Despite increasing public attention to the opioid crisis, capacity for addiction screening and treatment is still inadequate (
1,
2). It is critical to expand access to these services in primary care, particularly in underserved communities (
3). Federally qualified health centers, also known as community health centers or as health centers, can play a critical role as portals of access to addiction care in underserved communities, particularly for medication-assisted treatment (MAT) (
4,
5).
Health centers are well positioned to screen for addiction and identify at-risk patients, because they see millions of patients a year for primary, preventive, and prenatal services. They are required to provide primary care services on site, and many provide other services, including mental health and addiction treatment. Health centers are also leaders in the provision of integrated care, because behavioral health providers are part of the care team in many health centers (
6). Barriers to the provision of addiction treatment services in health centers include inadequate funding and lack of provider training. In addition, privacy protections, such as 42 CFR Part 2, can limit information flow related to addiction treatment and impede the provision of addiction treatment in primary care and the integration of services (
7).
Health centers serve all patients regardless of their ability to pay. Revenue from payers, especially Medicaid, is critical for supporting health center operations and increasing access to addiction services. In 2016, nearly half of health center patients (49.2%, N=12,543,818) were covered by Medicaid, and 23.4% (N=6,059,126) were uninsured (
8). Health centers receive cost-related reimbursement from Medicaid, but not all health centers are eligible to receive Medicaid reimbursement for behavioral health services. One reason is that not all health centers are licensed as behavioral health facilities by their state behavioral health authority, which may be required for Medicaid reimbursement. In addition, only certain provider types are eligible for reimbursement by Medicaid in some states, and providers in health centers may not be eligible. Furthermore, not all health centers and their providers are included in Medicaid managed care networks as addiction treatment providers. Finally, prohibitions on same-day billing for behavioral health and medical services in health centers persist in some states, despite federal efforts, including most recently in the 21st Century Cures Act, to clarify that same-day billing restrictions should not exist (
9,
10)
For health centers, another important source of funding are grants from the Health Resources and Services Administration (HRSA). In March 2016, the fiscal year (FY) 2016 Substance Abuse Service Expansion Awards were issued to support health center activities to build capacity for addiction treatment (
11). Health centers that received this funding must expand capacity for addiction treatment. In 2010, about half of health centers provided some type of addiction services on site (
4).
This study explored correlates of addiction screening and treatment in health centers that are members of the Midwest Clinicians’ Network (MWCN), a voluntary professional network of health centers in ten midwestern states. Surveys were used to examine several types of on-site addiction treatment capacity: the availability of psychiatrists and certified addiction counselors, addiction counseling, and MAT. This study was also the first to examine the issue of health center eligibility for Medicaid reimbursement for addiction treatment, providing a more granular picture of the important role of Medicaid funding in supporting behavioral health capacity in health centers. Health centers serve many Medicaid patients, but the fact that health centers do not always receive reimbursement for addiction treatment is not well understood and has not been explored previously in the literature.
Methods
This study used three data sources: two surveys that were fielded to federally qualified health centers that participated in the MWCN and the 2016 Uniform Data System, an administrative data set reported by each health center organization to HRSA annually. The MWCN is a not-for-profit corporation that supports community health centers by facilitating the exchange of best practices among clinical and quality improvement staff in the region. We mailed surveys to behavioral health leaders (Health Center Survey) (April to June 2016) and primary care providers (PCPs) (Behavioral Health and Diabetes Provider Survey) (October 2016 to March 2017) at 132 community health centers in ten midwestern states: Illinois, Indiana, Iowa, Kansas, Michigan, Minnesota, Missouri, Nebraska, Ohio, and Wisconsin. Responses were received from all states except Nebraska. The response rate for the Health Center Survey was 60% (N=77 health centers), and the adjusted response rate for the Behavioral Health and Diabetes Provider Survey was 55% (N=515 PCPs) (more information is provided in an online supplement to this article). The surveys were exempted by the University of Chicago Institutional Review Board.
