Introduction

Anxiety disordersFootnote 1 are among the most prevalent mental disorders, and are chronic, costly, and disabling (Kessler et al. 2005; Stein et al. 2015; Greenberg et al. 1999). Indeed, anxiety disorders are associated with poor quality of life and functional impairment (Olatunji et al. 2007), disability and impairment comparable to major depressive disorder (Wittchen 2002), and poorer functioning when comorbid with other psychiatric conditions (Mittal et al. 2006; Braga et al. 2005). Behavioral and cognitive-behavioral therapies (CBT) that involve gradual confrontation with feared stimuli (known as exposure) represent the most established and efficacious evidence-based psychosocial treatments for this group of disorders, showing large effects in both efficacy and effectiveness studies (Deacon and Abramowitz 2004; Roy-Byrne et al. 2010; Stewart and Chambless 2009) that are comparable to medication (e.g., SSRIs) in the short-term and are larger in the long-term (Hofmann et al. 2009). Exposure-based CBT also outperforms other forms of psychotherapy (Tolin 2010). These treatments are arguably the most effective existing treatments for any psychiatric disorder. As a consequence of the overwhelming evidence, CBT is considered the treatment of choice for anxiety disorders (Deacon and Abramowitz 2004).

Despite the overwhelming evidence for its effectiveness, access to and receipt of evidence-based treatments for mental health disorders, and exposure-based treatments for anxiety disorders in particular, remain shockingly low (Insel 2009; Santucci et al. 2012; Weissman et al. 2006; Hipol and Deacon 2013). Indeed, despite the existence of highly effective treatments for anxiety disorders, most individuals with anxiety disorders do not receive them (Young et al. 2001; Wolitzky-Taylor et al. 2015; Santucci et al. 2012). Given the significant public health burden of anxiety disorders, increasing accessibility to evidence-based psychosocial treatment for these disorders is urgently needed.

Little attention has been paid to the dissemination and implementation of evidence-based psychosocial treatment of anxiety disorders in U.S. adult populations, particularly in community mental health settings that serve low-income individuals. Some studies have made efforts to increase the accessibility of CBT for anxiety disorders by demonstrating that it is effective in real-world clinical settings when delivered by professionals with limited mental health training (Roy-Byrne et al. 2010) and via web-based modalities (Andrews and Titov 2010), and there has been a large body of research aiming to disseminate and implement motivational interviewing for adults with substance use disorders. However, studies are lacking that investigate strategies for successfully disseminating CBT for anxiety to community providers and implementating CBT for anxiety disorders in publicly funded community mental health settings. Indeed, large-scale work aimed at increasing accessibility and dissemination of CBT specifically for anxiety disorders to U.S. mental health settings has been largely limited to the treatment of children (Weisz et al. 2012) and military Veterans (McLean and Foa 2013). Although this important work has helped the lives of many individuals in vulnerable populations, the majority of individuals in the United States with anxiety disorders are non-Veteran adults, thus leaving a sizeable proportion of individuals under- or inadequately-treated for their anxiety disorders.

Despite the lack of large-scale dissemination studies of CBT for adult anxiety disorders, informal dissemination efforts could plausibly lead to CBT for anxiety disorders permeating usual care practices in community-based mental health clinics. However, our work examining the usual care practices at a large outpatient community mental health clinic that serves thousands of low-income, predominantly minority adult patients observed that nearly all patients with anxiety disorders received pharmacological treatment and supportive therapy for their anxiety disorder. A small minority received CBT but only a proportion of those (just over one-quarter) received any exposure therapy, arguably the most active component of CBT for anxiety disorders (Wolitzky-Taylor et al. 2015). This was the case even though many clinic staff had access to exposure-based CBT training and supervision, rendering this a better-case scenario. This study was the first to investigate usual care practices for anxiety disorders in a community mental health clinic in the United States. Although it remains possible that high quality exposure-based CBT is practiced in some community mental health clinics, there is no evidence to suggest that this occurs generally or that systematic efforts have been made throughout community mental health “safety-net” settings to support the training and delivery of CBT for anxiety disorders.

The reasons that predominantly low-income patients in community mental health settings generally fail to be offered and receive a highly effective and cost-effective treatment remain unclear. In line with conceptual models of implementation science (Aarons et al. 2011; Damschroder et al. 2009), and in particular the exploration and preparation phase of the Exploration, Preparation, Implementation, Sustainment (EPIS) model, we outline putative barriers deserving of consideration at multiple system levels (i.e., system, organizational, provider, and patient), both in the outer context (e.g., funding) and inner context (e.g., provider attitudes), that may impede the delivery of exposure-based CBT. By carefully considering the early stages of implementation, we position ourselves to develop strategies that increase the likelihood of adoption of exposure-based CBT into low-income community mental health settings. Although other putative barriers may emerge during the initial stages of implementation, here we focus on those that have received some empirical support. Note that although many common implementation science models do not characterize patients as adopters, we deliberately include this key stakeholder group in line with our patient-centered and community-based participatory research approach to include all stakeholders, including service seekers, as key participators in the implementation process. Indeed, without considering patient acceptability and needs, an otherwise well-adapted intervention ready for implementation is not likely to yield sustainable effects in clinical practice.

