The Journal of Behavioral Health Services & Research

, Volume 41, Issue 4, pp 460–472

The Relationship of State Medicaid Coverage to Medicaid Acceptance Among Substance Abuse Providers in the United States

Authors

    • College of Social WorkUniversity of South Carolina
Article

DOI: 10.1007/s11414-013-9387-2

Cite this article as:
Andrews, C.M. J Behav Health Serv Res (2014) 41: 460. doi:10.1007/s11414-013-9387-2

Abstract

The Affordable Care Act will dramatically increase the number of Americans with Medicaid coverage for substance abuse treatment (SAT). Currently, few SAT providers accept Medicaid, and consequently, there is concern that newly-eligible Medicaid enrollees will have difficulty finding SAT providers willing to serve them. However, little is known about why few SAT providers accept Medicaid. In response, this study examines how features of state Medicaid coverage for SAT, including benefits, eligibility, and oversight, are associated with Medicaid acceptance among SAT providers. Medicaid acceptance was positively associated with the number of SAT services covered, and the number of optional categorical expansions implemented by the state. Requirements for physician involvement were associated with lower odds of acceptance. The results suggest that more generous Medicaid coverage may encourage SAT providers to accept Medicaid, but regulatory policies may inhibit their ability to do so.

Introduction

The Affordable Care Act (ACA) represents the most profound policy shift to affect the substance abuse treatment (SAT) system since its inception, and is expected to transform the organization and financing of SAT services in the United States. Medicaid, the country’s health insurance program for low income citizens who are dependent, disabled, or aged, will be at the center of this transformation, as many Americans who receive specialty SAT services are low-income, and will become eligible for Medicaid in 2014. The ACA will dramatically increase Medicaid enrollment by providing states with the option to abolish categorical eligibility restrictions that limit enrollment to parents, children, elderly, and disabled individuals. Because the federal government will cover almost all of the cost of the expansion through 2019, many states have opted in, resulting in an estimated expansion of 12 million Americans.1,2 Moreover, the ACA will require all state Medicaid programs that participate in the expansion to provide coverage for SAT, and insure that limits on use of SAT services are no more restrictive than those placed on other medical and surgical services.3 Taken together, these provisions are expected to significantly increase insurance coverage for SAT in the US.4

Yet, it remains uncertain whether simply extending Medicaid coverage will result in increased access to SAT.4 There is widespread concern that the nation’s SAT system may lack the capacity to meet projected growth in demand for Medicaid-funded SAT services resulting from the ACA—a problem of great significance because at present, less than 20% of Americans who need SAT ever receive it.5 The number of Medicaid clients seeking SAT services is expected to grow significantly as a result of these changes.3 Medicaid enrollment among non-elderly adults is projected to increase by 82% among states that participate. But the expansion population exhibits high rates of substance use disorders, and as a result, demand for Medicaid-funded SAT services is expected to be especially pronounced.6 However, Medicaid acceptance among SAT providers is very low; only about half currently accept Medicaid, and in some states, such as California, acceptance is as low as 30%.7 While it remains unclear how the ACA will impact Medicaid acceptance in SAT, it is likely that some newly-insured Medicaid enrollees may face difficulties finding SAT providers who are willing and able to serve them.

In response, the present study examines how the design of Medicaid coverage for SAT is related to SAT providers’ acceptance of Medicaid enrollees. In general health care, research demonstrates that Medicaid coverage design is strongly related to provider acceptance of Medicaid.815 However, virtually nothing is known about how the design of Medicaid coverage may influence Medicaid acceptance in the SAT system—a system that differs vastly from general health care with regard to financing, structure, and case mix.16 To this end, this study brings together data on state Medicaid coverage and data from the National Survey of Substance Abuse Treatment Services (N-SSATS)—a population survey of all SAT providers in the US—to examine the relationship of key features of state Medicaid coverage for SAT to acceptance of Medicaid enrollees. As states will have broad discretion in structuring Medicaid coverage for SAT, knowledge of how its design may influence SAT providers’ willingness and capacity to serve the growing population of Medicaid enrollees is critical.