Measures
The Health Center Survey included items about types of behavioral health services, barriers to integrating medical and behavioral health care, funding for behavioral health services, behavioral health training for providers and staff, and health center characteristics. The survey asked, “Does your health center provide behavioral health services? (a) Yes, at all sites; (b) Yes, at some sites (how many sites?); (c) No.” The survey asked about specific behavioral health services provided at each health center by providing a list of services. For this study, data on provision of “substance use disorder counseling” and “medication-assisted treatment for opioids” were used. Addiction screening capacity was determined if the respondent chose the “substance use” response to the multiple-choice question, “Does your health center systematically screen for other behavioral health issues?” This study also used data from the yes-no question, “Does Medicaid allow for reimbursement of substance use disorder treatment at your center?” Medicaid expansion status in 2016 was determined on the basis of publicly available data (
12).
The Health Center Survey assessed training needs by asking for “the most pressing training needs for your primary care providers.” “Medication assisted treatment (MAT) of substance use disorders” and “screening, brief intervention, and referral to treatment (SBIRT) of substance use disorders” were options in multiple-choice answers. The impact of privacy protections on integrated care was captured by a 5-point Likert scale, ranging from strongly disagree to strongly agree, in reaction to the following statement: “Information privacy protections for substance use disorder treatment make providing integrated behavioral health care difficult.”
The Behavioral Health and Diabetes Provider Survey included items about care integration, behavioral health referrals, confidence in providing behavioral health care, depression screening and care, integration of care for diabetes and depression, and provider demographic factors. This study used data from the survey questions about training needs and each provider’s comfort level with providing counseling and MAT for addiction.
Survey responses were matched with publicly available data from 2016 on health centers’ patient populations and services from the Uniform Data System. Annually, each health center reports organizational characteristics, such as number of sites, urban or rural location, and receipt of funding for special populations, such as homeless individuals and migrant and seasonal farmworkers. Characteristics of each health center’s patient population are included in the data set.
Finally, with use of publicly available information from the HRSA Web site, a variable was created to reflect whether the health center received a FY 2016 Substance Abuse Service Expansion Award in March 2016 (
11).
Analysis
Health center characteristics, including addiction screening and treatment capacity, were summarized by means and SDs (for patient population data) and by frequencies and proportions (for other health center characteristics) by using SAS, version 9.4. Using unweighted data, univariate and multivariate logistic regressions and generalized linear mixed models via generalized estimating equations (GEE) were performed to test associations between health center characteristics and each addiction outcome. The Hosmer-Lemeshow test was applied to check logistic regression model fit. Due to the binary outcomes and the limited sample size, we selected nine predictors based on the literature that have potential effects on addiction treatment capacity into a single model for model selections (percentage of health center population without insurance, percentage of health center population with Medicaid insurance, number of clinic sites, location [urban or not urban], Medicaid expansion status, patient-centered medical home recognition at all clinic sites, Healthcare for the Homeless Program funding, FY 2016 Substance Use Service Expansion Award, and receipt of Medicaid reimbursement for addiction treatment).
Once all selected predictors were included in a single model, the backward model selection procedure was performed, with a p value cutoff of <0.2 from univariate results. The final model was selected on the basis of the smallest quasi-likelihood under the independence model criterion. The reference groups were health centers that did not receive Healthcare for the Homeless Program funding, those that did not receive a FY 2016 Substance Use Service Expansion Award, and those that do not receive reimbursement from Medicaid for addiction treatment. We also conducted a sensitivity analysis using multiple imputation to fill in missing values for 12 health centers, and the results were similar. Both logistic regressions and GEEs produced consistent results; we present the GEE results here. The GEE models considered within-state associations and used only health centers for which we had complete data.
Results
Of the 77 health centers responding to the survey, 29% received targeted grant funding from HRSA in March 2016 to expand addiction treatment capacity (
Table 1). Just over half (56%) of health centers indicated that they were reimbursed by Medicaid for addiction treatment, and 94% of health centers received Medicaid reimbursement for mental health treatment. Among the 77 health centers, 8% operated one clinic site, 60% operated between two and five sites, and 32% operated six or more clinic sites. In the sample, 34% of health centers were in a rural area.
