At a glance
Supported employment statistics from the Supported Employment Demonstration (SED) show a program that scaled across 30 sites, tracked fidelity closely, and produced sustained employment outcomes over three years (SED Final Process Analysis Report; SED Final Impact and Cost-Benefit Analysis Report).
The clearest pattern in the dataset is simple: implementation quality improved over time, staffing adapted, and employment remained materially stronger in the supported employment arms than in usual services (SED Final Process Analysis Report; SED Final Impact and Cost-Benefit Analysis Report).
Fast facts
- 30 sites delivered IPS supported employment services in the SED (SED Final Process Analysis Report).
- Average IPS fidelity increased from 4.4 of 5 in Year 1 to 4.7 of 5 in Year 3 (SED Final Process Analysis Report).
- 23 of 30 sites provided behavioral health services on site (SED Final Process Analysis Report).
- The study enrolled 2,960 eligible applicants from November 2017 through March 2019 (Enrollment in the Supported Employment Demonstration).
- The 3-year employment rate reached 74.0% for Full-Service participants and 74.4% for Basic-Service participants, versus 64.1% for Usual Services participants (SED Final Impact and Cost-Benefit Analysis Report).
- Total 3-year earnings averaged $17,925 for Full-Service participants, $17,556 for Basic-Service participants, and $13,547 for Usual Services participants (SED Final Impact and Cost-Benefit Analysis Report).
Table of contents
- What supported employment statistics measure
- Program scale and implementation
- Fidelity and staffing trends
- Enrollment and response numbers
- Employment outcomes over 3 years
- Earnings patterns and work history
- What the numbers suggest for supported employment programs
What supported employment statistics measure
Supported employment statistics are useful because they connect two parts of the same story: how faithfully a program is delivered and what participants experience afterward. In this dataset, the implementation side comes from the SED Final Process Analysis Report, while the outcome side comes from the SED Final Impact and Cost-Benefit Analysis Report, the Enrollment in the Supported Employment Demonstration report, and the SED PUF (SED Final Process Analysis Report; SED Final Impact and Cost-Benefit Analysis Report; Enrollment in the Supported Employment Demonstration; SED PUF).
That split matters. A site can enroll participants without delivering services consistently, and it can deliver services well without producing the expected outcome pattern. The SED numbers let you compare both sides at once.
Big number: The demonstration reached 30 sites across the United States and used repeated fidelity checks during the first three demonstration years (SED Final Process Analysis Report).
Why this dataset is useful
This set of supported employment statistics gives you:
- A large, multi-site implementation sample.
- Year-by-year fidelity and staffing changes.
- Participant enrollment counts and survey follow-up markers.
- Outcome comparisons across Full-Service, Basic-Service, and Usual Services groups.
That makes it easier to see not just whether supported employment worked, but how the service model behaved over time.
Program scale and implementation
The SED was not a small pilot. It used 30 sites across the United States to deliver IPS supported employment services (SED Final Process Analysis Report). That scale matters because supported employment programs often depend on local staffing, referral flow, and service coordination. A larger network can show whether a model holds up outside one or two favorable settings.
A few implementation details stand out:
- The fidelity assessment used a 25-item IPS fidelity scale at all 30 sites during the first three demonstration years (SED Final Process Analysis Report).
- The study also used an 18-item NCC fidelity scale in 2018 and 2019 (SED Final Process Analysis Report).
- 23 of 30 sites provided behavioral health services on site, which is three-quarters of the sites in the demonstration (SED Final Process Analysis Report).
- More than half of the 20 full sites, or 11 sites, had team leads fully committed to SED (SED Final Process Analysis Report).
These facts point to a program that was not only measured, but measured in a structured way. The repeated fidelity checks give the dataset a stronger implementation spine than a single endline snapshot would provide.
At a glance: implementation signals
| Indicator | Value | Source |
|---|---|---|
| Sites in the demonstration | 30 | SED Final Process Analysis Report |
| IPS fidelity scale items | 25 | SED Final Process Analysis Report |
| NCC fidelity scale items | 18 | SED Final Process Analysis Report |
| Sites with on-site behavioral health services | 23 of 30 | SED Final Process Analysis Report |
| Full sites with team leads fully committed | 11 of 20 | SED Final Process Analysis Report |
Fidelity and staffing trends
The fidelity pattern is one of the strongest signals in the dataset. Average IPS fidelity ratings rose from 4.4 out of 5 in Year 1 to 4.6 in Year 2 and 4.7 in Year 3 (SED Final Process Analysis Report). That is a steady increase, not a one-time jump.
The vocational unit meeting fidelity rating followed the same broad direction, moving from 4.1 in Year 1 to 4.7 in Year 3 (SED Final Process Analysis Report). In practical terms, the demonstration appears to have tightened service delivery as it matured.
