Evaluating Educational and Employment Services for Young People With Psychiatric Conditions: A Systematic Review

The functional impact of serious psychiatric conditions often includes significant disruptions to normative developmental tasks related to establishing a career, such as completing postsecondary education and honing skills and interests through early exploratory work and training activities. Thus, individuals with psychiatric conditions often have poorer educational attainment than their peers without psychiatric conditions, including lower college enrollment rates (1, 2), lower course completion rates (3), and a greater likelihood of dropping out (4, 5). Most individuals with serious psychiatric conditions also experience substantial challenges in maintaining gainful employment, with studies of workforce participation in this population indicating rates of approximately 15%−25% at any one time (6–9). Given the clear association between education level and employment outcomes both in the general population and among individuals with psychiatric conditions (10, 11), it is likely that disruptions in the attainment of education contribute in part to the poor vocational outcomes among people with these conditions.

Over the past two decades, there has been a growing appreciation of the importance of providing services that explicitly address the barriers to functional recovery, including successful vocational engagement (12–16). This increased emphasis on providing effective vocational services for individuals with psychiatric conditions has been motivated in part by the large body of research on supported employment (SE), in particular the individual placement and support (IPS) model of SE; this work has unequivocally demonstrated that IPS can significantly improve competitive employment outcomes in this population relative to other vocational approaches (17–19). IPS emphasizes the rapid initiation of a job search for competitive employment based on client preferences, the provision of ongoing individualized job supports, and the close integration of SE and mental health services (20). The research indicating its effectiveness has helped to fundamentally change the attitudes of practitioners and researchers regarding the ability of people with psychiatric conditions to successfully pursue competitive employment (21) and the potential for work to contribute to the process of recovery.

Thus, the IPS model has represented a critical advance in how vocational rehabilitation services for individuals with psychiatric conditions are approached. In recent years, however, some researchers in the field have recognized that the standard implementation of IPS, with its steadfast focus on rapid search for competitive employment, may not adequately foster educational or training experiences that facilitate longer-term career development (22, 23). Expanding the focus of vocational services to include an explicit emphasis on human capital development, especially through the attainment of postsecondary education, could promote greater access to employment in the primary labor market. Compared with low-wage entry-level jobs, positions in the primary labor market should more effectively buffer individuals from the impact of their psychiatric condition by providing increased job security, benefits, and greater financial self-sufficiency (21, 24–26). Vocational services that support these career-building pursuits are especially critical for young people, in that focusing on the early acquisition of human capital should aid in minimizing disruptions to the goal and skill development that typically occur during young adulthood and support employability over the long term. In this way, providing services that emphasize the successful completion of these formative career-related tasks to young people with psychiatric conditions may minimize or even prevent the vocational disability and poverty that are so often associated with these conditions (15, 27, 28).

In accordance with this view and the recognition that young adults are often interested in pursuing educational goals, there have been recent efforts to combine supported education (SEd) and SE services for young people with psychiatric conditions (14, 18, 29, 30), especially in the context of early intervention programs for first-episode psychosis (25, 31–33). Although findings to date suggest that this approach can improve vocational outcomes among young people when considering the combination of “work and/or school” activities (30, 31, 34), results specifically regarding educational outcomes have been mixed (30, 31, 35). In addition, given the relatively short follow-up periods of most of these studies, it is not yet clear whether vocational services that include SEd effectively improve career-related outcomes among young people, such as completing postsecondary education and securing positions in the primary labor market (21).

To inform these burgeoning efforts, we sought to summarize the evidence regarding the effectiveness of specific vocational intervention approaches for improving educational and employment outcomes among young people with psychiatric conditions. Multiple reviews regarding the efficacy of vocational services for adults with psychiatric conditions have been published (17–19, 36), and Bond and colleagues (35) recently published a systematic review examining employment and educational outcomes associated with early intervention programs for psychosis; however, there have been few attempts to synthesize the results of studies evaluating vocational services for the broader population of young people with psychiatric conditions (37). To address this literature gap, we conducted a systematic review of controlled evaluation studies of vocational interventions for young people with psychiatric conditions. In particular, we sought to determine the extent to which services were tailored specifically to the needs of young people, including through the incorporation of educational supports, and their efficacy to promote gains in postsecondary education and employment. Of note, for the purposes of this review, we use the terms “vocational intervention” and “vocational services” interchangeably and define these as any psychosocial treatment or service that specifically targets educational or employment outcomes (e.g., SE and SEd).