Background

Medicaid is the nation’s publicly financed health and long-term care coverage program for low-income Americans. Enacted in 1965 under Title XIX of the Social Security Act, it is an entitlement program that was initially established to provide medical assistance to individuals receiving cash assistance. Individuals gain Medicaid eligibility based on a combination of financial and categorical eligibility requirements. Beneficiaries must be low-income, although states make separate determinations of income thresholds. At present, states are required by federal law to cover children up to age 6 under 133% federal poverty level (FPL) and those aged 7–18 under 100% of FPL, pregnant women and infants up to 133% FPL, poor disabled and elderly (65+) recipients of Supplemental Security Income (SSI), and poor families with dependent children eligible for temporary assistance for needy families.17

Medicaid is a jointly financed intergovernmental program, largely administered by the states under a set of federal regulations. Federal spending levels for Medicaid are determined by the number of the people participating in the program, the extent to which enrollees utilize services, and the scope of services covered. Federal matching for state Medicaid programs is based on the average per capita income in each state, and as a result, matching rates vary. States are required to administer their own Medicaid programs, and each establishes its own eligibility standards, scope of services, reimbursement rates, and contract requirements within a set of federal guidelines. States also have the option to extend coverage to other populations and receive the federal contribution. Optional groups often include the same populations as above, but at higher income levels, as well as other groups such as working parents and the “medically needy,” defined as individuals with medical expenses especially high relative to their income.

Historically, states have had great discretion with regard to Medicaid coverage for SAT services. Federal guidelines have been vague: while they do not include coverage for substance abuse treatment, they also do not exclude it. Consequently, states vary greatly with regard to what coverage they provide and who can receive it. A few cover no SAT services, and among those that do, there is great variation in the types of SAT services covered and reimbursement rates for services. Moreover, eligibility for SAT services differs across states. Some states extend eligibility only to mandatory categories stipulated by federal program guidelines, while others have used Medicaid as a vehicle through which to extend health care coverage to a range of vulnerable populations, including low-income, medically needy, and working disabled individuals. Data indicates that states have increased coverage for SAT significantly through the Medicaid program over the past several decades. During the period from 1986 to 2003, Medicaid spending for substance abuse treatment increased from 10 to 21% of the total expenditures for SAT.18

The ACA’s state option to expand Medicaid to all individuals with incomes of up to 138% of the federal poverty line is expected to further increase Medicaid’s role in financing SAT in the US. This optional expansion will abolish existing categorical restrictions on eligibility, which limit enrollment to parents, children, elderly, and disabled individuals who meet income requirements. As the federal government will cover almost all of the cost of the expansion through 2019, many states are expected to opt in, resulting in an estimated expansion of 14 million Americans.19 While it is unknown exactly how this may influence demand for Medicaid-covered SAT services, the available estimates suggest that the proportion of individuals covered by Medicaid with behavioral health disorders is expected to increase substantially.6

Additionally, the ACA will require all states—regardless of their position in the Medicaid expansion—to provide basic coverage for SAT to all Medicaid enrollees. The ACA mandates all state Medicaid programs to provide SAT benefits comparable to their state’s essential health benefits benchmark plan, and to maintain parity between physical and behavioral health care with regard to service utilization and limitations. Finally, the ACA makes the integration of primary care, mental health, and substance abuse treatment an explicit priority. ACA provisions that promote the integration of substance use and mental health disorders with primary care include expanded funding for statewide implementation of screening, brief intervention, and referral to treatment programs and financial incentives to states to establish patient‐centered medical homes and accountable care organizations. These incentives, alongside broad expansion in eligibility and coverage for SAT services, are expected to result in dramatic expansion of Medicaid funding for substance abuse treatment.3

Medicaid design and SAT provider acceptance

Despite Medicaid’s growing importance for SAT, little is known about its influence on the SAT system. To date, only one study has examined SAT provider acceptance of Medicaid.20 Using N-SSATS, this study found that 55% of SAT providers in the US accept Medicaid. Using a dichotomous measure of state Medicaid coverage for SAT, they found that state Medicaid coverage for SAT was related to Medicaid acceptance, but only among SAT providers that received at least one other form of public funding. SAT provider acceptance of Medicaid was also positively related to organizational size, accreditation, public non-profit status, the availability of payment assistance, and the provision of mental or physical health services in addition to substance abuse treatment. Odds of acceptance were also positively associated with the percentage of residents who were White and the percentage of residents with income under the Federal poverty line in the state in which the provider was located.

