Adoption of the National CLAS Standards in State Mental Health Agencies: A Nationwide Policy Analysis

Abstract

Objective:

This study reports the extent to which states have adopted the National Culturally and Linguistically Appropriate Services (CLAS) Standards.

Methods:

Officials from public mental health agencies in the fifty states, Washington, DC, and Puerto Rico were contacted to obtain current policies between January and June 2016. Each policy was coded through thematic analysis to determine correspondence with any of the fourteen National CLAS Standards which are divided into three domains.

Results:

Officials from 47 states and territories (90%) responded. Eight states (17%) reported adopting all National CLAS standards. Ten (23%) had no policies, five (12%) had policies under one domain, three (7%) had policies under two domains, and 25 (58%) had policies under all three domains.

Conclusion:

Most states still do not have policies that meet all CLAS standards, raising questions on how they should be adopted.

Government and professional bodies have recommended cultural and linguistic competence training for mental health clinicians and organizations based on hundreds of studies documenting disparities with racial and ethnic minorities initiating, continuing, and completing treatments compared to non-Hispanic Whites (1–2). The causes of disparities include low patient literacy, clinician biases, the lack of language-matched services in organizations, and the unequal geographical distribution of health resources (3). This last cause justifies public mental health policies in disparity reduction.

Therefore, the US Department of Health and Human Services’ Office of Minority Health (OMH) released fourteen national standards for culturally and linguistically appropriate services (CLAS) in 2000 for clinicians, organizations, accreditation bodies, and state agencies (4). The standards encouraged clinician services compatible with patient cultural beliefs, practices, and languages; an organizational workforce representing the demographic diversity in local communities; culturally and linguistically competent services; and ongoing self-assessments for accountability (4). In 2013, OMH released fifteen enhanced National CLAS Standards with one overarching Principal Standard and fourteen specific standards under the three domains of (1) “Governance, leadership, and workforce,” (2) “Communication and language assistance,” and (3) “Engagement, continuous improvement, and accountability” (5, available at https://www.thinkculturalhealth.hhs.gov/content/clas.asp). The enhancements included a new definition for cultural identities beyond race and ethnicity (such as sexual orientation), an organizational blueprint of exemplary practices, and recommendations to adopt the standards in public health organizations (5). As in 2000 (4), OMH has urged but not mandated adoption of all standards due to the lack of research proving that any single standard reduces disparities (5). To address this research gap, Assistant Secretary for Health Howard Koh has encouraged studies on National CLAS Standards adoption in various contexts, including state mental health agencies (6).

In June 2016, OMH found that 32 states undertook 172 CLAS-type activities in 2014–2015, from holding conferences and creating CLAS-related media to incorporating standards within strategic planning; only nine states adopted CLAS-specific policies, procedures, and regulations (7). The report’s methodology consisted of searches for any type of CLAS activity mentioned in websites and research databases, examining activities across physical and mental health agencies, and defining CLAS adherence as adoption of all standards wholesale (7). We wanted to understand the extent to which mental health agencies across all states, Washington, DC, and Puerto Rico (n=52) have adopted the National CLAS Standards within policies. We undertook a more recent investigation, hypothesizing that policies could be better identified by contacting state officials since websites and research articles may be outdated.

Methods

We sought to answer two interrelated questions: Which states have adopted all National CLAS Standards? If states have not adopted all standards, under what CLAS domains would their disparities reduction policies fall? A list of each state agency administering public mental health services in the fifty states, Washington, DC, and Puerto Rico was compiled by searching Google. Among the sixteen US territories, only Washington, DC, and Puerto Rico were included since the US Congress has granted them rights over local self-government (8). Hence, the fifty states, Washington, DC, and Puerto Rico can formulate policies independently from the federal government, permitting study of local National CLAS Standard adoption. Because states/territories differ over whether a single agency provides mental health services – for example, some states such as New York have separate agencies overseeing services for patients with certain disorders such as autism spectrum disorder – we only included the agency responsible for general adult services to ensure sample consistency.

We contacted the official listed on each website to obtain cultural competence and disparity reduction policies. We made up to two phone calls and sent up to five emails. We introduced ourselves, sought specific policies on disparity reduction, and checked the accuracy of policies on each agency’s website. We used a standard script: “Is this the most accurate and up-to-date information for your state? If not, could you direct me to a more current website or send us the state’s policies?” Institutional Review Board approval was exempted since the subjects of analysis were state agencies, not humans.

We compared all policies on websites with those from officials. The first two authors analyzed all verified policies through thematic analysis (9), widely used in mental health services research (10). We relied on OMH’s own assumption that state policies are valid data in examining National CLAS Standards adoption (5, 7). Each policy was uploaded into NVivo and coded deductively by the first two authors by identifying text corresponding to the CLAS standards. Since only nine states adopted CLASrelated policies in 2016 (7), any policy corresponding to the CLAS standards was defined as fulfilling the broader domain. The first two authors coded policies for ten states (~25%) independently to achieve an interrater reliability of 80% before coding all remaining policies jointly. Two rounds of coding were required to meet this benchmark. To ensure reliability, we drafted analytical memos on coding differences and thematic patterns. Peer checking and debriefing meetings were held weekly during the project.