This review did not include evaluation studies of multicomponent programs that offer a suite of services, such as early intervention services for psychosis. Although such programs often include vocational services, evaluations of these programs as a whole do not allow one to draw conclusions regarding the specific impact of vocational interventions. Furthermore, given our interest in services that support young people at various developmental stages to promote their success in postsecondary vocational pursuits, we included studies with adolescent or young adult samples, defined as ages 14–35 years. However, because our focus was on postsecondary outcomes, we did not include studies that report only secondary educational outcomes, such as performance in or completion of high school. For a discussion of the importance of secondary education services and related issues, see Bradley et al. (38) and Killackey et al. (39).

Methods

To ensure adherence to standard systematic review procedures, we worked with a consultant from the Campbell Collaboration, a research organization that coordinates the production of systematic reviews and other types of evidence syntheses, when designing the methods for this review. In particular, this consultant assisted us with specifying our research question and developing standardized methods, described below, for identifying and reviewing eligible sources. Because this project involved reviewing published research findings, no institutional review board approval was required.

Article Identification and Retrieval

The following databases were searched with a set of standardized terms: PubMed, PsycINFO, Web of Science, Academic Search Premier, and ERIC. The search terms were grouped into three categories: population (e.g., “psychiatric condition,” “mental disorder,” and “psychosis”), intervention (e.g., “supported education” and “supported employment”), and educational and employment outcomes and other topical terms (e.g., “employment” and “college”). (We provide a complete list of the search terms used in the online supplement to this article.) Search terms within each category were separated with “OR,” and categories were combined with “AND.” All databases were searched for records of research articles published in English from 2000 to the beginning of June 2020, with the most recent searches conducted on June 1 and 2, 2020. Additional articles included for evaluation were suggested by team members and colleagues with expertise in this area (see figure in the online supplement).

Study Selection and Data Extraction

We reviewed all research articles identified for potential inclusion with a two-phase evaluation process. In phase 1, we examined the title and abstract of each article and used relatively broad criteria to determine whether it should be advanced to phase 2, which involved reviewing the full text of the article with more restrictive criteria. Articles eligible for inclusion were peer-reviewed reports of original research that evaluated a vocational service specifically, examined postsecondary educational or employment outcomes specifically for youths or young adults (ages 14–35 years) with psychiatric conditions (excluding primary neurodevelopmental, substance use, and personality disorders), included statistical results of putative treatment-related effects on educational or employment outcomes, and included both a treatment and comparison group. Both randomized controlled trial (RCT) and quasi-experimental (nonequivalent group) study designs were permitted. A complete list of the inclusion criteria for phases 1 and 2 is provided in an appendix of the online supplement.

Each article was reviewed for potential inclusion by at least one trained team member. Questions and coding discrepancies were resolved through consensus during team meetings. All team members participating in the review of identified articles were trained by senior authors (J.L.T., M.G.M.) in evaluating these reports with the phase-1 and phase-2 inclusion criteria. These team members included faculty members, graduate students, and research staff from several research groups. Consistent with the participatory action research approach that guided the larger project of which this review was a part, some of the team members were young adults with direct experience in receiving mental health services. To establish interrater reliability among team members, the first and last authors (J.L.T., M.G.M.) initially coreviewed and discussed multiple sets of articles to develop a standardized approach to applying the inclusion criteria. Thereafter, we required that all other team members establish interrater reliability with the first author before working independently. This involved demonstrating ≥90% agreement with the first author on proceed-or-exclude decisions for at least two consecutive sets of 25 titles and abstracts for phase 1 and demonstrating ≥85% agreement on include-or-exclude decisions for at least two consecutive sets of 15 full-text articles for phase 2.

All studies retained for inclusion in the final set were reviewed in detail by the first author, who used a data extraction form to guide this review. As a quality and accuracy check, a second team member independently reviewed the included studies and related data extracted by the first author. All questions regarding potential discrepancies were discussed and resolved through consensus.

Outcome Measures

Eligible studies were required to report outcomes related to integrated postsecondary educational programs or competitive employment (see online supplement). We excluded studies for which the only outcomes reported were those related to performance in nonintegrated settings, such as segregated classrooms or sheltered workshops.

