disorder (not Bipolar)      17%           16%
    PTSD                                  16%            7%
    Anxiety Disorder (not PTSD)            9%            9%
Any SUD                                   85%           73%
    Alcohol Abuse/Dependence              74%           68%
    Drug Abuse Dependence                 59%           21%
  Any Disabling Medical Disorder          44%           61%
Employed, 3 Years Prior to Admission       0%           54%
Competitively Employed at Discharge        1%            2%

                                        Bedford Subsample

Variable                                     Sample

n                                            228
Years of Education                           12.7 (1.6)
Gender--Male                                 98%
Race
    Caucasian                                79%
    African-American                         15%
    Hispanic                                  4%
    Other                                     4%
Marital Status
    Married                                   8%
    Separated/Divorced                       55%
    Widowed                                   2%
    Never Married                            35%
Receiving Non-wage Income
    Retirement Income                         2%
    SS or VA Disability Income               34%
Diagnosis
  Any Psychiatric Disorder                   56%
    Schizophrenia                             4%
    Other Psychotic Disorders                 2%
    Bipolar                                  11%
    Affective Disorder (not Bipolar)         32%
    PTSD                                     23%
    Anxiety Disorder (not PTSD)              19%
Any SUD                                      82%
    Alcohol Abuse/Dependence                 73%
    Drug Abuse Dependence                    47%
  Any Disabling Medical Disorder             54%
Employed, 3 Years Prior to Admission         57%
Competitively Employed at Discharge          17%

                                            Bedford Subsample

                                                 Age Group

Variable                                    20-54       55+

n                                           194 (85%)   34 (15%)
Years of Education                         12.7 (1.3)   12.7 (1.56)
Gender--Male                                97%         100%
Race
    Caucasian                               80%          75%
    African-American                        13%          25%
    Hispanic                                 3%           0%
    Other                                    4%           0%
Marital Status
    Married                                  7%          13%
    Separated/Divorced                      54%          60%
    Widowed                                  2%           6%
    Never Married                           37%          20%
Receiving Non-wage Income
    Retirement Income                        0%          11%
    SS or VA Disability Income              34%          38%
Diagnosis
  Any Psychiatric Disorder                  58%          44%
    Schizophrenia                            4%           3%
    Other Psychotic Disorders                2%           3%
    Bipolar                                 12%           0%
    Affective Disorder (not Bipolar)        35%          18%
    PTSD                                    25%           8%
    Anxiety Disorder (not PTSD)             19%          21%
Any SUD                                     84%          71%
    Alcohol Abuse/Dependence                74%          68%
    Drug Abuse Dependence                   50%          32%
  Any Disabling Medical Disorder            53%          61%
Employed, 3 Years Prior to Admission        59%          41%
Competitively Employed at Discharge         17%          15%
Table 2: Age and Participation in Vocational Rehabilitation

Age                 Length of Stay          Total           Hours per
Group       n           (days)            Earnings            Week

<25          81      84.6 (122.1)       $1386 ($2500)      21.6 (12.9)
25-29       565     101.4 (128.0)       $1822 ($2669)      25.8 (16.1)
30-34      2009     109.4 (142.4)       $2023 ($2939)      26.2 (15.6)
35-39      5120     108.1 (136.2)       $2006 ($3209)      26.5 (30.1)
40-44      7185     116.3 (142.8)       $2153 ($3142)      26.1 (21.7)
45-49      6760     127.3 (156.9)       $2362 ($3458)      26.2 (26.2)
50-54      3603     132.8 (157.3)       $2419 ($3242)      25.5 (24.2)
55-59      1518     139.6 (178.6)       $2472 ($3959)      25.2 (16.9)
60-64       611     150.7 (181.4)       $2294 ($3420)      24.2 (15.3)
65-69       219     144.1 (176.8)       $1757 ($2849)      21.8 (19.2)
70-74        95     197.8 (252.0)       $1381 ($2761)      16.9 (16.5)
>75          33     122.8 (152.5)       $1303 ($2964)      13.3 (10.2)

Age                  Mean Hourly         Percent In
Group                    Wage             Sheltered
                                          Workshop

<25                 $4.53 ($1.47)            49%
25-29               $4.83 ($1.94)            34%
30-34               $4.83 ($2.06)            32%
35-39               $4.82 ($1.93)            33%
40-44               $4.79 ($1.64)            33%
45-49               $4.79 ($2.97)            35%
50-54               $4.84 ($2.12)            34%
55-59               $4.65 ($1.78)            39%
60-64               $4.36 ($1.59)            46%
65-69               $4.01 ($2.42)            61%
70-74               $2.94 ($1.82)            72%
>75                 $3.31 ($2.27)            85%

References

(2001). What Are The Attitudes of Young Retirees and Older Workers. Washington, DC: National Academy on An Aging Society.

