Managed Care’s Dual Impact on Outpatient Substance Abuse Treatment: Methadone Maintenance vs. Drug Free

Abstract

The number and type of services offered at substance abuse treatment (SAT) facilities are important aspects of the quality of care. Managed care (MC) is a growing presence in SAT and has been shown to affect the provision of treatment. We expand on earlier work and examine the impact of managed care on the number and type of services offered by methadone maintenance (MM) and drug-free (DF) outpatient treatment facilities. We use the econometric technique of instrumental variables to address the issue of endogeneity of MC and service offerings, thereby allowing a causal interpretation of results. Using data from the 2000 National Survey of Substance Abuse Treatment Services, we find that MC significantly increases the total number of services offered in MM outpatient facilities by four, yet decreases the number by two in DF outpatient facilities. We also show how the impact on specific services differs by modality and provide explanations for our findings.

Introduction

The majority of substance abuse treatment (SAT) in the U.S. takes place in outpatient settings (Alexander, Lemak, & Campbell, 2003). Organizational and financial factors that affect outpatient care therefore can have a large impact on substance abuse treatment provided in the US. One such factor is managed care. Managed care has been growing recently in the general medical sector and increasingly so in drug treatment. Although there has been much research on managed care in the general medical sector, relatively little research has focused on the role of managed care (MC) and how it affects services offered in SAT, let alone outpatient SAT (OSAT). The goal of this paper is to analyze the causal impact of MC on the total number and range of services offered by OSAT facilities and to delineate the differential impact of MC across the methadone maintenance (MM) and drug-free (DF) modalities. Total services and detailed specific services have been found to increase the effectiveness of care and thus we consider them an indicator of the quality of care (McLellan, Arndt, Metzger, Woody, & O’Brien, 1993; Milby et al., 1996; McLellan et al., 1997; McLellan et al., 1998; Gould, Levine, & McLellan, 2000; Marsh, D’Aunno, & Smith, 2000; Smith & Marsh, 2002).

One goal of MC is cost containment; thus, its impact may be to reduce services offered due to cost constraints through the use of incentives and administrative requirements. However, MC is also concerned with quality, so its impact may be to increase service offerings by implementing appropriate clinical standards and oversight procedures. Thus, the net effect of MC on services offered at OSAT facilities is an empirical issue. A recent study (Olmstead, White, & Sindelar, 2004) finds that MC reduces total services offered at SAT facilities. This study examined the full spectrum of facilities, including inpatient, residential, and outpatient settings. However, pooling heterogeneous facilities may hide important differential impacts by type of facility. As outpatient care is the type most frequently provided to substance abusers, it is important to understand the impact of MC on service offerings in OSAT facilities. Furthermore, outpatient care is composed of two major types of treatment that differ dramatically: methadone maintenance and nonpharmacological counseling-based (i.e., drug free) treatment. Because these two types of outpatient treatment vary with respect to so many dimensions, it is likely that MC impacts each differently. In addition, problems and policy prescriptions vary across these two modalities, so it is critical to understand the impact of MC on each modality separately.

This paper contributes to the literature in several ways. It is the first paper to estimate separately the causal effect of managed care on services offered by outpatient MM and DF facilities. We look beyond pure association and use valid econometric methods (instrumental variables) to identify causal relationships between MC and services offered. We use facility-level data from the 2000 National Survey of Substance Abuse Treatment Services (NSSATS). These data allow us to examine the availability of 26 different services at over 6,000 OSAT facilities throughout the 50 States, thereby obviating the need to generalize from a sample.

Background

Managed Care

The prevalence of managed care in the general medical sector has grown substantially during recent years. Its growth has been documented in SAT as well.¹ MC aims to contain costs and to establish quality controls. Costs are controlled through a variety of mechanisms, including pricing, incentive based contracting, selective contracting, and utilization review. MC can affect quality through utilization review, staffing and service requirements, and other administrative controls and requirements. In addition, selective contracting allows MC organizations (MCOs) to choose only those facilities that meet their standards. Both quality controls and cost containment mechanisms may affect the services offered in a medical facility.

A contractual relationship between a facility and a managed care organization is the outcome of a dual decision: (1) the MCO must be interested in the facility, and (2) the facility must be willing to meet the price and review, reporting, and other administrative and services requirements of the MCO. The existence of an MC contract could mean that the facility has changed its prices and procedures to meet the requirements of the MCO, or it could mean that a SAT facility was attracted to the MC contract because it already had the appropriate characteristics. In the latter case of selective contracting, only a matching has occurred – no real change. In the former case, MC has resulted in changes. Studies of the impact of MC on SAT facilities must be careful to distinguish between a pure match and a causal change. Determination of a causal effect is a focus of our study and one means by which our study advances the field.