Patient-Level Factors

Awareness of Exposure-Based CBT

Exposure-based treatment represents the most scientifically supported psychosocial treatment for anxiety disorders, yet the majority of U.S. adults do not know it exists (Gallo et al. 2013; Arch et al. 2015). Indeed, in our work examining usual care practices at a large community mental health clinic, almost no patients requested exposure therapy or CBT; yet nearly all who were offered accepted it (Wolitzky-Taylor et al. 2015). Adults with anxiety disorders cannot make empowered health care decisions when they lack access to exposure-based treatment and knowledge of evidence-based psychosocial treatment approaches. We aim to explore whether community mental health patients have heard of CBT and exposure, and what they know (or do not know) about it in more detail. If lack of awareness is identified as a barrier to the receipt of CBT, these findings may inform the development of educational tools for patients and “direct to consumer” marketing about effective behavioral treatment approaches so that patients can advocate for their treatment preferences when visiting their providers (Santucci et al. 2012; Gallo et al. 2013).

Preference and Acceptability

Conventional wisdom might suggest that patients will not find deliberate exposure to feared stimuli acceptable, but research suggests that when given the choice, most adults prefer CBT over medication treatment for their anxiety disorders (Deacon and Abramowitz 2005; Feeny et al. 2009; Arch et al. 2015). Generally, low-income, predominantly minority patients report a preference for psychological treatment over medication (Dwight-Johnson et al. 2010). Also, we found that among the small minority who were offered CBT at a large community mental health clinic, nearly all accepted this treatment plan and began exposure therapy (Wolitzky-Taylor et al. 2015). These findings support the notion that patients will find CBT that includes exposure acceptable, and that patient preferences appear to be unmet in our community clinics. However, no research to date has specifically gathered feedback from low-income, predominantly minority adults with anxiety disorders to more thoroughly understand their preferences for treatment (e.g., CBT vs. other alternatives, exposure vs. traditional cognitive therapy; face-to-face sessions vs. Web-based treatment). We aim to explore these preferences, which may also elucidate whether we can preliminarily identify certain baseline characteristics that may be associated with more or less acceptability for exposure-based CBT. These findings may inform more formal moderator analyses to aid in the goal of developing personalized mental health care in community settings.

Clinician-Level Factors

Training

It is well-documented that most clinicians are not trained in evidence-based psychotherapies (Weissman et al. 2006; Berzin and O’Connor 2010; Pidano and Whotcomb 2012; Karekla et al. 2004), especially in CBT (Hipol and Deacon 2013). However, detailed information about training among adult mental health practitioners is lacking. Although insufficient training represents an organizational barrier, it may also be considered an individual adopter barrier because there are likely individual differences in the extent to which clinicians obtain and seek out evidence-based training and supervision. Significant research is needed about training and supervision practices that clinicians find most beneficial, specific knowledge gaps that require additional training or more experiential, barriers to ongoing training and supervision for those who do report having some CBT training (e.g., role playing specific skills with an expert trainer), and modalities of CBT (and its training) that are most acceptable to clinicians (e.g., flexible, modular approaches vs. manualized approaches; Weisz et al. 2012; Borntrager et al. 2009).

Beliefs and Knowledge About CBT and Clinical Issues

Additional potential barriers that have emerged from discussions with community providers include (a) negative beliefs about CBT and exposure that are not supported by empirical evidence, including beliefs that exposure-based CBT is harmful or only suitable for specific types of patients (e.g., simple, no comorbidity, no psychosocial barriers, no substance use); and (b) beliefs that anxiety disorders are not central to their patients’ primary psychopathology and thus are not important to treat. Beliefs such as these have been identified in prior studies investigating community therapists’ beliefs about exposure therapy for anxiety disorders (Deacon et al. 2013; Meyer et al. 2014).

Administrative-Level Factors

There is a paucity of data specific to anxiety disorder treatment in community mental health settings to shed light on administrative/system level barriers to the delivery of CBT. Similar to clinician-level barriers, we speculate that beliefs among clinic administrators and organizational decision-makers about short-term vs. long-term cost-effectiveness and prioritization of an individual’s psychotic or mood disorder over the same individual’s anxiety disorder are possible barriers in governmentally-funded, safety-net settings (e.g., Los Angeles County Department of Mental Health; DMH). That is, identifying and effectively treating anxiety disorders may be a low priority relative to mood and psychotic disorders due to administrative beliefs that anxiety disorders have lower impact and severity.

