Supported Employment What is the Evidence for This Evidence-Based

							  Supported Employment:
What is the Evidence for This
              u



 Evidence-Based Practice?
  Integrated Employment Institute
              UMDNJ
           Gary R. Bond
Indiana U. - Purdue U. Indianapolis
              11/28/01
          What Is an
Evidence-Based Practice (EBP)?

 Well-defined  approach
 Strong research evidence for
  effectiveness
 Track record showing it can be
  implemented in a variety of
  settings
   1998 RWJ Expert Panel
        Conclusion
Supported employment currently
is the only evidence-based
practice in the area of vocational
rehabilitation.
           (Drake, 2000)
Origins and Dissemination of SE
   Rehabilitation Act of 1986
   Drake and Becker (1993):
    Individual Placement and Support
    (IPS) -- Standardized principles
   Only 2% had access to supported
    employment in 1998 US survey
   Number of programs currently
    increasing (Bond, 2001)
Types of Jobs Obtained in
 Supported Employment
 Mainstream   job in community
 Pays at least minimum wage
 Work setting includes people who
  are not disabled
 Service agency provides ongoing
  support
 Intended for people with most
  severe disabilities
  Evidence-Based Principles
 Eligibilityis based on consumer choice
 Supported employment is integrated
  with treatment
 Competitive employment is the goal
 Rapid job search is used
 Job finding is individualized
 Follow-along supports are continuous
       Evidence for
 Supported Employment:
 Three Types of Evidence
 Day  treatment conversion studies
 Controlled studies
 Studies examining specific
  ingredients
      Evidence for
  Supported Employment

Day Treatment Conversion Studies
 Rehabilitative Day Treatment
 1000+   day treatment programs in US
 Work often seen a secondary goal
 Little evidence of long-term
  effectiveness (even in skill
  development or social functioning)
 Many clients stay for years
    Future of Day Treatment
 Resistanceto closing down programs:
–Financial: Centers would lose money
–Clinical: Fear clients will “fall apart”

          prospects for day treatment:
 Long-term
–Show usefulness or may be replaced
Day Treatment Conversions to
  Supported Employment
5 mental health centers in U.S.
 Discontinued day treatment
 Reassigned day treatment staff to new
  positions in center
 Replaced with supported employment

     Sources: Drake, 1994, 1996; Bailey, 1998; Becker, 2001
  Similar Results in All 5
 Day Treatment Conversions
 Large  increase in employment rates
 No negative outcomes (e.g., relapses)
 Consumers, families, staff liked change
 Overall, all former day treatment
  clients got out into community more
 Resulted in cost savings
    Evidence for
Supported Employment

    Controlled Studies
9 Controlled Studies of SE
  Drake (1996) -- New Hampshire
  Drake (1999) -- Washington, DC
  Mueser (2000) -- Hartford
  Lehman (in press) -- Maryland
  Bond (1995) -- Indiana
  Gervey (1994) -- New York
  Chandler (1997) -- California
  McFarlane (2000) -- New York
  Meisler (2000) -- South Carolina
      New Hampshire Study:
IPS vs. Group Skills Training (GST)
               (Drake, 1996)
   Compared    two well-regarded supported
   employment approaches:
    – IPS -- integrated at mental health
      center and used rapid job search
    – GST -- brokered program with group
      skills training before placement
          Design
 Two   mental health centers
  participated
 143 clients randomly assigned
  to two vocational conditions
 Data collected for 18 months
          Drake IPS/GST Study:
           18-Month Findings
                     IPS        GST
Comp etitive job    78 .1 %    40 .3 %   p<.0 01
20 -hr jo b         46 .6 %    22 .4 %   p<.0 03
Hou rs wor ke d      60 7       20 5     p<.0 01
Wage s             $3 ,3 94   $1 ,0 78   p<.0 01
Summary: Drake IPS/GST Study
   IPS did far better than GST on
    every objective employment
    measure
   GST did not have any measurable
    impact on career planning or work
    readiness
   Clients preferred IPS
     Drake (1999) Study
     in Washington, DC:
   Experimental Conditions
 IPS:  Supported employment
  following evidence-based principles
 EVR (“Enhanced Voc Rehab”):
  Brokered, stepwise approach using
  sheltered workshops
 (Enhanced = voc counselor added)
           Design
 152  clients randomly assigned
  to two vocational conditions
 Data collected for 18 months
 All clients received excellent
  clinical case management
 18-mo Competitive Employment
          Findings
                  IPS      EVR
Any comp job      61%       9%
Hours worked      322       28
Weeks worked       15        1
Days to first job 126      293
Mean % working 17% - 34%   <5%
            Findings:
   Competitive + Sheltered Jobs

                   IPS     EVR
Sheltered work       11%     71%
Any job              64%     78%
Total wages      $2,000  $2,005
DC Findings: Summary
 IPS had higher competitive
  employment rates
 EVR had higher rate of
  employment overall, if you
  include sheltered jobs
 Both groups earned similar
  wages, but sheltered workers
  worked more hours
        Interpretation