Results

Search Results and Included Studies

In total, 14,019 unique records of research articles were retrieved through the database searches, and an additional 103 were acquired on the basis of recommendations from team members and colleagues. Through our two-phase evaluation process, we identified a final set of 10 studies that met our inclusion criteria (see appendix and figure in the online supplement). Some of the most common reasons why studies did not meet the inclusion criteria were that they employed a general adult rather than an adolescent or young adult sample and did not report results separately for participants ages ≤35 years or that a vocational intervention was not evaluated.

We identified three studies (40–42) that assessed specific additions to, or components of, a given vocational intervention rather than the full intervention (e.g., the addition of motivational interviewing training for IPS staff [40]). Although these studies were not eligible for inclusion, we considered their results and summarize them in the online supplement, given their relevance to this review.

Of the 10 studies retained for review, five were based on RCTs (23, 30, 31, 43, 44), and another five were quasi-experimental studies (34, 45–48). Of note, we identified several additional studies (49–52) that were based on secondary analyses of data from RCTs included in our set; these studies were excluded because they overlapped with the retained RCTs (23, 31). The 10 vocational intervention studies included in this review are summarized in Table 1. (Additional details regarding the intervention and control conditions, treatment settings, and measurement of outcomes are provided in a table in the online supplement.)

Vocational Service Adaptations

A question of interest was the extent to which the interventions evaluated were tailored specifically to address the needs of young people, such as through the incorporation of SEd to assist with the attainment of education or an emphasis on career-building pursuits (e.g., career planning and vocational exploration activities). Six of the included studies assessed vocational services that were adapted in some way for this population (23, 30, 31, 34, 45, 46); all of these interventions were based on the IPS model of SE. We provide a brief overview of these vocational service adaptations in Table 2 and additional details regarding the interventions in the online supplement. Five of the six trials of vocational services adapted for young people were based in multicomponent treatment programs for recent-onset psychosis.

Vocational Outcomes

Given our aim to summarize the evidence regarding the effectiveness of the vocational services evaluated, we weighed the results of the five RCT-based investigations most heavily when considering the findings across studies. Four of the five RCT-based studies evaluated vocational services based on the IPS model, either in its standard form as an SE service for adults (43) or with adaptations for young people (23, 30, 31) (Table 2). The one RCT-based study that did not evaluate the IPS model in full, Burke-Miller et al. (44), examined data from a multisite RCT of SE for adults; the specific models on which services were based varied across study sites. Some of the sites implemented IPS, and all included, at a minimum, treatment components that aligned with several IPS features, namely, integrated vocational and clinical services, an emphasis on rapid job search for competitive employment based on client preferences, and the provision of ongoing vocational supports (16).

Only two of the five RCT-based studies required that participants be unemployed at baseline (43, 44), but all five included only participants who expressed at least a possible interest in working (31, 43, 44) or in working or pursuing education (23, 30). As indicated in Table 1, the studies varied in the diagnostic status of participants, the treatment settings in which the vocational interventions were delivered, and services provided to the control group (see also the table in the online supplement). The duration of service provision and follow-up also varied across studies but was relatively short (6–24 months) in all of the RCTs reviewed (Table 1).

Employment outcomes in RCT-based studies.

Despite the aforementioned variability across the RCT-based studies, all five observed a significantly higher employment rate among participants in the vocational intervention group than among those in the control condition. The four RCT-based studies of IPS interventions also assessed total duration of employment, and three of these studies found that the average number of weeks worked was significantly larger among participants receiving the IPS intervention than among those in the control group (30, 31, 43).

Burke-Miller et al. (44) observed differential treatment effects for their youth and young adult participant subgroups (Table 1), suggesting that youths in their late teens and early 20s may not benefit from the SE services evaluated to the same degree as young adults in their mid- to late 20s, at least in terms of securing competitive work. As the authors discuss, these results may reflect greater involvement in educational pursuits by their youth participants, which was not assessed in this study. Also, the relatively smaller sample size of the youth subgroup (Table 1) may have affected their results. Notably, these findings are not consistent with those of Bond et al. (43) (Table 1), who observed superior competitive employment outcomes among IPS participants in both their younger (ages 20–24) and older (ages 25–29) subgroups. Both of these studies evaluated SE services designed for adults receiving community mental health services (16, 43, 44), and both included young people with various psychiatric conditions. However, the two studies differed in several respects (e.g., in the type of SE assessed), which may have contributed to these discrepant results. (We discuss these issues further in the online supplement.) Of note, the results from the two RCTs of Killackey et al. (23, 31) are congruent with those of Bond et al. (43) in demonstrating IPS effects on employment outcomes among participants ages <26 years, specifically among young people receiving early intervention services for psychosis (Table 1).