On-Site Addiction Screening and Counseling Capacity
Of the 77 responding health centers, 70% in 2016 provided addiction screening, and 39% expressed a desire to add or expand addiction screening capacity (
Table 2). Sixty percent offered behavioral health services in all clinic sites. On-site addiction counseling was offered by 62% of health centers, 38% had a psychiatrist on staff, and 21% had a certified addiction counselor on site. MAT was offered by 26% of health centers. Seventeen percent offered neither addiction screening nor counseling.
Barriers and Training Needs
Barriers and training needs were identified by behavioral health leaders at 76 of the 77 responding heath centers. One in five (23%) considered information privacy protections for addiction treatment to be impediments to the provision of behavioral health care (agree or strongly agree) (
Table 3).
Half of the 76 behavioral health leaders (49%) indicated that PCPs needed more training in best practices for SBIRT. Of the 515 PCPs who responded to the Behavioral Health and Diabetes Provider Survey, SBIRT training was highlighted by one in three (32%) as a training need, and only one in three indicated that they felt confident providing addiction counseling (agree or strongly agree, 29%).
Of the 76 health center behavioral health leaders, one in three (35%) said that PCPs needed more training in MAT. Half of PCPs (49%) indicated that they would like to receive more training in the provision of MAT. Only one in five PCPs indicated that they were comfortable providing MAT (agree or strongly agree, 19%).
Factors Associated With Addiction Screening and Counseling Capacity
In multivariable analyses, the FY 2016 Substance Use Service Expansion Awards were associated with having a psychiatrist on site (adjusted odds ratio [AOR]=3.53, p=0.007) and providing MAT for opioid use disorder (AOR=20.87, p<0.001) (
Table 4). Receiving Medicaid reimbursement for addiction treatment was correlated with on-site addiction counseling capacity (AOR=5.06, p<0.001) and having a certified addiction counselor on site (AOR=6.74, p=0.032).
Discussion
This study found that Medicaid reimbursement for addiction services and targeted grant funding were associated with addiction treatment capacity in health centers. The correlations between Medicaid reimbursement for addiction services, addiction counseling, and certified addiction counselor capacity highlight the critical role of Medicaid in supporting addiction treatment capacity in underserved communities. Health centers that received FY 2016 Substance Use Service Expansion Awards were more likely than those that did not to provide MAT for opioid use disorders and to have on-site services from a psychiatrist. Barriers to the provision of behavioral health services in health centers included the need for training, as well as privacy protections for data about addiction treatment.
Similar to previous studies, this study found that Medicaid was an important payer for addiction treatment services in health centers (
13,
14). This study is the first to examine the policy variation in whether health centers are reimbursed for addiction treatment services provided to Medicaid beneficiaries, and we found that eligibility for Medicaid reimbursement was associated with behavioral health capacity in health centers. Efforts to ensure that state Medicaid officials are aware of the vital role that health centers play in creating addiction treatment capacity should include further research on the cost savings associated with treating patients at health centers (
15). State Primary Care Associations, which are organizations that receive funding from HRSA for providing training and technical assistance to health centers and other safety-net providers, are a natural partner to help health centers interface with state Medicaid agencies. A focus at the federal and state levels on parity enforcement means that, increasingly, gaining access to insurance means gaining access to affordable addiction treatment (
16). As part of the strategy to combat the opioid epidemic and increase addiction treatment capacity, health centers should be eligible for reimbursement from Medicaid for providing addiction services.
MAT capacity in health centers appears to be increasing over time. We found that MAT capacity was present in about one-fourth of health centers in our sample (26%, N=20), as compared with a previous study that found that 12.3% of health centers in 2010 (N=45) had on-site providers who were waivered to provide buprenorphine (
4). Because MAT is the gold standard for evidence-based treatment for opioid use disorder, expansion of MAT could greatly improve health outcomes for patients with opioid use disorder (
17). To combat the opioid crisis, recent investments to expand MAT capacity should be sustained.