Organizational Characteristics

A set of common organizational barriers to delivering evidence-based treatments in community settings with children, including insufficient time, training, supervision, and support for providers to adequately learn and deliver evidence-based treatment with high fidelity (Yoo et al. 2007) are likely to be common to adult community mental health settings. Broad ranges of training and competency in delivering CBT (including exposure), coupled with insufficient time for supervision to focus on mastery of CBT, may emerge as organizational barriers to delivering these interventions.

System-Level Characteristics

There is a lack of data specific to anxiety disorder treatment in community mental health settings to shed light on organizational- and system-level barriers to the delivery of CBT. In the outer context, we speculate that allocation of resources to training in evidence-based practices, funding mechanisms through the California Mental Health Services Act, and policies that may change the structure of the payer systems (e.g., moving to a capitated system of reimbursement of services rather than fee-for-service), all represent possible barriers at the systems level within governmentally-funded, safety-net settings.

Summary and the Current Study

Multiple patient-, provider- and system-level factors may contribute to the dramatic lack of uptake of evidence-based treatment for anxiety disorders in community mental health settings; however, no studies to date have attempted to discern which barriers are actually endorsed by leaders and clinicians in U.S. community mental health settings, e.g., which barriers objectively exist. The current study aims to address this striking evidence gap. Despite dissemination and implementation efforts made in community mental health settings for children (Weisz et al. 2012; Cohen and Mannarino 2008; Southam-Gerow et al. 2014), the vast majority of adults in community mental health settings do not have access to exposure-based CBT for anxiety disorders. This lack of access is particularly striking when considering that: (1) effective psychosocial treatments are more cost-effective than pharmacological interventions (Roberge et al. 2004) and the benefits are more enduring (Hofmann et al. 2009); and (2) across studies in which adults are provided education and choice about evidence-based treatments for anxiety, the majority prefer exposure-based psychosocial treatment over medication treatment of anxiety (Deacon and Abramowitz 2005; Feeny et al. 2009). Lacking knowledge of exposure-based treatment (Arch et al. 2015), adults in our communities with anxiety disorders cannot make informed decisions about their mental health care. They then risk investing time and resources on less effective or ineffective treatments (Lilienfeld et al. 2003). Given that most adults served in community mental health settings are low-income, they especially cannot afford to invest their limited resources in sub-optimal treatment.

The current study, supported in part by initiatives from the California Mental Health Services Act, sought to engage with multiple stakeholders in the community mental health system in Los Angeles County in order to gather input from each stakeholder type to better understand the barriers to delivering (or receiving) CBT. In a series of key informant interviews and focus groups with patients, providers, clinic administrators, and policy experts, this preliminary step was meant to inform a larger effort to more broadly assess barriers to delivering and receiving CBT for anxiety disorders in adult community mental health settings. The longer-term goal of this crucial exploratory and developmental work is to use the information gathered from key stakeholders to guide the development of recommendations, strategies, and materials for improving the dissemination and implementation of CBT for anxiety in low-income, safety-net settings. Because so little is known about the barriers to receiving and delivering evidence-based treatment, and in particular, exposure-based CBT in adult community mental health settings, qualitative research provides a foundation for beginning to understand the barriers. This formative work lends itself well to gathering detailed information that can ensure that future research is understanding the problem well enough to ask the right questions. Specifically, information from qualitative interviews and focus groups can be used to develop quantitative instruments that appropriately assess these putative barriers on a larger scale.

Method

Participants

Four clinics participated in the qualitative portion of this study. The four clinicsFootnote 2 all serve low-income, “safety net” patients on California’s state Medicaid system (i.e., Medi-Cal). These four clinics were chosen in part because they represent four major, distinct urban areas of the most highly-populated county in the United States, Los Angeles County (US Census 2010): West Los Angeles, South Los Angeles, East Los Angeles, and North Los Angeles, respectively. Participants included one policy expert in community mental health (i.e., the former Director of the Los Angeles County Department of Mental Health), five clinic administrators/directors (two from one site, and one from each of the other three sites), 30 providers across the four clinic sites, and 10 patients across two of the four clinic sites.