 Ifa rehab program offers
  sheltered employment, clients
  may participate at a high rate
 Sheltered jobs seem to
  substitute for competitive jobs
     Hartford Study
of Supported Employment
        (Mueser, 2000)
 204clients were randomly
 assigned to one of 3 groups:
  – IPS
  – Standard Services
  – Psychiatric Rehabilitation Center
            Standard Services
    Access to all standard vocational services
    Most clients linked with one of two
     programs:
    1. Off-site supported employment program
    2. Program offering protected jobs:
       o Government set-aside jobs
       o Enclaves providing on-the-job
         training and supervision
Psychiatric Rehabilitation Center
 Membership
 Prevocational training
 Transitional employment
 Competitive work with staff support
 Access to other rehabilitation
  activities
 Clubhouse atmosphere for informal
  socializing
           Number of people employed competitively (%)

80


70


60


50


40                                                        % employed (comp)



30


20


10


 0
     Standard             IPS          Psy Rehab Center
                 Mean total hours worked com petitively

400


350


300


250


200                                                       Mean hours worked (comp)



150


100


 50


  0
      Standard             IPS         Psy Rehab Center
      Figure 2. Competitive Employment Rates in 6 RCTs of Supported
                                   Employment
90%


80%


70%


60%


50%
                                                                             Supported    Control
                                                                             Employment
40%


30%


20%


10%


 0%

        Bond 95    Chandler   Drake 96    Drake 99   Gervey 94   McFarlane
                      97                                            00
    Evidence for Supported
    Employment Based on
      Critical Ingredients
 Studies examining general fidelity
 Studies examining specific program
  principles
    Working Hypothesis:

 Programs more closely
 implementing evidence-based
 practice have better outcomes
    Studies Examining
General Fidelity to Supported
        Employment
  Successful vs. Unsuccessful
   Programs (Gowdy, 2000)
 Surveyed   27 mental health centers in
  Kansas
 Identified 5 high-performing and 4
  low-performing centers based on
  employment rates
 Interviewed staff and consumers
   Integration of Supported
  Employment and Treatment
        Teams Crucial
 Inhigh performing programs, SE
  workers met often with case
  managers
 Case managers were involved in
  employment effort
   Other Gowdy Findings

 Higher performing programs
  –Did not use prevocational
   training
  –Emphasized consumer
   preferences in job selection
  Becker Study of 10 Vermont
Supported Employment Programs
   Conducted   site visits to all
    programs
   Rated each on fidelity of
    program implementation
   Examined overall employment
    rates for center
Vermont 10-Site Study: Correlations
 Between Fidelity and Employment
  Supported Employment Scale Correlations


  Total Scale                 .76**


  Factors
  Community-Based Services    .82**


  Employment Specialists      .69*
  Devoted Solely to Voc
  Zero-exclusion policy       .43
Studies Examining
 Specific Program
    Principles
     Evidence-Based Principles
1.   Eligibility is based on consumer
     choice
2.   Supported employment is integrated
     with treatment
3.   Competitive employment is the goal
4.   Rapid job search is used
5.   Job finding is individualized
6.   Follow-along supports are continuous
1. Eligibility Is Based on
   Consumer Choice
 Consumers  are not excluded
 because they are not “ready”
 or because of prior work
 history, hospitalization history,
 substance use, symptoms, or
 other characteristics.
Do Some Consumers Have Greater
   Potential for Employment?
   Many  factors predict poorer
    work outcomes:
    –Diagnosis
    –Disability payments
    –Negative symptoms
    –Cognitive deficits
    –Education
   But – Is this the key question?
 Do Client Characteristics Predict
Success in Supported Employment?
 Regardless   of client characteristics,
  clients generally do better in SE than
  in alternative programs:
   –Gender, ethnicity, diagnosis,
    hospitalization history, cognitive
    functioning, education
   –Substance use: Sengupta (1998)
  Problems with Traditional
Assessment for Work Readiness
 Screens  out people with mental
  illness at high rate
 Not cost effective
 Does not predict who can work
 Not tied to intervention

  Sources: Marshak (1990), Noble (1997),
              Anthony (1984)
    Self-Fulfilling Prophesy

      rating clients in
 Staff
 prevocational work crews rate
 them as less work-ready than
 when rating them in paid jobs.