Differences Between SAT and the General Medical Sector

There is a large body of literature on the impact of MC on services and costs in the general medical sector (see for instance, Glied, 2000; Dranove & Satterthwaite, 2000; Miller & Luft, 2002). However, findings from the general medical sector do not necessarily apply to the substance abuse treatment sector due to a variety of differences between the two sectors. One important difference is the financing of treatment. In SAT, employment-based financing plays a relatively small role because much of the funding comes from public financing, including block grants. Also, in the general medical sector, there is a concern that medical care will be overused or misused because insurance pays for a large part of the expense. In SAT, there is less concern with moral hazard due to the large positive externalities associated with reduced drug abuse (Jofre-Bonet & Sindelar, 2001). In fact, many individuals who need treatment do not seek treatment (Substance Abuse and Mental Health Services Administration [SAMHSA], 2003).

Another important difference is that many of those who need SAT have other medical, economic, and social problems in addition to the addiction itself (e.g., infectious diseases, welfare dependence, criminal activities). Given the multiple needs of SAT clients, it is not surprising that the provision of a broad array of services – including social, psychiatric, legal, housing assistance, and others – has been shown to increase the effectiveness of substance abuse treatment (McLellan et al., 1993; Milby et al., 1996; McLellan et al., 1997; McLellan et al., 1998; Gould et al., 2000; Marsh et al., 2000; Smith & Marsh, 2002). Thus, there are many reasons why lessons learned from the general medical sector do not necessarily apply to SAT services and that the impact of MC on SAT must be examined separately.

Literature on MC and SAT

There is a small but growing body of literature on the impact of MC on SAT. In an earlier study using the NSSATS data set, Olmstead et al. (2004) examine the availability of a broad array of services and find that (1) managed care causes SAT facilities to offer fewer services overall, and (2) this effect is concentrated primarily in tests for infectious diseases (i.e., tests for TB, HIV/AIDs, and STDs). This raises policy concerns that managed care may reduce treatment effectiveness by limiting the range of services offered to meet patient needs. Further, reduced onsite medical testing may contribute to the spread of infectious diseases, which poses important public health concerns. MC, as might be expected due to utilization review, also tends to increase the likelihood of offering assessments and a few other services. These findings are based on a study of all facilities taken together (i.e., inpatient, residential, and outpatient: MM and DF). In the current study, we seek to understand how MC affects outpatient service offerings, differentiating between the MM and DF modalities.

In two county-based surveys, McNeese-Smith (1998) reports that 38% of surveyed SAT program directors in Los Angeles County eliminated some services in response to managed care, and Rivers (1998) reports that 20% of surveyed SAT program directors in Miami-Dade County claim MC caused reduced availability and accessibility of SAT services.

Lemak & Alexander (2001) find that MC is associated with (as opposed to causes) shorter treatment durations and fewer individual and group therapy sessions in OSAT. They use a data set of about 600 facilities, and their sample is drawn from a national database that is similar to the one that we use. They collect additional data not found in the national sample and are thus better able to measure specifics of managed care, including MC behaviors such as authorized visits and other utilization controls. In related OSAT studies, Alexander and Lemak (1997a, 1997b) find that (1) MC results in additional administrative burdens due to compliance with utilization review, and (2) MC is most likely to be found in hospital-based care.

Methadone Maintenance vs. Drug-Free Treatment

The MM and DF modalities differ along a number of important dimensions, including their approach to treating addiction, financing, prevailing profit status, and locational demographics. For example, MM primarily treats opiate abusers while DF treats a variety of other addictions. The recommended length of stay in MM exceeds one year while the length of stay in DF is often measured in weeks. MM facilities are more likely than their DF counterparts to be for profit, located in metropolitan areas, and to accept Medicaid, but less likely to accept private health insurance. In addition, medical testing to prevent the spread of infectious diseases (e.g., HIV/AIDS, STDs, TB) is more of a concern in MM due to the greater incidence of intravenous injections associated with opiate abuse. These differences suggest that the impact of managed care may vary by modality and thus MM and DF facilities should be examined separately in empirical analyses of the impact of MC on OSAT service offerings.

Methods

Analytic Approach

Our aim is to estimate the causal effect of MC on OSAT service offerings. We have data that allow us to examine empirically the relationship between MC and service offerings. However, if we merely regress services on MC we will obtain an association, but not necessarily a causal relationship. This is because both MC and service offerings are likely to be correlated with other factors that affect both variables. For instance, if MCO’s prefer to contract with facilities that are capable of responding to requests for utilization review, providing care of sufficient quality and providing a range of services to covered MC patients, then those facilities that are found to have MC contracts are also likely to provide a broad set of services. In this scenario, a regression could reveal a positive association between MC and service offerings, but not a causal relationship. Controlling for additional factors could help to alleviate the problem. Thus, we control for those relevant factors for which we have data. Nonetheless, there are likely to be unobserved factors that would produce a spurious correlation. Administrative sophistication, for example, is unobserved in our dataset and likely to be positively correlated with services offered and MC, thus creating a spurious correlation between MC and services.