Table 1 Demographics in focus groups by stakeholder type and clinic
Full size table

Patient participants were primarily female (70%). 40% were African-American, 10% Hispanic/Latino, 30% multiracial, and 20% Caucasian. 30% had full- or part-time employment, with the remaining participants unemployed. Provider participants across the four clinics were primarily female (80%), had an average age of 35.14 (SD = 12.19), and were ethnically/racially diverse, with 37.9% Caucasian, 17.2% Hispanic/Latino, 10.3% African-American, 10.3% Asian-American, 20.7% multiracial, and 3.4% other race. Providers had an average of 5.28 (SD = 10.29) years of experience. 43.3% of providers had an MSW, 26.7% had an MFT, 26.7% had an MD (with a specialty in psychiatry), and 3.3% had a PsyD. Administrator participants (who typically served as clinical supervisors with administrative duties) were primarily female (66.6%), with a mean age of 54.33 (SD = 10.13). Administrators had an average of 24.83 (SD = 10.08) years of experience in their field. Administrators included two MSWs, two PhDs (in clinical psychology), one DSW, and one MD (psychiatrist). Table 1 presents the demographic data distinguished by clinic for the patient and provider focus groups.

Measures

Separate sets of standardized interview and focus group questions were developed for the (a) administrator and policy-level interviews; (b) provider focus groups; and (c) patient focus groups. Questions covered (a) beliefs about anxiety disorders and the importance of treating anxiety disorders, (b) anxiety disorder treatment practices, (c) knowledge, beliefs, and preferences about different treatment options for anxiety disorders, with a focus on cognitive and behavioral treatment components, and (d) perceived barriers to delivering (or receiving) CBT for anxiety disorders. Each topic area began with open-ended questions, with additional prompts and follow-up questions presented as the discussion developed. Questions were geared toward the stakeholder type being questioned. For example, with regard to insufficient training and competency in CBT as a potential barrier, patients were asked whether their therapists delivered certain skills (e.g., “helping you face avoided situations that make you anxious”), providers were asked whether they believed they had adequate training and supervision in specific techniques, and clinic directors and administrators were asked about CBT training and supervision practices in the clinic and whether they had sufficient resources to train in CBT for anxiety disorders. The full scripts of interview and focus group questions are available upon request.

Procedures

The study was approved by the university’s Institutional Review Board and the Los Angeles County Department of Mental Health Research Committee.

Recruitment Procedures

The first author (KWT) had initial meetings with administrators to discuss the project and conducted interviews at that time. Administrators then made verbal and email announcements to their clinical staff about the study and worked with the research team to identify potential provider participants. After the provider focus groups at two of the clinics, the first author discussed with providers ways in which they could help to facilitate patient recruitment and let patients know about the study. Therefore, recruitment was done sequentially and in collaboration with community stakeholders. Flyers were also posted in clinic waiting rooms for the patient focus groups.

Data Collection Procedures

After providing informed consent and basic demographic information, stakeholders participated in their respective interviews and focus groups, which were led by the first author (KWT). Specifically, one provider focus group was conducted at each of the four clinics and one patient focus group was conducted at two of the four clinics. Focus groups were conducted in English and at the clinic sites. Focus group participants signed a confidentiality agreement. All interviews and focus groups were audio recorded and transcribed. Interviews lasted approximately 45 min and focus groups lasted approximately 1 h. Refreshments were provided. Gift card compensation was provided to the patient participants.

Data Analysis Procedures

Transcripts were analyzed using a grounded theory approach, with the project’s specific aims serving as sensitizing concepts (Padgett 2012; Charmaz 2006). Coding occurred in two stages. First, two doctoral student level researchers (fourth and fifth authors) independently analyzed each transcript. These researchers specialized in qualitative methods and were not involved in the project’s data collection process. These authors created an initial codebook of emerging, distinguishable themes and then debriefed with the first author to refine and verify the codes. Second, using NVivo 11, they independently analyzed all transcripts based on immersion in the data (Patton 2015). Coding decisions were discussed and discrepancies were resolved through ongoing discussions until consensus was reached. In addition to line-by-line coding, meaningful and descriptive passages were also extracted to illustrate the emergent themes. In order to reduce bais in interpretation of the data, the first author had minimal contact and virtually no influence on the analysis itself. Specifically, the first author had one brief phone conversation with the data analysts to explain the purpose of the study and the barriers that were being explored in the interviews and focus groups (but not hypotheses). The first author had no contact with the analysts during the data analysis portion of the study, and had one additional discussion after they had completed their results to discuss how they suggested the results be interpreted.

Methodological rigor was enhanced using several strategies. First, the qualitative researchers and first author regularly debriefed to discuss theme development and coding decisions (Padgett 2012). Second, to enhance reliability, the qualitative researchers independently co-coded 100% of the transcripts and discussed discrepancies in the co-coded documents until consensus was reached (Padgett 2012). To assess inter-rater reliability, kappa coefficients were calculated using NVivo 11. Across the codes, kappa coefficients ranged from .66 to .93, indicating “substantial” to “almost perfect” agreement, and the overall kappa was .79, indicating “substantial” agreement (Viera and Garrett 2005). Finally, the qualitative researchers used memos in NVivo to create a detailed audit trail that describes coding decisions and theme construction (Charmaz 2006; Padgett 2012).