While little research has examined Medicaid acceptance among SAT providers, there exists a broad base of research on factors influencing provider acceptance of Medicaid in general health care.815 Specifically, this literature indicates that broader eligibility criteria, a greater number of covered health services, and more generous reimbursement rates are linked to increased likelihood of Medicaid acceptance by health care providers.815 In contrast, limits on service utilization, use of utilization review measures, and reliance on capitation and bundled payment systems have been linked to decreased likelihood of Medicaid acceptance by health care providers.9,10,15 Prior studies have also found that residential segregation is linked to the providers’ likelihood of accepting Medicaid; they are more likely to occur in low-income communities, perhaps because more Medicaid clients live in there.14,21,22 More recently, Greene, Blustein, and Weitzman found that physicians were significantly less likely to participate in Medicaid in areas where the poor are nonwhite as well as in areas that are racially segregated.23

However, it is unclear whether similar dynamics are at work in the SAT system, which differs from general health care in several important ways.11 First, the financing structure is unique. Unlike medical and surgical services, SAT services are funded predominantly by public grants, consisting of the substance abuse prevention and treatment block grants funds and other general state revenue. The majority of SAT providers rely heavily on block grants.24 Private insurance for SAT services made up less than 25% of funding for all SAT services provided.25 Second, SAT providers that rely primarily on paraprofessionals and people in recovery may face challenges in meeting Medicaid requirements stipulating involvement of medical and other certified health care professionals in the provision of SAT services. In 2004, less than 50% of SAT providers had either a physician or a nurse on staff, and less than 50% of SAT staff possessed a professional or other master’s degree.26,27

Third, the SAT system lags far behind general health care with regard to the overall administrative and technological sophistication among providers. Medicaid program requirements related to medical billing and reporting require significant financial investment in information technology among providers that do not already have such capabilities. In a nationally representative sample of SAT providers, McClellan, Carise, and Kleber found that 20% had no information technology services (including voice mail), and 50% had information technology services available only for administrative staff.26 In light of these differences, there is a need for research that explicitly examines how the design of Medicaid coverage may encourage—or deter—Medicaid acceptance within the SAT system. While the study by McElrath, Chriqui, and McBride examines the relationship of any state Medicaid coverage for SAT to provider Medicaid acceptance, it does not examine how different facets of coverage, including coverage, eligibility, and provider oversight, may influence providers willingness and ability to accept Medicaid.20

Methods

Data sources

State-level data

Information on state Medicaid policies was culled from a report issued by the Substance Abuse and Mental Health Services Administration in28 (SAMHSA) contract with Abt Associates and the National Academy for State Health Policy (NASHP) to produce the report, which includes in-depth state-by-state information regarding Medicaid program benefits, eligible populations, oversight, and administrative practices. With assistance from several Medicaid policy experts, NASHP developed an instrument through which to cull data from public records describing state Medicaid programs. Information was gleaned from state Medicaid plans approved by the Center for Medicare and Medicaid Services. Additionally, Abt staff conducted internet searches of each state’s website for additional information on Medicaid benefits and eligibility related to mental health and substance abuse services. Once completed, the profiles were sent to state Medicaid program directors to confirm accuracy. Program directors also provided specific enrollment and cost data. While the report includes data for all 50 states and the District of Columbia, 8 states did not review their profiles. Data was collected in 2003–2004. Information about the percentage of state residents with alcohol or other drug use disorders was taken from the 2004 wave of the National Survey on Drug Use and Health.29

County-level data

County-level demographic data on population, race, ethnicity, and poverty was drawn from the 2000 Census.30

Provider-level data

Data on SAT providers was drawn from the 2004 wave of N-SSATS, which includes the population of all substance abuse treatment providers in the US. N-SSATS is designed to collect information from all facilities in the US, both public and private, that provide substance abuse treatment. The objectives of N-SSATS are to collect multipurpose data that can be used to assist the SAMHSA and state and local governments in assessing and analyzing general treatment services trends. Data are collected on topics including ownership, services offered (assessment, substance abuse therapy and counseling, pharmacotherapies, testing, transitional, ancillary), primary focus (substance abuse, mental health, both, general health, other), methadone/levo-alpha-acetylmethadol dispensing, type of treatment provided, number of clients, types of payment accepted, sliding fee scale, special programs offered, facility accreditation and licensure/certification, and managed care agreements.