Fidelity trend table
| Measure | Year 1 | Year 2 | Year 3 | Source |
|---|---|---|---|---|
| Average IPS fidelity rating | 4.4 | 4.6 | 4.7 | SED Final Process Analysis Report |
| Vocational unit meeting fidelity rating | 4.1 | not stated | 4.7 | SED Final Process Analysis Report |
Staffing data adds another layer. Average IPS specialist staffing rose from 2.1 positions in Year 1 to 2.9 in Year 2, then eased to 2.7 in Year 3 (SED Final Process Analysis Report). That suggests a program that expanded staffing early, then stabilized.
At the same time, IPS specialist time allocated to SED declined from 81% in Year 1 to 80% in Year 2, 77% in Year 3, and 66% in Year 4 (SED Final Process Analysis Report). That pattern is worth noting because staffing counts and time allocation do not always move together. A program can have enough positions on paper while the share of time devoted to the initiative gradually loosens.
Care manager and NCC time moved differently:
- Care manager time allocated to SED increased from 60% in Year 1 to 71% in Year 2, then to 77% in Year 3 and 76% in Year 4 (SED Final Process Analysis Report).
- NCC time allocated to SED was 64% in Year 1, 72% in Year 2, 67% in Year 3, and 63% in Year 4 (SED Final Process Analysis Report).
That is a useful contrast. Care manager commitment rose and stayed high, while NCC time peaked in Year 2 and then drifted down.
Staffing and time allocation snapshot
| Measure | Year 1 | Year 2 | Year 3 | Year 4 | Source |
|---|---|---|---|---|---|
| IPS specialist staffing | 2.1 | 2.9 | 2.7 | not stated | SED Final Process Analysis Report |
| IPS specialist time allocated to SED | 81% | 80% | 77% | 66% | SED Final Process Analysis Report |
| Care manager time allocated to SED | 60% | 71% | 77% | 76% | SED Final Process Analysis Report |
| NCC time allocated to SED | 64% | 72% | 67% | 63% | SED Final Process Analysis Report |
Turnover and continuity
The process team also tracked IPS specialist turnover. Twelve sites had no IPS specialist turnover from Year 1 to Year 2, 11 sites had no turnover from Year 2 to Year 3, 17 sites had no turnover from Year 3 to Year 4, and 20 sites had no IPS specialist turnover by the end of Year 4 (SED Final Process Analysis Report).
That matters because continuity is often a hidden driver of performance. A site can look stable in headline staffing counts while still losing expertise. Here, the later years look more stable than the earlier transition periods.
Enrollment and response numbers
The enrollment figures show how much participant traffic supported employment programs may need before results start to matter at scale. The SED recruited participants over 16 months, from December 2017 through March 2019 (SED Final Impact and Cost-Benefit Analysis Report).
Recruiters attempted to contact 21,003 applicants near participating community mental health agencies (Enrollment in the Supported Employment Demonstration). From there, the study enrolled 2,960 eligible applicants from November 2017 through March 2019 (Enrollment in the Supported Employment Demonstration). Among potentially eligible enrollees, 26.2% enrolled (Enrollment in the Supported Employment Demonstration).
The participant pool itself was large enough to support downstream analysis:
- The SED PUF contains data on 2,944 enrollees (SED PUF).
- Response rates among eligible participants stayed above 70% for the first two years of study enrollment (SED PUF).
- By Quarter 12, 65.3% of eligible enrollees completed the survey (SED PUF).
- By Quarter 12, 83 participants were deceased and 50 participants had formally withdrawn (SED Final Impact and Cost-Benefit Analysis Report).
Enrollment funnel table
| Stage | Count or rate | Source |
|---|---|---|
| Applicants contacted | 21,003 | Enrollment in the Supported Employment Demonstration |
| Eligible applicants enrolled | 2,960 | Enrollment in the Supported Employment Demonstration |
| Enrollment rate among potentially eligible enrollees | 26.2% | Enrollment in the Supported Employment Demonstration |
| Enrollees in the PUF | 2,944 | SED PUF |
| Survey completion by Quarter 12 | 65.3% | SED PUF |
Employment outcomes over 3 years
The most important supported employment statistics in the dataset are the outcome comparisons. Across 3 years, both supported employment arms outperformed usual services on the main employment metric.
The 3-year employment rate was 74.0% for Full-Service participants and 74.4% for Basic-Service participants, compared with 64.1% for Usual Services participants (SED Final Impact and Cost-Benefit Analysis Report). That is a clear gap, and it holds the supported employment arms above the usual-services baseline.
The year-by-year pattern adds texture:
- Year 1 employment was 53.9% for Full-Service participants, 58.1% for Basic-Service participants, and 50.1% for Usual Services participants (SED Final Impact and Cost-Benefit Analysis Report).