Results

Table 2 lists the themes and subthemes dervied from the raw codes that emerged during the analysis, and illustrates the proportion of sources that contained relevant text for each subtheme. As shown in Table 2, knowledge about anxiety and its treatments, barriers to accessing or offering evidence-based treatments for anxiety, and facilitators and solutions for accessing or offering these treatments were discussed in the interviews and focus groups. Below, we describe and synthesize key barriers to using and accessing evidence-based anxiety treatment in more detail, as that was the primary focus of our investigation.

Table 2 Description of raw themes and subthemes
Full size table

Patient-Level Barriers

Provider and Clinic Administrator Perspectives

Key patient-level barriers that emerged during the provider and clinical administrator interviews and focus groups were (1) competing treatment priorities, (2) characteristics of the client, (3) ability to grasp CBT, (4) logistical issues, and (5) resistance to CBT. Both providers and administrators frequently cited competing priorities for treatment as a barrier. Competing priorities included psychosocial stressors (e.g., unemployment), more “severe” diagnoses (e.g., psychotic or bipolar disorders), and homelessness. As one provider stated, “…yes, maybe 90% of our clients have anxiety, but the homelessness issue, the depression, the suicidality- that’s what we need to focus on.” Anxiety kind of gets put off till the end and at that point.” Existing patient characteristics, including medical or substance comorbidities, cognitive impairment, and cultural or language differences, could also hinder the delivery of evidence-based anxiety treatment. For example, one provider described how culture affected client receptivity to CBT: “There’s that cultural barrier also… they’re just not used to talking to anyone about their problems. They want to come here and tell you, ‘I have anxiety can you just fix it so I can go.’”

A third patient-level barrier reported by providers and administrators was difficulty grasping CBT-related concepts. This may be an issue, for example, when patients are not able to explain their thoughts or core beliefs. One provider described,“I had one patient you could try to sit there and explain things and she would never understand the basic concepts.” Additionally, providers identified patient logistical issues, such as lack of childcare, limited finances, and transportation difficulties, as directly interfering with patient participation in treatment. For example, one provider mentioned, “It’s just not feasible for them to be coming back every week. They only have enough money for the bus one time a month… CBT should be weekly or biweekly, but it’s just not.” Finally, providers described how client resistance or lack of buy-in to anxiety treatment or CBT could prevent providers from offering and clients from benefiting from treatment. One provider recalled, “I had such trouble trying to sell that [CBT] to the client… I was like ‘OK, I guess I’ll switch tactics.’”

Patient Perspectives

Some commonly mentioned patient experiences with regard to patient-level barriers aligned with provider and administrator perspectives. For example, similar to providers and administrators, two themes patients identified as barriers to receiving CBT were (1) competing treatment priorities, including other psychiatric disorders and (2) logistical issues. With regard to comorbidity and competing treatment priorities, one patient explained, “With anxiety you’ve got a lot of things going on. It’s not just the anxiety, it’s a whole lot.” Describing one of the logistical constraints as a barrier to receiving regular CBT sessions, another patient shared, “It was hard, because I was taking the bus, three buses to get here.”

Another area of convergence between patient perspectives and provider/administrator perspectives related to patients’ ability to understand and articulate their experiences with anxiety and subsequent treatment. While providers discussed patients’ difficulty grasping CBT-related concepts, patients expressed a lack of awareness and knowledge about their symptoms that may have hindered their ability to grasp CBT concepts. As one patient explained, “I didn’t realize I was being treated for anxiety, and I didn’t realize some of the things we were doing may have fallen into CBT type programs.”

Provider Level Barriers

Provider and Administrator Perspectives

From the perspectives of the providers and administrators, barriers included intra-provider characteristics, which are related to professional work culture but are described separately as an organizational barrier below. With regard to intra-provider characteristics, some administrators identified provider competence, both with CBT and with general therapeutic skills, as a barrier. For example, one administrator stated, “I think some of the newer clinicians might say ‘oh, this is depression’ and won’t have the experience to say ‘this is an anxiety disorder.’” Additionally, providers reported that personal discomfort with using anxiety treatments and CBT influenced how likely they were to offer these treatments to patients. As one provider explained,

I feel like “oh maybe this would be appropriate for this person, because that’s a panic attack and this is one of the interventions and I just learned about it.” But, I think that probably was my level of comfort and maybe not knowing the right approach to help the client buy into it, and so they were just very uncomfortable and didn’t even want to try it.