Results

Fifty-two state officials (one per state/territory) were contacted. Officials from Delaware, Hawaii, Kansas, Maine, and Puerto Rico (n=5) did not respond; these states were excluded because we could not verify policies on their websites. Officials from 46 states and Washington, DC (n=47, 90%) responded. Of these 47, officials from Colorado, Georgia, Kentucky, and New Jersey (n=4) had no policies since these were in development at the time of study; these states were excluded from further analysis. Results are reported for 43 state/territory agencies whose policies either on websites or written were verified as current. The supplementary table specifies whether state/territory policies were current on websites and provides illustrative examples of policies coded according to CLAS domain.

These 43 states/territories fell into four categories depending on the number of domains covered in their policies. Ten (23%) reported no policies corresponding to any National CLAS Standard domain, five (12%) reported policies corresponding to one domain, three (7%) under two domains, and 25 (58%) under all three domains. Of these 25, eight states (17%) reported adopting all fifteen CLAS standards.

States were almost evenly divided in adopting policies under each domain. Thirty states adopted policies under the domains of “Governance, leadership, and workforce” and “Communication and language assistance.” Twenty-eight states adopted policies for the last domain.

No respondent refused to provide information and all were familiar with the National CLAS Standards. Respondents held various positions, ranging from Chief Community Relations Officer to Director of the Division of Behavioral Health. Officials forwarded our emails to others when they did not know their policies. Twenty-two officials (47%) confirmed that policies on websites were current and 25 (53%) noted differences between current policies and state websites since websites were outdated or current policies were not online.

Discussion

This is the first nationwide study to examine how mental health agencies have adopted the enhanced National CLAS Standards by obtaining policies from officials. As of June 2016, most agencies have taken an all-or-nothing approach. About a third have adopted all standards or policies under all CLAS domains. The remaining two-thirds had no policies or policies under three domains. Most agencies still have not adopted the National CLAS Standards three years after release.

Our study also reveals methodological challenges in examining CLAS adoption. OMH’s (2016) method of relying predominantly on websites would have produced inaccurate data in almost half of our sample. Our method of contacting state officials was more comprehensive but some did not respond, raising questions around best practices. To study disparity reduction, some differentiate policy adoption from service implementation (11). Instead of assuming that all National CLAS Standards must be adopted wholesale, an alternative approach could explore why states adopt certain policies over others. This framework can generate research evidence through case studies. Case studies on states with no policies can illuminate barriers to incorporating CLAS standards within policy making. Similarly, research on states with CLAS-related policies could identify facilitators and examine policy implementation within service settings to assess patient, provider, and organizational outcomes. We have focused on state mental health agencies, but methodological problems may exist with agencies administering general medical services. The CLAS standards also pertain to private for-profit and non-profit entities (5) where independent researchers may have limited access to information. Future work could recommend methodological best practices in these contexts.

The differentiation between adoption and implementation has emerged in the few studies on disparity reduction in public mental health settings. Clinicians and administrators may resist disparity reduction policies due to perceived costs in the time needed for asking patients “cultural” questions or interpreters (12). Private and non-profit clinicians and organizations look to public systems to implement reimbursable practices, and state agencies have historically pioneered cultural competence training practices (13). Hence, state mental health agencies can model National CLAS Standard adoption.

Our work has limitations. First, our data are valid for the period of investigation, and adoption status may have changed, though this is unlikely for the 43 states/territories from which we received policies. Second, our data may be restricted as states may have policies beyond the agency contacted, though it is unlikely that states would adopt CLAS-related policies in agencies for some mental health populations but not others.

Third, discrepancies between OMH’s study and ours may be due to methodology. We focused on mental health agencies rather than all medical agencies, and our study occurred in 2016 compared to OMH’s (7) which finished in early 2015. Both studies defined adoption differently, and we found that whereas only eight states adopted all National CLAS Standards, 33 had policies under at least one CLAS domain. To give states the benefit of the doubt, we focused on adoption by domain which prevented us from detecting differences in adoption for each standard within the domain, a focus for future work. Finally, some officials may have neglected to mention pertinent policies, especially since states like New Jersey have long engaged in cultural competence initiatives (14).

Nonetheless, our study provides the first nationwide analysis comparing state disparity reduction policies to the National CLAS Standards. OMH has repeatedly recommended adoption of these standards (4, 5). Most states have not followed this approach, and current scholarship is limited. Instead, studies of how state officials adopt and implement the standards within policies to close disparities may generate greater evidence. Treating the National CLAS Standards as an innovation meriting study through formal implementation and dissemination frameworks could organize future research (15). Such studies may help determine the efficacy of specific policies. For example, it is unknown whether certain policies are more effective in reducing disparities over others and whether states adopting certain numbers or types of standards are better positioned to reduce disparities than others. Finally, future work could study for whom policies are intended – state personnel, county agencies and providers, or independent contractors receiving state funding – and the mechanisms states can draw on to enforce policy implementation.