In their most recent RCT of IPS, Killackey et al. (23) assessed both immediate intervention effects and persistence of treatment-related gains over a 12-month posttreatment follow-up (Table 1). Consistent with their previous RCT (31), they observed a significantly higher rate of employment among participants receiving IPS than among control participants over the 6-month intervention period (23). However, the groups did not differ significantly during the posttreatment follow-ups (Table 1). As the authors discuss, a complication for interpreting these results with regard to durability of IPS effects is the unexpectedly high employment rate attained by the control group in this sample. Both this and Killackey et al.’s earlier RCT of IPS were based in the Early Psychosis Prevention and Intervention Centre (EPPIC), a specialized program for first-episode psychosis. The authors speculate that among other factors, changes in EPPIC staff attitudes regarding their clients engaging in vocational pursuits, as well as an increase in the skills of their clinicians in providing general vocational support, may have contributed to the higher-than-expected employment rate in the control group in this recent IPS trial (23).

Employment outcomes in quasi-experimental studies.

Results from the three quasi-controlled trials of IPS-based interventions (34, 45, 46), all of which were tailored for young people (Table 2), are consistent with those of the aforementioned RCTs in pointing to superior employment outcomes among participants in the vocational intervention group (Table 1). One of these studies assessed Vocational Support Within Early Intervention (VIBE), an intervention that explicitly incorporated several IPS components (34) (see table in the online supplement). Participants who had access to VIBE were significantly more likely to begin or return to competitive work or education than those who had no access to VIBE (Table 1); however, the descriptive results provided by the authors suggest that this finding was driven by gains in employment rather than education (34). Likewise, the primary results of Hegelstad et al. (46) focused on a composite measure of education “and/or” competitive employment (Table 1). However, additional results provided for work and school separately indicate that participants who received their IPS-based intervention, Job- and SchoolPrescription, were more likely than participants in the control group to be employed ≥20 hours a week at the end of the 1-year intervention, suggesting specific intervention effects on employment. These results must be viewed tentatively, however, because the authors did not control for baseline group differences in education level in these additional analyses. Furthermore, as the authors discuss, their intervention and historical-control groups may have differed on other features as well (46). Of note, Hegelstad et al. also assessed outcomes during a 12-month posttreatment period, and their findings were similar to those of Killackey et al. (23), discussed above. Specifically, the superior vocational outcomes among Job- and SchoolPrescription participants observed at the end of the 1-year intervention were no longer evident at the 2-year follow-up (46) (Table 1).

The remaining two quasi-experimental studies (47, 48) reported employment results from large-scale evaluations of vocational services that were not explicitly aligned with IPS principles nor specifically adapted for young people. Honeycutt et al. (47) used archival data from state vocational rehabilitation (VR) agencies to examine employment outcomes associated with the receipt of college or vocational training support, such as financial assistance for educational expenses. Among young people with psychiatric conditions who received VR services (Table 1), both the receipt of college support and the receipt of vocational training support were significantly associated with being employed for at least 3 months at VR exit, suggesting that these types of VR supports (see table in the online supplement) may have a beneficial effect on employment outcomes. A caveat regarding these findings, however, concerns the nonrandom assignment of participants to specific VR services. Thus, participants who received such supports may have differed from those who did not on factors such as goals and perceived likelihood of benefiting from such services. Educational outcomes were not examined in this study.

Mattila-Holappa et al. (48) assessed employment outcomes over a 5-year period among young adults in Finland who were initially on work disability because of a psychiatric condition. Participants whose treatment plans specified both work-oriented interventions (see table in online supplement) and psychotherapy were significantly more likely to be employed at follow-up (Table 1), and planned work-oriented interventions that included rehabilitative courses and training were associated with a significantly earlier entry into work. Limitations of this study were that the authors examined planned interventions without verification of service engagement and, similar to the study by Honeycutt et al. (47), that participants for whom work-oriented interventions were planned may have differed from those for whom no interventions were specified.

Educational outcomes in RCT-based studies.