The Inventory of Substance Abuse Treatment Services provides the sampling frame for N-SSATS, and includes two main categories of treatment facilities. The largest group of facilities includes those that are licensed, certified, or otherwise approved by the state substance abuse agency to provide substance abuse treatment. The second group represents treatment facilities that state substance abuse agencies do not license or certify, which typically include private for-profit providers, small group practices, or hospital-based programs. Detailed information regarding the data collection protocol can be found elsewhere.31 The N-SATSS response rate in 2004 was 96%, and the final sample size included 11,379 SAT providers (government-operated facilities were excluded) from all 50 states. A total of 1,905 counties included, which represented all counties that had at least 1 private SAT provider. Forty-nine percent of respondents completed the mail survey, 24% completed the survey via telephone, and 27% completed the survey using a web-based questionnaire.

Study variables

Among the study’s explanatory variables, generosity of benefits was measured using two variables. The first is the Medicaid Fee Index, a measure of the relative differences in reimbursement rates for Medicaid-financed services across states.32 These data were collected from state websites, and supplemented by a state survey conducted by the Urban Institute. The second variable measuring generosity of benefits is a count measure of the total number of seven possible substance abuse treatment-related services covered by states’ Medicaid programs. Possible services include case management, needs assessment, individual counseling, group counseling, intensive outpatient/day treatment; methadone maintenance; and buprenorphine and other drug administration and monitoring.

Scope of eligibility was measured using a count measure of the total number of “optional” groups covered by states’ Medicaid programs. Optional groups are defined as populations for whom states can elect to provide coverage under federal requirements, but are not required to do so. Some optional groups include the same populations as mandatory populations, such as parents, elderly, and disabled individuals, but at higher income levels. Other optional groups include new categorical populations, such as working parents and medically needy individuals. In this study, optional categorical expansions included income eligibility expansions for pregnant and parenting women, allowance for state-based SSI eligibility for Medicaid, existence of a medically needy program, use of the Omnibus Budget Reconciliation Act of 1986 option to expand income eligibility for elderly and disabled individuals, and existence of a work incentives program for employed individuals with disabilities.

Finally, intensity of oversight of state Medicaid program is operationalized using three variables. These measures attempt to gauge states’ involvement in monitoring and regulating substance abuse treatment providers’ provision of services. First is a measure of state programs’ use of managed care arrangements. This variable indicated whether managed care—through a traditional health maintenance organization, behavioral health carve out, prepaid health plans, or primary care case management plan—was the predominant structure of financing SAT within the state Medicaid program. The second measure gauged whether the state Medicaid program required physicians to provide, oversee, or approve Medicaid beneficiaries’ receipt services were coded affirmatively. The third measure indicated whether the state Medicaid program required state licensure for reimbursement.

The study’s outcome is a dichotomous variable measuring whether the SAT provider accepted Medicaid in the most recent fiscal year (beginning or ending in 2003). Control variables at the provider-level included ownership status (for-profit and non-profit, with the latter as referent), organizational affiliation (mental health center and freestanding, with the latter as referent), modality (residential and outpatient, with the latter as referent), provision of detoxification, provision of methadone, accreditation by the Joint Commission on Accreditation of Healthcare Organizations, state licensure status, whether the provider was a single individual (as opposed to an organization), and total past-year admissions, transformed into an ordinal variable. At the county-level, controls were included to measure the number of SAT providers in the county, population size (logged), the percentage of residents living under the federal poverty line, and variables measuring the proportion of residents who were a member of a racial or ethnic minority group. At the state level, measures of prevalence of alcohol and drug abuse were included, drawn from the National Survey on Drug Use and Health.30

Statistical analysis

Analyses were performed using Stata MP, version 11. Comparisons of all study variables were made for two groups as follows: providers that accept Medicaid and providers that did not. Analysis of variance were used to test for statistical significance; logistic regression with state-level random effects was used to examine potential relationships between state Medicaid program features and SAT provider acceptance of Medicaid.33 Results are reported as odds ratios with 95% confidence intervals. For modeling purposes, multiple imputation was used. Seven of the 21 variables in the study have missing data. The greatest proportion of missing data for a variable was 4.3%. Multiple imputation fills in the missing values by assuming the data were missing at random.31 Unlike other forms of imputation, multiple imputation represents uncertainty about the right value to fill in and thus overcomes the problem with single imputation. In the multiple imputation procedure, each missing value was replaced with five plausible values using the Markov Chain Monte Carlo method.34,35 The resulting five imputed datasets were merged for further analysis. A measure of goodness of fit was calculated that accounted for the use of the multiple imputation.36 Predicted probabilities were calculated for state-level policy variables significantly associated with outcome using the post-estimation tools for the xtlogit command in Stata MP, version 11.

