• Using an intent to treat analysis, subjects' maintained employed rate is measured using the Employment Index. [ Time Frame: 12 months ] [ Designated as safety issue: No ]
  

• Using an intent-to-treat analysis, the change from baseline to endpoint in PTSD symptoms is compared between groups using the Clinician Administered PTSD Scale (CAPS). [ Time Frame: 12 months ] [ Designated as safety issue: No ]
  

Hypotheses or Key Questions

I. Primary Hypothesis (obtained + maintained employment): Subjects assigned to IPS will have a higher maintained employed rate for the 12-mo observation period than their standard VRP counterparts.

• I.a. First Component Primary Hypothesis (obtained employment): Subjects assigned to IPS will have higher rates of obtained employment for the 12-mo observation period than the VRP subjects.
• I.b. Second Component Primary Hypothesis (maintained employment): Among those subjects employed, those assigned to IPS will have worked more weeks for the 12-mo observation period than the VRP subjects.

II. Secondary Hypothesis: Those subjects who obtain competitive employment will have significantly reduced symptoms of PTSD from baseline to endpoint compared to those subjects who do not obtain competitive employment.

III. Confirmatory and Correlational Hypotheses: Compared to VRP subjects, those subjects assigned to IPS will have:

• greater work intensity (# weeks, days, # hours, wages earned)
• higher total earnings
• greater improvement in quality of life outcomes

Supplemental hypothesis (moderator): Because of the superior integration with the mental health (MH) treatment team, social work services, and community resources that the IPS-SE employment specialist provides as part of this model, we hypothesize that the IPS-SE model will have greater success in the maintenance of competitive employment compared to standard VRP for veterans with PTSD challenged in one or more of these psychosocial domains. From a research perspective, a variable that characterizes subjects for whom a particular treatment will be effective is called a moderator of treatment.

The hypothesized moderators of IPS-SE and standard VRP for subjects with PTSD are: 1) Transportation, 2) Housing, 3) Financial Means, and 4) Family Care Burden. We hypothesize that these variables will have moderating effects for each of two outcomes: 1) number of weeks the participant is competitively employed during the 12-month study period and 2) their score on the Sheehan Disability Scale.

Several principles are important in defining a moderator. First, the moderator must precede treatment. In our RCT, these psychosocial domains are baseline characteristics of subjects enrolled in the study. Second, the moderator must be uncorrelated with treatment choice. This will be the case in this RCT due to a 1:1 randomization, which should balance baseline characteristics between the two groups. Third, a moderator of treatment must have a differential effect of treatment for each level of the moderator. The treatment effect (e.g., active vs. comparator) must be shown to be greater for one level of the moderator (i.e. IPS-SE will show a greater effect size than standard VRP in subjects with a moderator present). Also, an index that combines these domains will be explored and formulated to identify subjects that may benefit to greater extent from IPS-SE.