As indicated in Table 1, three of the RCTs (23, 30, 31) assessed both employment and educational outcomes; all were evaluations of adapted IPS-based interventions that included educational supports (Table 2). In their first RCT of IPS, Killackey et al. (31) reported significantly greater gains on a combined measure of employment and education (i.e., work, school, or both) among IPS participants. However, the details provided in this report strongly suggest that this effect was driven by the number of participants who became employed rather than enrolled in education. Thus, in contrast to the results the authors observed for employment, neither this nor their later RCT (23) yielded evidence for treatment effects on educational engagement (Table 1).

Results from an 18-month RCT conducted by Nuechterlein et al. (30), however, indicate treatment-related gains in school participation (Table 1). IPS was expanded to incorporate SEd and was combined with the Workplace Fundamentals Module (WFM [53]), a group-based skills training program that was adapted to promote skill development for both school and work. Participants were randomly assigned to either this IPS-WFM intervention or to brokered VR (i.e., VR provided by an outside agency) plus a social skills training group, which served as the control condition (see table in the online supplement). In addition to experiencing superior employment outcomes, IPS-WFM participants were significantly more likely than participants in the control group to return to school during the first 6 months (when treatment was most intensive), and the mean number of weeks in which they were in school across the 18-month follow-up was significantly greater than that of participants in the control group (30). Of note, >80% of those participating in school were in postsecondary education (primarily 2- and 4-year colleges).

Educational outcomes in quasi-experimental studies.

Two of the five quasi-controlled trials also assessed educational outcomes (34, 46); both evaluated adapted IPS-based services (Table 2). Of note, however, they both used a composite measure of employment and education in their primary analyses (Table 1). Whereas the results of Major et al. (34) did not yield evidence of specific effects on education, as noted above, those of Hegelstad et al. (46) suggested that compared with a historical-control group, participants who received Job- and SchoolPrescription were more likely to be in school ≥20 hours a week at the end of the 1-year intervention. However, as with the employment-specific outcomes in this study, these education results must be viewed with caution, in part because the authors did not control for baseline group differences in education level for their separate work and school analyses (46). At the same time, both the education and employment results of Hegelstad et al. are encouraging and warrant further study of this novel IPS adaptation.

Discussion

The most consistent finding that emerged from this review was that services based on the IPS model were significantly more effective at improving employment outcomes among young people with psychiatric conditions than both treatment as usual and comparison vocational services (e.g., VR and unspecified brokered SE). These results are congruent with a large body of work evaluating IPS for adults with serious psychiatric conditions (17, 19, 36). Moreover, across the studies reviewed, these findings were observed among both recent-onset psychosis and mixed psychiatric diagnosis samples and with IPS adapted specifically for young people and as delivered in its standard form as an SE service for adults. In contrast, the smaller set of studies that specifically assessed educational outcomes yielded mixed results, with only one RCT of an adapted IPS-based intervention for young adults with recent-onset schizophrenia providing clear evidence of treatment-related gains in education (30).

The substantial treatment effects of IPS-based services on employment rates observed in several of the trials in our set (e.g., 30, 31, 43), along with the consistency of these effects across the studies reviewed, are notable. Given the relatively short follow-up periods of these studies, a critical area for further inquiry is the longer-term vocational functioning of young people receiving IPS-based services. In addition to examining the stability of employment, it will be important for future longitudinal studies to incorporate assessments that probe aspects of career development over time, including occupational advancement, the attainment of personally meaningful work, and financial self-sufficiency. Further research is also needed to determine whether IPS adaptations that include components specifically designed to enhance functioning and persistence of young people in vocational settings, such as Nuechterlein et al.’s IPS-WFM intervention (30), yield greater gains over the long term compared with more standard approaches.

Relatedly, given that neither of the two studies that evaluated vocational outcomes during posttreatment follow-ups found evidence for the persistence of IPS effects on employment beyond the intervention period (23, 46), additional research is needed on the impact of the duration of IPS-based services in this population. For example, it is possible that the short duration of IPS in the Killackey et al. RCT (23) (i.e., 6 months) contributed to the apparent loss of the IPS treatment effect during the postintervention follow-up period (see also Hegelstad et al. [46]). One of the principles of IPS is that employment supports are individualized and ongoing (54). However, community mental health providers often face resource constraints that hinder the provision of such services (55), and programs targeting young people in particular are typically time limited. Thus, a better understanding of these factors should inform efforts to develop and deliver employment supports to maximize long-term vocational success.