Findings

Bivariate comparisons

About 55% of private SAT providers accepted Medicaid. Table 1 presents descriptive statistics for all state, county, and provider characteristics included in the model, as well as bivariate comparisons of SAT providers that did and did not accept Medicaid. There were significant differences among the two groups across all variables. On average, SAT providers that accepted Medicaid were located in states that covered a greater number of SAT services, offered higher reimbursement rates, and extended program eligibility to a greater number of optional populations. SAT providers that accepted Medicaid were more likely to be located in states in which the majority of services were administered through managed care arrangements, and less likely to be located in states that required state licensure or physician involvement in the provision of SAT. All county- and provider-level variables were significantly related to Medicaid acceptance. On the whole, providers that accepted Medicaid were more likely to be larger, non-profit, and multi-service programs. They were more likely to be located in smaller, lower-income counties with fewer racial and ethnic minority residents.
Table 1

Descriptive and bivariate statisticsa

  

Accepts Medicaid

 

Total

Yes

No

State-level variables

 Number of SAT services covered by Medicaid

3.636

3.840

3.408

 Medicaid FFS reimbursement index

9.956

10.070

9.852

 Number of optional groups covered by Medicaid

3.399

3.441

3.345

 Managed care predominant in Medicaid program

0.403

0.478

0.317

 Medicaid requires state licensure

0.676

0.693

0.659

 Medicaid requires physician involvement

0.449

0.393

0.510

 Percent of state residents with alcohol use disorders

7.822

7.791

7.844

 Percent of state residents with drug use disorders

2.946

2.942

2.952

County-level variables

 County population (logarithm)

12.819

12.584

13.067

 Percentage of county residents under FPL

0.155

0.155

0.154

 Percentage of country residents racial/ethnic minority

0.274

0.245

0.304

 Total number of SAT providers in county

0.401

0.340

0.465

Provider-level variables

 For-profit privately-owned

0.309

0.195

0.431

 Provides mental health services

0.337

0.426

0.241

 Solo/private-practice provider

0.061

0.024

0.103

 Number of SAT clients served in past year

4.560

4.985

4.118

 JCAHO accreditation

0.231

0.325

0.131

 Licensed by state to provide SAT

0.853

0.870

0.835

 Provides residentially-based treatment

0.275

0.203

0.351

 Provides detoxification services

0.203

0.215

0.191

 Provides methadone maintenance services

0.098

0.120

0.075

aAll bivariate comparisons in the above table are statistically significant at p < 0.01

Model results

Table 2 includes the results of the logistic regression model (R2 = 0.260). The results indicate that state Medicaid program coverage and eligibility were significantly related to SAT provider acceptance of Medicaid. Increased odds of Medicaid acceptance were associated with the number of SAT services covered by Medicaid (AOR 1.133; CI 1.019, 1.259) and the total number of optional eligibility expansions implemented by states (AOR 1.287; CI 1.065, 1.556). Conversely, residing in a state with a requirement for physician involvement in SAT was associated with decreased odds of Medicaid acceptance (AOR 0.636; CI 0.401, 0.994).
Table 2

Results of random effects logistic regression predicting Medicaid acceptance among private substance abuse treatment (SAT) providers in the United Statesa,b,c

 

AOR

LCI

UCI

State-level variables

 Number of SAT services covered by Medicaid

*1.133

1.019

1.259

 Medicaid FFS reimbursement index

0.950

0.854

1.058

 Number of optional groups covered by Medicaid

**1.287

1.065

1.556

 Managed care predominant in Medicaid program

1.574

0.982

2.522

 Medicaid requires state licensure

1.542

0.942

2.524

 Medicaid requires physician involvement

*0.636

0.401

0.984

 Percent of state residents with alcohol use disorders

1.118

0.891

1.403

 Percent of state residents with drug use disorders

0.621

0.267

1.443

County-level variables

 County population (logarithm)

***0.784

0.742

0.828

 Percentage of county residents under FPL

***1.021

1.008

1.035

 Percentage of country residents racial/ethnic minority

1.004

0.999

1.008

 Total number of SAT providers in county

***1.003

1.002

1.004

Provider-level variables

 For-profit privately owned

***0.324

0.291

0.361

 Provides mental health services

***2.552

2.289

2.844

 Solo/private-practice provider

***0.405

0.325

0.505

 Number of SAT clients served in past year

***1.077

1.055

1.099

 JCAHO accreditation

***1.865

1.644

2.116

 Licensed by state to provide SAT

***1.420

1.240

1.624

 Provides residentially-based treatment

***0.479

0.429

0.534

 Provides detoxification services

1.061

0.923

1.218

 Provides methadone maintenance services

***2.100

1.748

2.523

aThe dependent variable is a dichotomous measure indicating whether the SAT provide accepted Medicaid

bLCI and UCI refer to lower-bound and upper-bound confidence intervals, respectively

cAdjusted odds ratios with an asterisk are statistically significant at p < 0.05; two asterisks at p < 0.01; and three asterisks at p < 0.001