To overcome the potential problem of spurious correlation, we use an econometric approach called the instrumental variables (IV) approach.² The IV approach is used frequently in economic applications and is becoming increasingly common in healthcare evaluation (McClellan & Newhouse, 2000). We use the IV approach to account for possible endogeneity (or spurious correlation) between a facility’s involvement in MC and its service offerings. The instrument is a variable or set of variables selected to be highly correlated with MC and uncorrelated with service offerings. With valid instruments, the IV approach accounts for endogeneity and produces consistent estimates of the causal effect of managed care on service offerings (Davidson & MacKinnon, 1993). We use the following two-equation model:

$$S=\beta X+\gamma MC+\epsilon$$ (1)

$$MC=\alpha X+\kappa Z+\mu$$ (2)

where S is the outcome variable “services offered” (i.e., either total number of services at the facility or the presence of a specific service), MC is a dummy variable indicating the presence of managed care at the facility, X is a vector of facility and environmental characteristics, ε is the error term for equation (1) that captures unobserved determinants of S, Z is a vector of instruments that influence the presence of managed care but are uncorrelated with ε, and μ is the error term for equation (2).

We use two types of outcome variables to measure service offerings at a facility: the total number of services offered at the facility and the presence of a specific service at the facility. When the outcome variable in (1) is total number of services, the two-equation system is estimated using a full-maximum-likelihood “treatment effects” model (Maddala, 1983) that considers the effect of an endogenously chosen binary treatment (managed care) on another endogenous, fully-observed, continuous variable (number of offered services). When the outcome variable in (1) is the presence of a specific service (a binary indicator), the two-equation system is estimated using a maximum likelihood bivariate probit model (Maddala, 1983). In both models, facilities are clustered by county to account for possible spatial correlation amongst facilities operating in the same geographic area.

We estimate the system of equations separately for the methadone maintenance (MM) and drug-free (DF) modalities. We estimate separate models because we think that the independent variables are likely to differ by modality in their impact on service offerings.

The IV approach requires finding one or more variables to be used as instruments, Z, that substantially affect MC but have no direct impact on S. We use two different county-specific instruments in our model: number of Medicare MC enrollees and number of HMO enrollees. Given that we control for county population and density, the number of Medicare MC (or HMO) enrollees is, in essence, a percentage.

Both instruments plausibly satisfy the first correlation condition inasmuch as OSAT facilities in counties with a large number (“percentage”) of Medicare MC (or HMO) enrollees are more likely to have agreements or contracts with managed care than their counterparts in counties with relatively few Medicare MC (or HMO) enrollees. Similar county-specific environmental factors (e.g., local tastes, state/local laws, demographics) could cause this correlation. Spillover effects of many kinds could increase the favorable climate for managed care in OSAT. As for the second condition, it seems plausible that the number of Medicare MC enrollees in a given county is unlikely to directly influence the service-offering decisions at OSAT facilities in that county. Medicare covers primarily those 65 or older and OSAT facilities treat very few elderly (approximately 1.5% of OSAT clients are 65 or older). Similarly, there is likely to be relatively little overlap between HMO enrollees and OSAT clients (about 11% of the total population in the U.S. is drug dependent and only a fraction of them seek substance abuse treatment). We assess formally the validity of our instruments by testing for the presence of bias due to weak instruments and conducting a standard test of overidentification.

Data

We use the 2000 National Survey of Substance Abuse Treatment Services (NSSATS), which is designed to collect data on the characteristics of all SAT facilities and services in the U.S. (United States Department of Health and Human Services [USDHHS], 2002b).³ Of the original census of 14,622 SAT facilities, 13,749 (94%) completed the survey. Of these, we select only outpatient SAT facilities whose primary focus is either substance abuse or “substance abuse and mental health” (n = 7,498). We exclude facilities located in other U.S. jurisdictions such as American Samoa and Guam (n = 16), owned by federal or tribal governments (n = 38), or offering detoxification services only (n = 3). An additional 1,428 facilities are excluded due to missing data, resulting in a final study sample comprising 6,013 OSAT facilities (610 are MM and 5,403 are DF). We supplement the facility-level 2000 NSSATS data with county-specific data from the 2002 Area Resource File (USDHHS, 2002a), a database containing information on population and managed care activity in the U.S.