Organizational Level Barriers

Provider Perspectives

Providers and administrators identified barriers related to (1) training, (2) supervision, and (3) time constraints. Providers described barriers in terms of the quantity and quality of training and supervision. With regard to training, one provider stated:

I did get some prolonged exposure therapy training- it was brief. I think trainings that had been successful are ones where you get consistency provision and more course materials and it was just kind of like a quick weekend training but no ongoing supervision. It’s been a while since I’ve taken it so I don’t feel comfortable in using it.

With regard to supervision, one provider stated: “I definitely feel like I need more ongoing supervision, I think I got the basics down from the coursework, but definitely ongoing supervision and maybe a consultation…Yeah it’s different than like role playing in class.”

Providers also identified time constraints as an important organizational-level barrier to delivering anxiety treatment. Providers’ time was limited by other professional commitments, scheduling difficulties, and the limited time allotted to therapy sessions. For example, one provider stated, “We have certain time frames [for completion in a program], so once they meet that, we have to move them on to the next program.”

Administrator Perspectives

Three primary themes emerged in the administrator interviews with regard to organizational-level barriers: (1) insufficient ability to support training in CBT; (2) uneven CBT competence across providers within an organization; and (3) time constraints. While providers discussed barriers related to receiving training and supervision, administrators discussed these barriers in terms of the organization’s ability to provide and support training. Administrators described the limited funds available for training and the ways in which training could interfere with the day-to-day operations of the organization. One administrator stated.

I have a certain amount of CBT dollars. Training in this particular CBT program is a lot of work. I wouldn’t put two clinicians [in the training], so it’s definitely taking away…I would love to offer it to everyone but I imagine that my bosses wouldn’t love that. I mean, I would love to take the training.

This particular statement also highlights a point that was made in two of the four clinics. Specifically, both providers and administrators noted instances in which the supervisors did not get CBT training, so the clinicians who received CBT training were supervised by those with no CBT training themselves.

Another barrier that hinders an organization’s ability to provide high quality anxiety treatment, reported by providers and administrators, is uneven CBT competence across different staff members. As one administrator explained, “…the social workers, it varies. So right now I think we have one who’s pretty confident in delivering CBT that we would be speaking the same language. The other one I think is more eclectic.”

Also, like providers, administrators also identified time constraints for their providers as a barrier to delivering CBT for anxiety disorders. For example, describing the difficulty with offering CBT to large numbers of patients, one administrator noted, “One of the challenges is being able to work with their schedules because they also have to see a number of return patients.”

Patient Perspectives

One key area of divergence across the perspectives was turnover. While high provider turnover was not identified as a barrier by providers and administrators, this issue was emphasized by patients and was the primary theme that emerged in the patient focus groups with regard to organizational barriers. One patient described how turnover inhibited the therapeutic process: “Every three months, I’ve had a different therapist. I don’t like that. How could you get to know me if every 2-3 months I have to go to someone different.”

Service System Level Barriers

Administrator and Provider Perspectives

Features of the service system can also interfere with the delivery of evidence-based anxiety treatment. These barriers included (1) level-of-care issues, (2) funding and reimbursement constraints, and (3) referral issues. With regard to level of care and service eligibility, administrators described the issue with the perception that anxiety disorders may not be sufficiently impairing to access public services. For example, one administrator stated,

Typically someone suffering from anxiety disorders at an early stage wouldn’t think of going to the public system. The second barrier is if they did think of it, they might be turned away because they don’t meet Medi-Cal necessity because their level of life impairment isn’t significant enough yet.

With regard to funding streams, administrators and providers frequently described how policies preventing organizations from serving clients with higher functioning or prior treatment histories often blocked clients with anxiety from accessing CBT treatment. They also explained that payment and reimbursement restrictions could preclude delivering individual therapy generally, and CBT specifically. One administrator stated,

I can think of something which has nothing to do with CBT or this clinic at all but just the general funding as I see it. There’s providing funding for clients who have never been in therapy and clients with quite high needs of care and there’s quite a gap in the middle. Those people often don’t have access to service here, I don’t know what their other options are but that is a barrier to being able to receive anxiety treatment, CBT, or any treatment at all.

Finally, both providers and administrators mentioned barriers related to successfully referring patients to other community-based providers for CBT. These referral barriers included lack of confidence in community providers to maintain treatment standards, limited CBT availability, and lack of information about community practitioners offering CBT. For example, one provider commented, “I didn’t feel confident I was doing great CBT but I felt like I was doing more than they would get anywhere else, most places anyways. So I didn’t feel bad about doing bad CBT.”