The change in the predicted probability of Medicaid acceptance for each of the variables significantly related to outcome, for a hypothetical provider at the “mean” for all variables, are as follows: (1) an increase of 2.8 percentage points in the predicted probability of Medicaid acceptance for each one unit increase in the number of substance abuse treatment services covered by the state Medicaid program; (2) an increase of 5.5 percentage points in the predicted probability of Medicaid acceptance for each one unit increase in the number of substance abuse treatment services covered by the state Medicaid program; and decrease of 11.0 percentage points in the predicted probability of Medicaid acceptance for providers located in states which physician involvement in the Medicaid program is required.

At the county level, the size of the population in the county in which the provider was located was associated with decreased odds of Medicaid acceptance (AOR 0.784; CI 0.742, 0.828), while the total number of providers in the county (AOR 1.003; CI 1.002, 1.004), and the proportion of county residents, were under the FPL was associated increased odds of Medicaid acceptance (AOR 1.021; CI 0.008, 1.035). At the provider level, greater odds of Medicaid acceptance were associated with having a primary focus on mental health services (AOR 2.552; CI 2.289, 2.844), being accredited (AOR 1.865; CI 1.644, 2.116), obtaining state licensure (AOR 1.420; CI 1.240, 1.624), and providing methadone (AOR 2.100; CI 1.748, 2.523). Additionally, past-year admissions were associated with increased odds of Medicaid acceptance (AOR 1.077; CI 1.055, 1.099). Lower odds of Medicaid acceptance were associated with being for-profit (AOR 0.324; CI 0.291, 0.361), a solo/private practice provider (AOR 0.405; CI 0.325, 0.505), and providing residentially based SAT (AOR 0.479; CI 0.429, 0.534).

Discussion

As described above, the findings of the study suggest that the decisions states make about coverage, eligibility, and provider oversight in their Medicaid programs may influence SAT providers’ willingness and capacity to accept Medicaid clients. SAT providers are likely to assess the financial viability of Medicaid program participation by considering potential profits and the size of the Medicaid population, i.e., potential customers. However, while some SAT providers may perceive Medicaid participation as profitable, regulatory policies such as physician involvement requirements may serve as barriers to Medicaid acceptance. Taken together, the findings suggest that states seeking to increase SAT provider participation in Medicaid may be able to use the design of the program itself to achieve this end. In particular, expanding the number of SAT services covered by Medicaid, and lifting requirements for physician involvement in treatment, when appropriate, holds the greatest promise for encouraging greater Medicaid acceptance among SAT providers.

The total number of SAT service covered was significantly associated with odds of Medicaid acceptance, suggesting that when providers are able to receive reimbursement for a broader number of services, they may be more likely to participate in the program. This finding builds upon the work of McElrath, Chriqui, and McBride, who found that any Medicaid coverage for SAT was associated with greater likelihood of participation in the program.20 The findings of this study suggest that the extent of coverage is also important. Moreover, unlike several other studies on participation, no relationship was found between reimbursement rates and participation in Medicaid. There are at least a couple of possible reasons for this. One is that the measure of reimbursement used was not sensitive enough to distinctions in rates in substance abuse treatment. The study relied upon a general index of differences in state Medicaid fees that was not specific to behavioral health care. Further research is needed that uses a more detailed, SAT-specific measure of reimbursement. Another possibility is that Medicaid reimbursement remains attractive to many SAT providers regardless of variation in fees, in light of the steadily declining availability of private insurance for SAT, and limited funding from public grants.

