Distribution and Correlates of Assertive Community Treatment (ACT) and ACT-Like Programs: Results From the 2015 N-MHSS

All data were collected by the Substance Abuse and Mental Health Services Administration (SAMHSA) between April 30, 2015, and January 19, 2016, as part of the N-MHSS. Data collection methods are described extensively in survey documentation (18). Briefly, the N-MHSS is a voluntary, nationwide survey assessing mental health facility characteristics and offered services. N-MHSS has been described as the “only source of national and state-level data on the mental health service delivery reported by both publicly operated and privately operated specialty mental health facilities” (18). The purpose of the N-MHSS, as stated in the SAMHSA data manual, is to assist governments and other entities in assessing the nature, scope, and distribution of services and to keep an updated national directory of relevant facilities. Participant facilities were identified from a national inventory maintained by SAMHSA and updated annually from state reports. Facilities were screened for eligibility via telephone. Facility supervisors had the option of answering the surveys via a secure Web site, telephone, or mail.

SAMHSA staff provided reminders and technical support and contacted the facilities directly to address response inconsistencies or missing data. To assess whether a facility had ACT, the following question was asked, “Which of the services and practices [listed] are offered at this facility, at this location?” ACT, one of the choices listed, was defined for the respondents as “a multi-disciplinary clinical team approach—helps those with serious mental illness live in the community by providing 24-hour intensive community services in the individual's natural setting” (18). Because of the broad nature of this definition, we could not be certain that each facility that reported having ACT had a program that met the standard definition of this treatment model or whether some programs incorporated only some aspects of the ACT model. A standard ACT program is typically characterized by the presence of core services, high intensity and frequency of contact, a team-based approach, and assertive engagement mechanisms, although only the services aspect could be assessed with the N-MHSS data. Some of the programs that reported offering ACT meet the definition of standard ACT by offering all the core services and meeting other requirements, but others may not and are best described as “ACT-like” programs. This study included all facilities that reported having an ACT program, including programs more accurately described as “ACT-like.”

Department of Defense facilities, jails, and small practices not licensed as clinical facilities were excluded from the survey.

Of the 17,486 known facilities, 16.7% (N=2,913) were found to no longer be eligible or had closed. Of the remaining 14,573 facilities, 87.7% (N=12,826) were included in the study, 8.1% (N=1,185) were nonresponders, and 4.2% (N=562) did not meet survey criteria or did not provide clinical services. Most facilities (80.4%, N=10,315) responded via Internet, another 15.4% (N=1,976) via telephone, and 4.2% (N=535) by mail. Missing values were imputed by SAMHSA investigators by using 2014 N-MHSS data. The data were made publicly available via the SAMHSA Web site. Because N-MHSS is a public data set and does not identify individual patients, no IRB approval was required.

Analyses were conducted in three stages. First, facilities that reported having ACT were compared with facilities without ACT with regard to ownership, funding, licensing, available income subsidies, and treatment focus. Geographic distribution of these facilities was also explored. Availability of ACT was compared with need for ACT programs was further assessed by using the 2015 National Survey of Drug Use and Health (NSDUH) to estimate national and state populations of individuals with serious mental illness. A formula from previous research was used to establish the proportion of individuals with serious mental disorders that would be eligible for ACT at the state and national level based on severity of mental illness and number of hospitalizations in the past year (19).

Second, clinical services offered at facilities that reported offering ACT were assessed. Core services, thought to be critical to ACT programs, as well as secondary services important for individuals with serious mental illness were included. Core services were defined by using the DACTS as well as published studies and guidelines, and they included pharmacotherapy (20–22), individual therapy (20, 23, 24), peer services (20, 25), case management (1, 20, 22), crisis intervention (20, 21), employment services (20, 22–24, 26), housing services (1, 20, 22), and co-occurring disorders services (4, 20) (a model that combines treatment for mental illness and substance abuse from the same clinician and includes counseling specifically designed for patients with co-occurring disorders [18, 20). The secondary services included suicide prevention (27), tobacco treatment (28, 29), diet and exercise counseling (30, 31), integrated primary care (32, 33), educational services (30, 31), and family psychoeducation services (20, 24–26). Telemedicine, although not tested, was also included because it may provide a new venue for bridging primary care and specialty services for patients living in the community.

In the third stage, the characteristics and geographic distribution of facilities that self-reported ACT were compared on the basis of whether they did or did not offer all the core ACT services. All comparisons were conducted by using unadjusted logistic regression models and SPSS software, version 23.