- Year 2 employment was 60.0% for Full-Service participants, 58.1% for Basic-Service participants, and 48.6% for Usual Services participants (SED Final Impact and Cost-Benefit Analysis Report).
- Year 3 employment was 54.3% for Full-Service participants, 53.2% for Basic-Service participants, and 42.9% for Usual Services participants (SED Final Impact and Cost-Benefit Analysis Report).
Employment comparison table
| Measure | Full-Service | Basic-Service | Usual Services | Source |
|---|---|---|---|---|
| Year 1 employment | 53.9% | 58.1% | 50.1% | SED Final Impact and Cost-Benefit Analysis Report |
| Year 2 employment | 60.0% | 58.1% | 48.6% | SED Final Impact and Cost-Benefit Analysis Report |
| Year 3 employment | 54.3% | 53.2% | 42.9% | SED Final Impact and Cost-Benefit Analysis Report |
| 3-year employment rate | 74.0% | 74.4% | 64.1% | SED Final Impact and Cost-Benefit Analysis Report |
What stands out in the employment data
A few points are especially notable:
- The supported employment arms are not separated by a huge amount at the 3-year mark. Full-Service and Basic-Service are close: 74.0% versus 74.4% (SED Final Impact and Cost-Benefit Analysis Report).
- Both supported employment arms sit well above Usual Services at 64.1% (SED Final Impact and Cost-Benefit Analysis Report).
- Year 2 is the strongest employment year for Full-Service at 60.0%, while Basic-Service stays at 58.1% in both Year 1 and Year 2 (SED Final Impact and Cost-Benefit Analysis Report).
- By Year 3, the usual-services group falls to 42.9%, which makes the gap to both supported employment groups more visible (SED Final Impact and Cost-Benefit Analysis Report).
Earnings patterns and work history
Employment rates tell only part of the story. Earnings data give another angle on the same supported employment statistics.
Total 3-year earnings averaged $17,925 for Full-Service participants, $17,556 for Basic-Service participants, and $13,547 for Usual Services participants (SED Final Impact and Cost-Benefit Analysis Report). That means the supported employment arms both outperformed usual services on earnings as well as on employment incidence.
Work status at enrollment also mattered:
- Participants working at enrollment had total 3-year earnings of $37,002 in Full-Service, $35,729 in Basic-Service, and $28,690 in Usual Services (SED Final Impact and Cost-Benefit Analysis Report).
- Participants not working at enrollment had $13,436, $12,395, and $10,272 respectively across the same three groups (SED Final Impact and Cost-Benefit Analysis Report).
Earnings comparison table
| Group | Total 3-year earnings | Working at enrollment | Not working at enrollment | Source |
|---|---|---|---|---|
| Full-Service | $17,925 | $37,002 | $13,436 | SED Final Impact and Cost-Benefit Analysis Report |
| Basic-Service | $17,556 | $35,729 | $12,395 | SED Final Impact and Cost-Benefit Analysis Report |
| Usual Services | $13,547 | $28,690 | $10,272 | SED Final Impact and Cost-Benefit Analysis Report |
The pattern here is straightforward: participants with employment at enrollment earned much more over 3 years in every group, but supported employment still showed higher earnings than usual services in both working and not-working baselines (SED Final Impact and Cost-Benefit Analysis Report).
What the numbers suggest for supported employment programs
These supported employment statistics point to a few practical takeaways.
First, implementation quality improved over time. IPS fidelity moved from 4.4 to 4.7, vocational unit meeting fidelity improved from 4.1 to 4.7, and site-level staffing stabilized enough to support sustained delivery (SED Final Process Analysis Report).
Second, staffing commitment was not static. IPS specialist staffing rose early and then adjusted, while care manager time devoted to SED climbed from 60% to the mid-70s and stayed there (SED Final Process Analysis Report). That suggests the program’s operating model evolved rather than staying fixed.
Third, outcomes were not just marginally positive. Over 3 years, both supported employment groups held a clear employment advantage over usual services, and that same advantage showed up in total earnings (SED Final Impact and Cost-Benefit Analysis Report).
Fourth, the dataset shows why supported employment is worth measuring across multiple dimensions. Enrollment volume, response retention, fidelity, staffing, and earnings all help explain the final employment picture. If you only looked at one layer, you would miss the shape of the program.
Fifth, the difference between Full-Service and Basic-Service is small in several places. In this dataset, the more important comparison is supported employment versus usual services, not one supported employment arm versus the other (SED Final Impact and Cost-Benefit Analysis Report).
For readers comparing supported employment models, the strongest signal is the combination of rising fidelity, high site coverage, and better participant outcomes. The SED numbers do not suggest a flashy one-metric win; they show a system that improved operationally and then translated that stability into stronger employment and earnings results across the study window.