   Sources: Bond & Friedmeyer, 1986;
         Schultheis & Bond, 1993
 2. Supported Employment Is
Integrated with Mental Health
          Treatment
   Employment  specialists meet
   frequently with the mental
   health treatment team.
 Evidence for Critical Role
of Integration of Vocational
and Mental Health Services
 In 7 controlled studies, the
  more successful program was
  integrated, while comparison
  program was not.
Qualitative Studies of Programs Using
       “Brokered” Approaches
   Breakdowns    in communication
   Referral process works poorly
   Meetings hard to schedule
   Clients perceived differently
   Meds, housing out of sync with voc
   Responsibility for follow-up unclear
   Employment staff may get caught up
    in crisis work (case manager role)
3. Competitive Employment
       Is the Goal
 Agencies devote resources to
 supported employment services
 rather than to day treatment or
 sheltered work.
 Evidence for Importance of
    Making Competitive
   Employment the Goal
 Day  treatment conversion
  studies
 DC and Hartford studies:
  Adding sheltered employment
  seems to dilute effort
4. Rapid Job Search Is Used
  Theservice agency avoids
  lengthy pre-employment
  assessment, training, and
  counseling.
             Issues in Stepwise
                Approaches

 What   do consumers want?



 Do   they make progress?
   Bond and Dincin (1986):
  Accelerated Placement Study
 Accelerated:   Immediate placement
  in temporary community jobs
 Gradual: 3 months of prevocational
  training first
        Thresholds Accelerated Placement Study:
              15-Month Outcomes (N = 107)

50%
45%
40%
35%
30%                                                   Accelerated
25%
20%                                                   Gradual
15%
10%
 5%
 0%
      Own Job   Transitional Prevocational   No Voc
                Employment Work Crews         Goals
     Role of Pre-Placement
Counseling Groups (Rogers, 2000)
  Experimental   study comparing
   –PVR: Psychiatric Vocational
     Rehabilitation (Career planning in
     classroom before job placement)
   –Control: Referral to state vocational
     rehabilitation counselor
  135 clients followed for 2 years
 Rogers(2000) -- Study Findings


          PVR Control
9 months 12%   12%
18 months 32%  36%
24 months 20%  26%
   5. Job Finding Is
     Individualized

 Jobfinding is based on
 consumers’ preferences,
 strengths, and work
 experiences, not on a pool of
 jobs that are available.
 Job Preference Studies

 Most  clients have stable and
  realistic job preferences.
 Clients matched to initial
  job preference stay in job
  twice as long as those not.

   (Becker, 1996; Gervey, 1995)
6. Follow-Along Supports
   Are Continuous and
     Time-Unlimited
 Supported employment staff
 continue to stay in regular
 contact with consumer and
 (when appropriate) the
 employer without arbitrary time
 limits.
         Employment Rates
  3.5 Years After Starting Supported
            Employment
Clients who continued receiving support:
        71% were still working

   Clients who discontinued support:
           28% still working

             (McHugo, 1998)
Supported Employment:
Limitations of the Research
    and of the Practice
Commonly Mentioned Limitations
   of Supported Employment
     Too  expensive
     Too hard to implement
     Too many clients do not benefit
     Jobs are too often short term
     Little impact outside of work
      Too Expensive?
 Annual  per-client cost of
  supported employment less
  than $2000
 Alternatives to supported
  employment (day treatment,
  sheltered work) often cost
  more
  Too Hard to Implement?
 Appears to be one of the easier
  evidence-based practices to
  implement (Drake)
 Program elements are clearly
  described in the literature
 Supported Employment
  Implementation Kit from
  SAMHSA EBP Project should
  help
Too Many Clients Do Not Benefit?

 What  about >30% of consumers in
  supported employment studies who
  do not obtain work?
 SE Clients Who Do Not Work
 Rate  decreased by informational groups
  (e.g., Drake, 1994)
 Not necessarily failures -- Some get new
  information and choose not to work
 In day treatment conversion studies, those
  who do not work benefit in other ways
  (Torrey, 1995)
 “Individualized job engagement” may help
  (Ahrens, 1999)
 Jobs Are Too Often Short Term?
 Similarly high turnover rates in entry
  level jobs for general working
  population (Adams-Shollenberger,
  1996)
 Jobs seen as transitions, OK to have
  multiple jobs (Becker, 1998)
 Educational options should also be
  considered (Unger, 1998)
 More attention needed on career
  planning (Baron, 2000)
Little Impact Outside of Work?
 Controlled  studies have generally
  found few experimental differences
  in nonvocational outcomes
 One important finding: Supported
  employment does not lead to higher
  rates of psychiatric hospitalization
  (as feared by clinicians)
     Indirect Impact on
   Nonvocational Outcomes
 Among     those who sustain
  competitive employment,
  nonvocational outcomes
  (self-esteem, symptom control,
  life satisfaction) are superior to
  those who do not work
 Not true for sheltered employment
             (Bond, 2001)
Change in Self-Esteem Over Time




    (Bond, 2001)
Impact of Work on Symptoms                                no work
                                                          minimal work
               46                                         sheltered work
                                                          competitive work
               44

               42
  BPRS Score




               40

               38

               36

               34

               32

               30
                    Baseline   6 months   12 months   18 months


(Bond, 2001)
      Overall Conclusion:
  Supported Employment is an
    Evidence-Based Practice
 Its effectiveness is supported by
  9 randomized controlled trials and
  3 day treatment conversion studies
 It is based on well-defined
  principles, each of which also has
  evidence for its importance