This study highlights an important potential strategy for combating the opioid crisis. Because health centers are strategically located in rural and urban underserved communities—which are disproportionately affected by the opioid crisis—targeted grant funding for expanding addiction services at health centers could help curb the opioid crisis. Previous research highlights the association between grant funding and Medicaid reimbursement and treatment capacity in health centers (
14,
18). Capacity for delivering behavioral health services, in particular, has been shown to be sensitive to grant funding, and thus this might be an effective way to invest in increasing access to behavioral health services in underserved areas (
13). In addition, investing in health centers is an efficient way to increase access to care, because these sites already provide transportation, case management, and other “enabling” services. Health centers also play a key role in their communities, assisting with enrollment in health insurance, the Special Supplemental Nutrition Program for Women, Infants, and Children, and other social services. These services are particularly valuable for patients with addiction, who often face barriers to treatment initiation and retention related to the social determinants of health. HRSA continues to release funding to target the opioid crisis, including $200 million to federally qualified health centers, $111 million to rural organizations, and $70 million for workforce expansions (
19).
Previous research highlights the impact of workforce availability on health center behavioral health capacity; in areas with more behavioral health providers per capita, health centers were more likely to have behavioral health capacity (
9). Future research should aim to improve technical assistance to health centers in the recruitment and retention of addiction specialists, in addition to using peer support specialists to help increase addiction screening and treatment. Previous research has also highlighted the importance of using standardized screening tools, and health centers might need assistance with identifying the best screening tools to use (
5). Supporting team-based care can maximize the impact of the addiction workforce in health centers; in addition to providing direct patient services, addiction specialists can support PCPs as they screen for addiction and provide brief interventions.
The analysis had limitations, including the fact that the survey was fielded only to health centers in ten midwestern states affiliated with the MWCN. This group of health centers may be more engaged in quality improvement efforts because of its long-standing affiliation with University of Chicago and experience with quality improvement research (
20–
22). In addition, the Health Center Survey was administered starting in April 2016, only a month after some of the health centers received the targeted grant funding in March 2016; these results cannot be interpreted as causal because of the timing. Instead, this analysis describes the characteristics of health centers that received the targeted grant funding. Although an association was found between the receipt of targeted grant funding for addiction services and the availability of MAT, we do not know how much the grants improved capacity. Before receiving the targeted grants, health centers that received the grants might have already had higher addiction treatment capacity than other health centers. These findings might not be generalizable to the entire United States, and causal inferences cannot be made because the data are cross-sectional. It is also important to note that the survey data and the information from the Uniform Data System were collected from grantee organizations, not from individual health center clinic sites. Each grantee organization operates several care delivery sites; thus this analysis underestimated behavioral health capacity at the clinic level, because each grantee is likely to have behavioral health capacity in only some of its clinic sites.
Future research should include implementation science work to examine how to encourage broader adoption of evidence-based practices, such as MAT and targeted screening for at-risk populations. Another important topic is the impact of having behavioral health providers on site in primary care settings and strategies to convince providers to undergo the necessary training to obtain the Drug Enforcement Administration waiver enabling them to prescribe buprenorphine. Research shows that even among providers with the waiver, patient volume is low, and thus further qualitative and quantitative research should examine the barriers to prescribing buprenorphine. Future research should focus on how to increase access to addiction screening and treatment for special populations, particularly homeless individuals. Finally, research should focus on further exploring the positive spillover effects of addiction treatment, including increased odds of receiving preventive screenings (
23).
Conclusions
Improving access to addiction screening and treatment in primary care is a critical part of combating the opioid use disorder epidemic. Federally qualified health centers provide access points to addiction screening and treatment services in underserved communities, but health centers without addiction staff on site were less likely to screen patients for addiction and less likely to provide on-site MAT. Both targeted grant funding and Medicaid funding were found to be associated with addiction capacity in health centers.