Given the mixed findings across the five studies of IPS-based interventions that examined school participation (23, 30, 31, 34, 46), we cannot draw firm conclusions regarding the effectiveness of this approach to promote gains in education specifically (see also Bond et al. [35]). However, the results of Nuechterlein et al. (30) suggest that IPS-based services that are adapted to include comprehensive educational supports and skills training can significantly improve both employment and educational outcomes in this population relative to brokered VR services. Although additional research is needed to determine which element(s) of the IPS-WFM intervention (14, 30) contributed to gains in school participation, considering this intervention alongside those evaluated by the studies reviewed that did not yield improvements in education (23, 31, 34) further suggest that adapting IPS to include these additional components may help promote educational attainment in this population (see Discussion section in Nuechterlein et al. [30]). Specifically, based on the intervention descriptions in these reports, we note that the SEd provided as part of IPS-WFM (14, 30) may have been more comprehensive than the educational supports of these other IPS-based interventions (23, 31, 34). Likewise, it seems that the WFM component (30) may have been a more formalized and systematic approach to vocational skills training than that provided by the other services (23, 27, 34). Given the tentative but encouraging education-specific results of Hegelstad et al. (46) discussed above, it is worth noting that like Nuechterlein et al. (30), Hegelstad et al.’s IPS-based intervention included a formal incorporation of SEd, along with staff with SEd-specific expertise. Ultimately, however, because the reports of these studies (14, 23, 27, 30, 31, 34, 46) varied considerably with regard to how much information they provided about specific educational supports and training components, it is difficult to confidently compare and contrast them.

As discussed by Nuechterlein et al. (30), it is also possible that differences in the samples of these studies (23, 30, 31, 34, 46), such as in the proportion of participants interested in education (vs. only in employment) or clinical characteristics (23, 30, 34), contributed to the mixed education results. For example, there is evidence that the IPS sample of Nuechterlein et al. (30) may have had a greater proportion of participants interested in pursuing education than the sample of Killackey et al. (23, 31, 56). Despite these caveats, the findings of Nuechterlein et al.’s well-controlled trial (30) provide the field with important guidance for efforts to develop services that promote gains in higher education among young people with psychiatric conditions.

Relatedly, our literature search identified several other sources that assessed, in some form, SEd services or combined SE-SEd programs (e.g., 25, 29, 57, 58). These services included IPS adaptations for young people that emphasized the strengthening of relevant skills and the attainment of education as a means of promoting career development. Although these sources did not meet our inclusion criteria because of the outcomes reported or the lack of a control group, they provide valuable information on how to effectively tailor services to this population, including for young people with a range of psychiatric symptoms and life experiences. At the same time, they highlight the need for more rigorous evaluations of these promising approaches through well-controlled trials. Likewise, we identified several studies of early intervention programs, including the RAISE NAVIGATE Program, Connection Program, and OnTrackNY, that describe the implementation of IPS-based SE-SEd (SEE) in the context of multidisciplinary recovery-oriented services (32, 33, 59, 60). Although the designs of these studies do not allow one to draw conclusions regarding specific effects of SEE, their results suggest a positive impact of this intervention on vocational functioning and can provide guidance on strategies to maximize service engagement and related outcomes (32). It will be important for future work not only to add to the empirical base of Nuechterlein et al. (30) and the other studies in our set through RCTs that assess specific treatment effects on work and school outcomes, but also to extend these findings by evaluating whether such approaches effectively promote outcomes related to longer-term academic success and career development, such as graduation rates and movement into the primary labor market.

As summarized in the online supplement, we identified one RCT that evaluated the impact of training early intervention staff who were providing IPS in motivational interviewing (MI) techniques (40). The aims of the MI training were both to address the ambivalence and fears that staff themselves may hold about their clients pursuing competitive employment and to equip staff to implement MI-based strategies with their clients. The results of this study suggest that training IPS clinicians in MI for these purposes can improve both employment and educational outcomes among the young people they serve. Such findings are congruent with recent speculations of Killackey et al. (23) concerning the potential influence of changes in EPPIC staff attitudes regarding vocational engagement among their clients on the results observed in their recent IPS trial (see also 46, 61).