Acceptance of Medicaid was also related to the breadth of eligibility for Medicaid among the total population of SAT providers. The finding aligns with prior research in general health care demonstrating that providers will be more likely to participate in public insurance programs as the proportion of potential patients with public coverage increases.15 Due to economies of scale, a provider may be willing to accept relatively unattractive reimbursement rates if the number of potential clients to be served is high. This finding may also be related to the demographic changes associated with such eligibility expansions, which have often served to extend Medicaid eligibility to higher-income, less stigmatized populations including women, adolescents, and disabled individuals.37 This finding suggests the possibility that the expansion of Medicaid in many states as a result of the ACA could lead to an increase in the number of SAT providers that accept Medicaid-insured individuals. The increase in the number of potential Medicaid clients may enhance some providers’ perception of the potential profitability of serving this population. This finding is promising, in light of widespread concerns about whether the supply of SAT providers will be adequate to meeting expected increases in demand resulting from the expansion of Medicaid in many states.3

Of particular importance is the finding that SAT providers in states that required physician involvement in SAT has substantially lower odds of accepting Medicaid. As described above, SAT providers are less likely to employ professionals, and relatively few have a physician on staff.23,24 These findings open up the possibility that some SAT providers—particularly those in the non-profit sector—want to participate in Medicaid, but cannot meet with program’s staffing requirements for certification. Many studies have documented problems of underfunding within the substance abuse treatment, particularly among private non-profit providers, many of which lack funds to purchase basic voice and information technology.27 The cost of recruiting and retaining an addiction medicine specialist may be out of reach for many.

Further research is needed to understand the role of physician involvement in limiting participation in Medicaid. Studies are needed that focus on the particular challenges such policies may pose for providers, as well as potential technical assistance strategies that could be used to improve participation rates among this group. Research might also consider whether physician involvement is necessary in all Medicaid-funded SAT settings. While physician involvement may be necessary in settings that provide medications to treat addiction, it is unclear whether such involvement is needed in some other kinds of programs, such as recovery support. Such work may be of particular importance in light of the growing understanding of addiction as a chronic disease for which a broad-based continuum of care is essential. In light of the strong relationship of these physician mandates to provider Medicaid acceptance, policymakers might explore whether lifting these requirements may increase provider acceptance of Medicaid, and if so, whether there may be circumstances under which doing so would be appropriate and beneficial to treatment access.

Interestingly, neither of the other oversight variables in the study were significantly associated with Medicaid acceptance. Use of managed care as the predominant mechanism for financing SAT showed no association with Medicaid acceptance by SAT providers. This finding adds to the findings of several studies that have examined the impact of the introduction of managed care oversight of Medicaid-funded SAT on beneficiaries SAT access in single states.3842 The findings of these studies are divergent—in some states, Medicaid-managed care had a positive impact on SAT access, while in others, the impact was negative. In this cross-state study, the lack of a significant relationship between managed care and Medicaid acceptance may be due to the lack of nuance in the measurement of this complex construct. Further study that explicitly accounts for the complex features of individual managed care arrangements may help to further explain the complex role that managed care appears to play in Medicaid acceptance and service access.

State Medicaid program requirements for state licensure also showed no relationship to Medicaid acceptance. While the reasons for this are also unclear, it may be due to the fact that rates of state licensure are relatively high—approximately 85%—and thus do not reflect a significant barrier to Medicaid certification. The lack of association of state Medicaid program requirements for state licensure may also reflect the fact that for-profit SAT providers are both less likely to seek licensure and more likely to cater to the privately insured population.

Study limitations

The findings of this study must be considered within the context of several limitations. First, the study relies on cross-sectional data. As a consequence, it is not possible to establish the directionality of the relationships observed in this study. It is possible that SAT providers that participate in Medicaid could have influenced the design of coverage in their states through advocacy activities. Thus, the purpose of this study is limited to an exploration of the potential relationships between coverage and participation; further research is needed to understand the causal mechanisms that may underlie the relationships observed in this study. Moving forward, there is a need for longitudinal research to examine how provider participation is affected by changes in the design of Medicaid coverage for substance abuse treatment. Such research may also consider the role of providers as influential advocates in the design of coverage, particularly with regard to the generosity of benefits and reimbursement rates.

Second, the study is limited to examination of any acceptance of Medicaid enrollees—it cannot shed light on the extent of providers’ participation in Medicaid. Participation is a rather blunt measure of involvement; some providers may participate in Medicaid, but the proportion of Medicaid clients they serve may be low. There is a need for additional research examining the impact of coverage on the number of Medicaid enrollees served and, relatedly, the proportion of SAT providers’ caseload that Medicaid enrollees represent. Such research could have important implications for questions about the readiness of the substance abuse treatment system to respond to the impending expansion of Medicaid under the ACA.