Of the 12,826 participant facilities, 13.4% (N=1,720) reported offering ACT or ACT-like programs at the physical location surveyed (Table 1). There was considerable variability among respondents (including the District of Columbia) in availability of ACT and ACT-like programs (Figure 1A). New Hampshire, South Dakota, and Texas appeared to have the highest proportion of facilities with ACT or ACT-like programs. In most states (N=35, 69%), no more than 15% of licensed facilities offered ACT or ACT-like programs. There were no clear regional trends in availability of these programs (Figure 1, panel A). Based on the 2015 estimated population of ACT-eligible persons, 4,266 ACT programs would be needed nationally to provide optimal coverage, assuming that each team would typically serve 100 clients. Nationally, the 1,720 facilities that provided ACT or ACT-like programs supplied this needed coverage to only 40.7% of ACT-eligible patients, with wide variations across states (Figure 1, panel B). [Estimated capacity for ACT-eligible clients by state is available in an online supplement to this article.]

Compared with facilities without ACT or ACT-like programs, facilities with these programs had higher odds of being publicly owned (OR [odds ratio]=2.18), federally funded (OR=1.53), or funded by a grant (OR=2.03) (Table 1). Facilities with these programs had higher odds of offering low-income subsidies by providing either sliding-scale services (OR=1.48) or pro bono services (OR=1.76). Furthermore, facilities offering ACT or ACT-like programs had higher odds of being hosted by community mental health clinics (CMHCs) (OR=2.44 ) or by U.S. Department of Veterans Affairs (VA) or military facilities (OR=2.55). Facilities with these programs also had higher odds of offering services for co-occurring disorders (OR=1.87) (Table 1).

Just over 19% of facilities with ACT or ACT-like programs reported offering all of the core services thought to be crucial to the ACT model (Table 2). A majority (N=1,392, 80.8%) of programs that reported providing ACT (10.8% of facilities included in the analyses) did not offer all of the core services. There was substantial variability among facilities in individual core services, with the percentage of facilities offering various services ranging from 44.0% for peer services to 94.9% for individual therapy (Table 2). The proportion of facilities offering all the core services varied widely from state to state (Figure 1, panel C). Facilities with ACT or ACT-like programs had higher odds of offering secondary services, although availability of these services varied considerably among programs (Table 2).

There were significant differences in characteristics and availability of secondary services between facilities that offered all core ACT services and those that did not (Table 3). Facilities offering all the core services had higher odds of being publicly owned or operated (OR=2.27) than facilities with fewer ACT or ACT-like services. Facilities offering all the core services also had higher odds of being funded by federal (OR=3.33) and grant funds (OR=1.67). Overall, there was a trend for facilities that offered all the core ACT services to have higher odds of funding across all sources other than Medicaid (Table 3).

Facilities with ACT or ACT-like programs that offered all the core ACT services had higher odds of providing sliding scale (OR=1.98) and pro bono services (OR=2.05) and had higher odds of being a CMHC (OR=1.34) or a VA or military facility (OR=4.47) (Table 3). Finally, facilities with ACT or ACT-like programs that offered all the core ACT services had higher odds of offering secondary services (Table 3).

As a clinically effective, evidence-based practice model for individuals with serious mental illness (1, 2, 4), ACT has gained national and international recognition (1, 4). However, this study’s results show that availability of ACT or ACT-like programs varies widely from state to state. Additional analyses with 2015 NSDUH data highlighted this further, suggesting that only 40.7% of individuals eligible for ACT nationwide could receive such services, based on the self-reported number of facilities with ACT, with significant variability among states.

Compared with facilities without ACT, facilities with ACT have higher odds of being owned and operated by public agencies, perhaps suggesting that in certain states, the legislative and administrative environments are more favorable to ACT programs by directing public agencies to promote the dissemination of these programs. Facilities with ACT also had higher odds of receiving federal funds, including military and VA funding, and higher odds of being funded by grants. Funding sources have long been thought to influence the distribution of certain types of services, including ACT (34). The VA system has funded a national network of programs in an effort to disseminate ACT services to veterans (35). It is also possible that the increase in funding stability and consistency that comes with grants might promote dissemination of ACT, although programs that rely exclusively on grant funding might not be sustainable in the long term. This possibility underscores the importance of increased state funding, for example through Medicaid. Relying on Medicaid funding has traditionally been a challenge for ACT in most states, given that many of the services provided by ACT are not covered by Medicaid (36). Some states have specifically revised their regulations to better align Medicaid reimbursement with ACT services, but they remain in the minority (36). Overall, the wide variability in ACT availability across states suggests that in parts of the country, significant challenges to wider ACT adoption and availability remain.