A limitation of this review was that we did not conduct a meta-analysis to quantitatively evaluate outcomes. The decision to conduct a qualitative review only was based on the heterogeneity across studies in the services evaluated and length of follow-up. Furthermore, we restricted our review to RCTs and controlled quasi-experimental studies. Our requirement that studies include a comparison group aligned with our aim to summarize the evidence regarding the effectiveness of vocational services for young people, but it limited the formal review to a relatively narrow set of findings. We did not include a formal evaluation of study quality, although we weighed the consideration of results on the basis of the strengths and notable caveats related to study design. In addition, we did not include unpublished or gray literature, which increased the risk for publication bias.

Furthermore, we focused on a relatively narrow set of interventions and outcomes, without considering other relevant intersecting factors such as immigration status, discrimination on the part of employers and education providers, family involvement, comorbid substance use disorders, socioeconomic status, and disability benefits status. Regarding the latter, multiple findings suggest that receipt of disability benefits, such as Supplemental Security Income, may have persistent detrimental effects on workforce participation (44, 62, 63) and, furthermore, that providing services that help people with psychiatric conditions avoid the need for initial enrollment may be the most effective means of reducing long-term dependence on these benefits (64). Thus, systematic work to evaluate the effectiveness of interventions to reduce early reliance on disability programs will further inform efforts to develop services that improve long-term vocational trajectories in this population.

Finally, we did not consider organizational- and system-level barriers to community implementation of SE-SEd services for transition-age youths and young adults, such as funding and billing issues (65). As noted above, limitations of the literature reviewed also affected the conclusions that could be drawn, including the variability across reports in the level of detail provided regarding interventions, the relatively short follow-up periods, and the small number of controlled studies of educational supports and related outcomes in this population.

Conclusions

There has been increasing recognition that the disruptive impact that serious psychiatric conditions typically have on developmental tasks, such as the completion of postsecondary education and early work activities, may be a potent contributor to the chronic disability associated with these conditions. Thus, providing services to young people with psychiatric conditions that actively support the completion of these formative tasks may be a powerful strategy to improve long-term outcomes (15, 26, 39, 66).

In accord with this view, recent adaptations of vocational interventions for young people have incorporated SEd and other components with the aim of better addressing their educational and employment support needs. The results reviewed here provide evidence for the effectiveness of IPS-based services to improve employment outcomes in this population, with findings to date suggesting that at least in the short term, such IPS-related effects do not depend on specific adaptations for this age group. Although fewer controlled studies have focused on educational outcomes, their findings suggest that additional service components that enhance standard IPS may be important for efforts to improve postsecondary educational outcomes in this population (23). The elements most clearly suggested by the findings of this review include well-specified, staged educational supports in addition to those for work (14, 57); comprehensive skills training that includes academic, vocational, and problem-solving areas (14, 30, 57, 58); and training staff in MI to assist in addressing fears and ambivalence regarding the pursuit of vocational goals (40). Additional well-controlled studies are needed to evaluate such enhanced approaches, including how they compare with more standard IPS interventions. Such work will help determine how, and to what extent, to adapt the IPS model to best meet the needs of young people (43, 67, 68). In parallel, further research is needed to identify the factors related to variability in vocational treatment response in this population. This will advance our understanding of how to tailor the combination of service components (69) and the overall intensity of services according to individual needs to maximize the effectiveness and efficiency of resources (23).

Finally, a secondary set of observations from this review relates to the lack of standardization across the reviewed studies with regard to the specification of treatment models, description of intervention components, and measurement of potential treatment effects, especially with regard to educational supports and outcomes. This lack of descriptive information, which contrasts with that provided for the clearly delineated model that guides IPS-based SE (20, 31, 70), hinders the process of comparing studies and treatment approaches (35), as well as attempts to replicate initial promising findings.

As discussed by Mueser and Cook (21), specifying a formalized framework of principles and service elements for interventions that incorporate SEd and other career-development components would serve to guide the implementation and evaluation of such approaches (see also Killackey et al. [23]). For example, the recently developed IPS Fidelity Scale for Young Adults (67) includes items related to both employment and educational supports. Thus, it should aid IPS programs in their efforts to design and deliver services adapted for young people in a manner consistent with IPS principles, as well as facilitate the assessment of these efforts. Similarly, the adoption of standardized descriptions of participant samples, including the proportion of participants interested in education, employment, or both by treatment group, along with standard reporting of employment and education outcome measures (23), including the type of employment or educational programs entered into or completed by group, would facilitate the rigorous evaluation of such interventions. These steps are critical for advancing the literature and building the evidence base of effective educational and employment interventions for young people with psychiatric conditions.