Third, data used in this study were collected in 2003, and as a consequence, do not reflect possible changes in the SAT system that may have occurred since that time. Such changes may have altered the nature of the relationship between state Medicaid coverage features and SAT provider acceptance of Medicaid. However, review of NSSATS data from 2000 to 2010 indicate that participation rates have remained steady at just over 50%. Finally, measurement of state Medicaid programs’ reimbursement rates in this study is not ideal. Detailed information regarding state differences in Medicaid reimbursement rates for SAT was not available. As such, the study relies upon a fee index devised for general health care. It is possible that reimbursement rates in general health care differ significantly from those in SAT.

Implications for Behavioral Health

The findings of this study have implications for those charged with preparing for implementation of key provisions of the ACA in 2014. Policymakers, public administrators, and other stakeholders need to be aware that decisions about what is covered, who is covered, and how coverage is administered may make a real difference in SAT providers’ decisions regarding Medicaid acceptance. The findings suggest that SAT providers may weigh the potential incentives and disincentives of Medicaid program participation within the context of their organizational capacity and state policy environment. In particular, in states where Medicaid covers a broad array of SAT services, a significantly greater proportion of SAT providers are willing to accept Medicaid enrollees. This knowledge is significant in light of concerns that current rates of Medicaid participation are not high enough to meet the expected growth in demand for SAT generated by the program’s expansion under the ACA.

The findings of the study also suggest that the expansion of benefits and eligibility under the ACA may not result in increased participation if Medicaid certification requirements, such as physician involvement in the provision of SAT, prove difficult for existing SAT providers to attain. This could result in an expanded role in providing SAT for those providers which already accept Medicaid, and perhaps a diminished presence in the system for those that cannot. Moreover, as the ACA includes new incentives to promote care coordination and service integration through accountable care organizations and patient-centered medical homes, it is likely that larger, administratively sophisticated health and mental health providers will play an increasing role in the provision of SAT. To stay viable, SAT providers must be cognizant of the growing role of Medicaid in financing SAT services, and consider strategies to achieve certification.

Conclusions

The results of this study suggest that the design of state Medicaid coverage for SAT is significantly associated with provider acceptance of Medicaid enrollees. All three of the components of Medicaid coverage measured in the study were linked to odds of Medicaid acceptance including benefits, eligibility, and oversight. Providers in states that elected to cover a broader range of SAT services had greater odds of participating in Medicaid. However, actual rates of reimbursement were not associated with acceptance. Medicaid acceptance was also related to the breadth of eligibility for Medicaid among the total population of SAT providers. Accounting for other features of the design of coverage for substance abuse treatment, as well as organizational and county characteristics, SAT providers in states that covered a greater number of optional Medicaid populations were more likely to accept Medicaid-insured individuals.

With regard to oversight practices, Medicaid acceptance was significantly associated with state Medicaid program requirements for physician involvement in SAT, but not with predominance of Medicaid managed care or requirements for state-based provider licensure. All else equal, SAT providers in states with Medicaid program requirements for involvement of a physician in the provision or oversight of SAT were less likely to accept Medicaid patients, suggesting that such mandates may prove as a deterrent to Medicaid acceptance. Of all of the state policy variables included in the study, the requirement to involve physicians in treatment was most strongly associated with provider acceptance. In contrast, no such relationship between Medicaid acceptance and broad use of managed care by state Medicaid programs. Similarly, providers in states with Medicaid programs that required SAT providers to obtain state licensure to provide services were not less likely to accept Medicaid than providers in states without such requirements.

The results also suggest that characteristics of both the providers and the counties they are situated in were also linked to odds of Medicaid participation. Large, multiservice providers were more likely to accept Medicaid. Acceptance was positively related to both the number of clients served, as well as the provision of other mental health services in addition to substance abuse treatment. Additionally, the kinds of services provided were also important. Provision of methadone was associated with greater odds of acceptance, while provision of residentially-based treatment was associated with lower odds of Medicaid acceptance. Measures of quality were also important—being accredited by the Joint Commission and possessing state licensure to provide SAT were both associated with significantly greater odds of Medicaid acceptance. Finally, ownership structure was strongly related to acceptance, with for-profit providers of SAT being substantially less likely to accept Medicaid. Providers situated in counties that were smaller, lower income, and home to a smaller number of SAT providers had greater odds of accepting Medicaid.

Conflict of Interest

The author has no conflicts of interest to report.

Copyright information

© National Council for Behavioral Health 2014