Perhaps availability is as important as fidelity to the ACT model (5, 7, 8). Using parameters that approximate DACTS, this study found wide variability in the availability of required services for ACT in facilities that offered it. Across the United States, only 19.2% of facilities with ACT (or 2.6% of all surveyed facilities) reported offering all core ACT services, with significant between-state variability, whereas 80.8% of programs that reported providing ACT (or 10.8% of all surveyed facilities) did not. Notably, about one in seven facilities with ACT did not offer pharmacotherapy, despite ample evidence supporting its use (21, 22, 37), and one in six did not offer any type of case management.

The outcomes of programs not offering all core services may be different from high-fidelity ACT programs, although there is some evidence that ACT teams may be effective even without full ACT staffing (15, 38). Nevertheless, the large discrepancy between the facility administrator reports of offering ACT services and the availability of core services in these facilities is concerning.

Patients with severe mental illness often have unique and complex mental and physical health needs (39–42), and they may benefit from additional services not captured by commonly used fidelity models. Facilities with ACT are uniquely positioned to address such needs, and as a group they had higher odds of offering important secondary services.

A number of significant differences were found among facilities that reported having an ACT or ACT-like program, depending on whether the program offered all core services or just some. Facilities offering all the core ACT services appeared to have higher odds of being publicly owned and operated, suggesting the potential role of state-level legislative, regulatory, and administrative environments. Certain states might be particularly adept at providing training and administrative oversight to ACT programs. Additionally, some states have established centers of excellence via federal block grants that can foster higher-fidelity ACT.

Facilities with ACT offering all the core ACT services also had over three times the odds of receiving funding from federal sources, including military and VA funds. The military and VA have historically funded and promoted specialized professional training to help attract staff to their clinical facilities (43). Federal funding of clinical services has at times been used to incentivize certain evidence-based practices (44, 45). Additionally, federal funds may have unique spending, oversight, or auditing requirements that might promote higher quality of care. For example, Medicare funding requires certain parameters in care provision and documentation (46, 47). Finally, federal funds have been used in the past to directly promote certain services, such as rehabilitative services (48), supported employment (48), and telemedicine (49).

Facilities with ACT or ACT-like programs that reported offering all of the core ACT services also had higher odds of being funded by grants. Certain types of grant funding could promote higher quality of care directly (44–47, 49). Additionally, grants might promote focus on quality by allowing programs to shift their focus away from billing and budgetary considerations, although only when the grants are active.

The odds of providing important secondary services were higher at facilities that provided all core ACT services compared with programs that reported having an ACT program but did not provide all core services. Notably, telemedicine was offered at just under two-thirds of facilities that offered all the core services.

The study’s findings should be considered in the context of its limitations. First, the unit of analysis was facility, not program. Some facilities may have housed other programs along with ACT. Some of the services offered in facilities with ACT may have been part of other programs in these facilities. Such a misclassification error would result in overestimation of ACT fidelity. The true fidelity of ACT programs might indeed be lower than reported here. Similarly, secondary services reportedly offered at facilities with ACT may have been part of other programs in those facilities. Fewer services may have been attached to the ACT programs than reported here.

Second, the self-report nature of the survey means that the number of facilities with ACT was only a rough estimate. Self-report surveys are prone to social desirability bias. Additionally, some facilities may have housed more than one ACT program, although a facility was unlikely to house more than one ACT program at the same physical address. It is also notable that the number of sites reporting ACT in each state was roughly similar across the 2010 to 2015 N-MHSS survey years (data not shown).

Third, because individual DACTS scores were not available for comparison, ACT fidelity could be approximated only by examining the services offered at each facility. However, the services used to define higher-quality ACT were derived from DACTS and relevant literature. Nevertheless, the DACTS likely does not capture all the services relevant to individuals with severe mental disorders, and implementation of newer fidelity assessments, such as the Tool for Measurement of Assertive Community Treatment (TMACT), is increasing (50).

Finally, unmeasured parameters, such as being part of an academic center or the size of the programs, which may independently affect availability of services, could potentially explain some of the differences in service offerings.

Although this study used a very broad definition of ACT, it is remarkable that few facilities across the country offered these programs and even fewer offered all the core services required by the ACT model. Furthermore, there was marked variability across states with regard to availability of programs, unmet need for these programs, and availability of core ACT services. The findings call for greater attention to the availability and quality of ACT. State legislators and regulators can contribute to improving the quality of programs through incentivizing and monitoring adherence to fidelity measures at the state level and modifying their Medicaid coverage to promote ACT, particularly higher-quality ACT. States might also take advantage of available block grants to establish and support centers of excellence in order to foster higher-fidelity ACT. Individual program administrators and ACT clinicians could contribute to improved quality of care for the vulnerable patient population potentially served by these programs by lobbying for reimbursement of services that would meet benchmarks of quality.