Another system-level barrier worth noting was the idiosyncrasy with which evidence-based practices were put on an authorized DMH “list” and then the lack of standardization across clinics in terms of choosing which evidence-based practices to train providers to use from the list. The policy expert administrator described this process, noting,

Then in some programs where you have to use evidence-based practices like the Department of Mental Health, they may be incorporated but there really is not a specific EBP for anxiety disorders I believe that the Department of Mental Health sanctions it as a standalone evidence-based practice. Partially because I believe they limit the number to the ones they feel impact the largest number of people in the service area. Each service area selects around 5 evidence-based practices. While anxiety may have been on the original list, it may not have made it to the top 5, so it’s not one that’s offered.

Discussion

This study provides an important first step toward understanding the barriers to delivering and receiving evidence-based treatment for adult anxiety disorders in U.S. community mental health settings. Our key informant interviews and focus groups with a diverse group of patients, providers, and administrators at four community mental health clinics representing each major demographic area of Los Angeles County revealed some promising findings, including recognition of the problem of anxiety disorders and a desire to improve the uptake of exposure-based CBT in these clinics. However, several consistent barriers emerged within and across stakeholder groups at a variety of levels of the system.

At the patient level, consistently reported barriers included complex clinical presentations with multiple comorbidies (e.g., more severe diagnoses) and psychosocial/environmental stressors (e.g., homelessness), logistical issues making session attendance difficult, and difficulty understanding therapeutic components of CBT. Despite patients describing their anxiety as disabling and distressing, providers and administrators noted that their typical patients have complex psychiatric, medical, and psychosocial issues, often making anxiety a low priority issue to treat. Consistent with prior research on other clinical populations (Alegría et al. 2008; Scheppers et al. 2006), the various stakeholder groups consistently raised environmental barriers such as transportation, childcare, and finances when describing difficulties patients would have in attending regular sessions for a course of CBT for anxiety disorder(s). Finally, providers and patients both described how some patients had difficulty understanding CBT concepts, particularly those with severe psychiatric conditions (e.g., psychotic disorders).

These preliminary patient-level findings suggest that patients may benefit from their providers conducting a fuller diagnostic assessment that includes an assessment of which problems/diagnoses cause significant distress and impairment. This would facilitate developing treatment plans that do not overlook potentially distressing anxiety disorders that may not be as evident as a severe mental illness, substance use disorder, or major psychosocial stressor such as homelessness. In addition, additional training may be needed to help providers clearly communicate core CBT concepts to patients with deficits in cognitive functioning (e.g., clear ways to explain the rationale for exposure, focusing on the concrete behavioral component of CBT). Finally, using modular or transdiagnostic, principle-based protocols that are flexible and not tied to specific anxiety disorders, but rather processes (e.g., behavioral avoidance) and components that target them (e.g., exposure) may give providers the tools they need to flexibly use strategies to treat anxiety without the need for specific differential diagnostic information (e.g., Chorpita et al. 2004; Farchione et al. 2012). Finalizing or implementing these recommendations, however, requires additional research using quantitative instruments to determine the frequency and scope of these barriers in larger samples (see Wolitzky-Taylor et al. 2018).

At the provider level, the main emerging barrier was intra-provider characteristics related to competence and comfort with delivering CBT for anxiety disorders. Interestingly, we found that providers generally had positive beliefs about evidence-based practices and about CBT. In contrast to previous work (Olatunji et al. 2009; Becker et al. 2004; Deacon and Farrell 2013), negative attitudes about exposure-based CBT for anxiety did not emerge as a major barrier. In fact, most providers expressed interest in receiving training in exposure-based CBT, in further developing CBT skills they may have learned in a brief training, and receiving ongoing CBT supervision. To speculate, this divergence from previous findings could stem from a number of factors—possible growing awareness of the efficacy of CBT among practitioners (at least within Los Angeles County) since these prior studies were published, the particular voluntary sub-group of providers who participated in the study (who by virtue of volunteering for this study, may have been more open-minded), or the adult community mental health practice context, in which the size and intensity of the caseloads leads clinicians and administrators to be more open to learning something new (like CBT). Regardless of the cause, as a consequence, the most prominent themes to emerge at the organizational level pertained to supervision and training in treating patients with anxiety disorders using CBT. Both providers and administrators described insufficient organizational structure and support for providers to consistently learn and develop as CBT clinicians, which presumably leads to the lack of confidence providers expressed in treating patients with anxiety disorders. Patients noted that the frequent turnover of therapists served as a barrier to the therapy process. Yet training and supervision in CBT would give therapist providers the opportunity for professional development, which a study showed was the largest single predictor of lowered intention to quit among community mental health social workers (a major contributor to therapist turnover; Acker 2004). Thus, CBT training and supervision could enhance patient care in multiple ways.