Aquino, J. A., Russell, D. W., Cutrona, C. W., & Altmaier, E. M. (1996). Employment status, social support, and life satisfaction among the elderly. Journal of Counseling Psychology, 43(4), 480-489.

Drake, R. E., McHugo, G. J., Becker, D. R., Anthony, W. A., & Clark, R. E. (1996). The New Hampshire study of supported employment for people with severe mental illness. Journal of Consulting and Clinical Psychology. 64(2), 391-399.

Drebing, C. E., Rosenheck, R., & Penk, W. (2001). CWT National Survey: Summary of Project Director Responses. Bedford, MA: New England MIRECC.

Drebing, C. E., Van Ormer, E. A., Schutt, R. K., Krebs, C., Losardo, M., Boyd, C., Penk, W., & Rosenheck, R. (in submission). Client Goals for Vocational Rehabilitation: Distribution, Sources, and Consequences. Psychiatric Rehabilitation Journal.

Jacobs, R. H. (1978). Re-employment and unemployment in old age. Journal of Geriatric Psychiatry, 11 (1), 79-80.

Kashner, T. M., Rosenheck, R., Campinell, A. B., Suris, A., Crandall, R., Garfield, N. J., Lapuc, P., Pyrcz, K., Soyka, T., & Wicker, A. (in submission). Impact of work on health status among homeless, substance dependent veterans: A randomized controlled trial. Archives of General Psychiatry.

Kaye, L. W., Alexander, L. B., & Kauffman, S. (1999). Factors contributing to job quality and satisfaction among ethnically diverse, lower income, elderly part-timers. Journal of Gerontological Social Work, 31 (1-2), 143-166.

Lassiter, R. A., & Lassiter, M. H. (1999). The development of the rehabilitation enterprise in America: An early history of the rehabilitation movement. In G. L. Gandy, E. D. Martin, & R. E. Hardy (Eds.), Counseling in the Rehabilitation Process: Community Services for Mental and Physical Disabilities (pp. 52-73). Springfield, IL: Charles C. Thomas Publisher Ltd.

Levinson, D. J. (1986). A conception of adult development. American Psychologist. 41 (1), 3-13.

Losardo, M. L. (1999). Veterans in need of services: The history and development of the Edith Nourse Rogers Memorial VA Hospital Compensated Work Therapy program. University of Massachusetts, Boston.

Maloney, T. W., & Paul, B. (1990). Enabling Older Americans to Work. New York:: The Commonwealth Fund.

Martin, E. D. (1999). Foundations of Rehabilitation. In G. L. Gandy, E. D. Martin, & R. E. Hardy (Eds.), Counseling in the Rehabilitation Process (pp. 5-31). Springfield, IL: Charles C. Thomas Publisher LTD.

Mor-Bank, M. E., Scharlach, A. E., Birba, L., & Sokolov, J. (1992). Employment, social networks, and health in the retirement years. International Journal of Aging and Human Development, 35(2), 145-159.

Mutchler, J. E., Burr, J. A., Massagli, M. P., & Pienta, A. (1999) Work transitions and health in later life. Journals of Gerontology: Series B: Psychological Sciences and Social Sciences, 54B(5), S252-S261.

Noble, J. H., Honberg, R. S., Hall, L. L., & Flynn, L. M. (1997). A legacy of failure: The inability of the federal-state vocational rehabilitation system to serve people with severe mental illnesses. Arlington, Virginia: National Alliance for the Mentally Ill.

Seibyl, C. L., Rosenheck, R., Corwel, L., & Medak, S. (2000). Third Progress Report on the Compensated Work Therapy (CWT)/Veterans Industries (VI) Fiscal Year 1999. WestHaven, CT: Northeast Evaluation Program.

Soumerai, S. B., & Avorn, J. (1983). Perceived health life satisfaction, and activity in urban elderly: A controlled study of the impact of part-time work. Journal of Gerontology. 38, 352-362.

Taylor, H., & Leitman, R. (1989). Older Americans: Ready and Able to Work. New York:: Louis Harris and Associates.

United States Census Bureau. (2000). Census 2000 Summary Files: QT.P# Race and Hispanic or Latino for United States and Massachusetts. Available: http://factfinder.census.gov.

Wright, G. N. (1980). Total Rehabilitation. Boston: Little Brown. Xie, H., Dain, B. J., Becker, D. R., & Drake, R. E. (1997). Job tenure among persons with severe mental illness. Rehabilitation Counseling Bulletin, 40(4), 230-239.

Charles E. Drebing, Ph.D., Psychology Service--Bedford VAMC 200 Springs Rd., Bedford, MA 01730. Email: [e-mail]

Charles E. Drebing
Boston University School of Medicine

Marylee Losardo
Alice Van Ormer
Christopher Krebs
Walter Penk
Elizabeth Nasser

Bedford VA Medical Center

Heidi Ray
Amherst College

Catherine Seibyl
Robert Rosenheck

Yale School of Medicine