At the system level, several interesting themes emerged. First, the idiosyncratic manner in which the DMH selects EBPs to include on its list from which clinics can choose to train their providers (using MHSA funds) reportedly omits exposure-based CBT for anxiety disorders, arguably among the most effective treatments for any psychiatric condition (Deacon and Abramowitz 2004). Second, the DMH funding streams support “early intervention” treatment (i.e., first-time service seekers) and “severe mental illness” treatment (i.e., typically disorders associated with severe disability such as schizophrenia and bipolar disorder), which appears to overlook a presumably sizeable proportion of those suffering from anxiety disorders - the most prevalent class of mental disorders a class which by DSM definition causes distress and impairment (as reported also by participating patients). Third, providers reported that, due to these system-level barriers (i.e., providers’ time constraints and level of care eligibility issues), their desire was to provide appropriate referrals for their patients to receive CBT when the patients were unable to receive it in their clinics. However, providers reported difficulty making appropriate referrals to CBT for anxiety disorder treatment, identifying yet another system level barrier for patients to receive appropriate care for anxiety disorders.

Potential solutions recommended across the provider and organizational levels point primarily toward offering training in exposure-based CBT for anxiety disorders to providers and their supervisors, followed by support to facilitate its use (e.g., individual supervision of the first few cases, monthly consultation groups). Across the patient and provider levels, establishing a strong therapeutic alliance prior to exposure was recommended as well as ensuring a strong match between the presenting problem and treatment goals. The latter would be enhanced by more careful and systematic assessment of anxiety disorders at intake. At the service system level, ensuring leadership buy-in and creating system incentives that focus on better patient outcomes both served as facilitators. Possibilities for meeting these recommendations include seeking and cultivating an on-the-ground champion of CBT for anxiety disorders within a given clinic or system, tracking and incentivizing better outcomes from anxiety disorder cases. Long-term goals include working with the DMH system to demonstrate the importance of financially reimbursing exposure-based CBT treatment for anxiety disorders and training/supervision for providers thereof.

Study Strengths and Limitations

This study sought rich and detailed information from patient, provider, and administrator stakeholders to gain a preliminary understanding of the barriers to receiving and delivering CBT for anxiety disorders in safety-net settings for adult mental health. We used the knowledge gained herein to develop questionnaires that tapped into a variety of putative barriers uncovered through these interviews and focus groups to gather quantitative data from larger samples (see Wolitzky-Taylor et al. 2018). As with many exploratory studies using qualitative methods, there are some limitations. First, results from this study are subjective and subject to biases in interpretation. To address this limitation, we developed a thematic coding scheme by using a widely used theoretical approach (grounded theory Padgett 2012; Charmaz 2006); moreover, two coders independently analyzed all of the data and afterwards worked to resolve disagreements through discussion until they reached consensus. Further, we employed standardized sets of questions for each stakeholder type. Second, results are based on small samples of stakeholders. Although we made efforts to sample from a diverse group of stakeholders in four distinct geographic communities in Los Angeles, it is difficult to ascertain the extent to which these results will generalize to other to stakeholders in these communities. Research with larger samples across more clinics and agencies are needed. Relatedly, stakeholders self-selected for participation in which they knew they would be discussing treatment for anxiety disorders at their clinic. Thus, there could have been some selection bias. Third, although we used standardized sets of questions for each stakeholder type, the very nature of focus groups and interviews is less reliable as a measurement tool than questionnaires or previously validated interview questions. However, the primary purpose of the study was to explore a variety of putative barriers in order to identify repeating ideas and themes from which to generate hypotheses, which can be used to develop appropriate quantitative instruments to test those hypotheses. Given the dearth of previous work in this area, no existing standardized instruments were available to assess many of the topic areas we covered in our interviews and focus groups.

Conclusions

This study represents the first to engage diverse stakeholders in adult safety-net mental health clinic settings to gain an initial understanding of the barriers to delivering and receiving exposure-based CBT for anxiety disorders. We thus took a vital first step toward specifying the factors that underlie the dramatic gap between the unparalleled evidence base of CBT for anxiety disorders (Tolin 2010; Hofmann and Smits 2008; Stewart and Chambless 2009) and the reality that very few anxiety disorder patients in community mental health clinics appear to receive it (Wolitzky-Taylor et al. 2015). Common barriers identified included a scarcity of CBT training and supervision for providers and supervisors, logistical barriers that prevent patients from attending weekly CBT sessions, a lack of institutional incentive including limited institutional means to be reimbursed for such treatment, and patient perception of frequent provider turnover. Importantly, we leveraged the knowledge generated by this qualitative investigation to create, distribute, and quantify the frequency and severity of each potential barrier identified in the current study (Wolitzky-Taylor et al. 2018). Thus, step-by-step, we advance our understanding of what prevents community mental health patients from accessing one of the most effective and evidence-based treatments for any mental disorder (Deacon and Abramowitz 2004).