Tables 1 and 2 provide definitions and summary statistics for the variables used in our study. Statistics are presented for the full sample, as well as separately by modality (MM and DF) for facilities with and without relationships with managed care.

Table 1.

Summary Statistics of Outcome Variables¹

			Modality
		Full Sample	Methadone		Drug Free
Variable Name	Definition	N = 6,013	MC = 0 N = 378	MC = 1 N = 232	MC = 0 N = 2,198	MC = 1 N = 3,205
Total Services	Total number of services offered at SAT facility	12.85 (4.47)	14.73 (4.54)	17.45 (4.04)	11.66 (4.51)	13.12 (4.11)
Assessment
Substance abuse assessment	0-1 dummy variable, = 1 if facility offers substance abuse assessment	0.97 (0.17)	0.94 (0.24)	0.97 (0.17)	0.95 (0.21)	0.99 (0.10)
Mental health assessment	0-1 dummy variable, = 1 if facility offers mental health assessment	0.41 (0.49)	0.19 (0.39)	0.37 (0.48)	0.31 (0.46)	0.50 (0.50)
Therapy/Counseling
Family counseling	0-1 dummy variable, = 1 if facility offers family counseling	0.80 (0.40)	0.50 (0.50)	0.74 (0.44)	0.73 (0.44)	0.89 (0.31)
Group therapy	0-1 dummy variable, = 1 if facility offers group therapy	0.90 (0.30)	0.63 (0.48)	0.89 (0.32)	0.90 (0.30)	0.93 (0.25)
Individual therapy	0-1 dummy variable, = 1 if facility offers individual therapy	0.96 (0.19)	0.96 (0.21)	0.99 (0.11)	0.94 (0.24)	0.98 (0.16)
Pharmaco therapy	0-1 dummy variable, = 1 if facility offers pharmaco therapy	0.37 (0.48)	0.71 (0.45)	0.82 (0.38)	0.21 (0.41)	0.40 (0.49)
Relapse prevention groups	0-1 dummy variable, = 1 if facility offers relapse prevention groups	0.77 (0.42)	0.58 (0.49)	0.83 (0.38)	0.75 (0.44)	0.80 (0.40)
Aftercare counseling	0-1 dummy variable, = 1 if facility offers aftercare counseling	0.82 (0.38)	0.53 (0.50)	0.73 (0.45)	0.79 (0.41)	0.89 (0.31)
Medical Testing
Blood alcohol test	0-1 dummy variable, = 1 if facility offers blood alcohol testing	0.42 (0.49)	0.44 (0.50)	0.74 (0.44)	0.36 (0.48)	0.44 (0.50)
Drug/alcohol urine screen	0-1 dummy variable, = 1 if facility offers drug/alcohol urine screening	0.77 (0.42)	0.99 (0.11)	0.99 (0.09)	0.69 (0.46)	0.79 (0.41)
Hepatitis	0-1 dummy variable, = 1 if facility offers hepatitis testing	0.15 (0.36)	0.65 (0.48)	0.78 (0.41)	0.08 (0.27)	0.09 (0.29)
HIV test	0-1 dummy variable, = 1 if facility offers HIV testing	0.22 (0.42)	0.67 (0.47)	0.76 (0.43)	0.19 (0.39)	0.15 (0.36)
STD test	0-1 dummy variable, = 1 if facility offers STD testing	0.14 (0.35)	0.70 (0.46)	0.68 (0.47)	0.08 (0.28)	0.08 (0.27)
TB screen	0-1 dummy variable, = 1 if facility offers TB screening	0.25 (0.43)	0.94 (0.23)	0.93 (0.26)	0.16 (0.37)	0.18 (0.38)
Transitional
Assist in obtaining social  services	0-1 dummy variable, = 1 if facility offers assistance obtaining social services	0.43 (0.50)	0.54 (0.50)	0.69 (0.46)	0.35 (0.48)	0.46 (0.50)
Discharge planning	0-1 dummy variable, = 1 if facility offers discharge planning	0.77 (0.42)	0.76 (0.43)	0.91 (0.28)	0.70 (0.46)	0.81 (0.39)
Employment training	0-1 dummy variable, = 1 if facility offers employment training	0.29 (0.45)	0.47 (0.50)	0.42 (0.50)	0.27 (0.44)	0.28 (0.45)
Housing assistance	0-1 dummy variable, = 1 if facility offers housing assistance	0.21 (0.41)	0.26 (0.44)	0.36 (0.48)	0.19 (0.39)	0.21 (0.41)
Referral to other transitional  services	0-1 dummy variable, = 1 if facility offers referrals to other transitional services	0.81 (0.39)	0.88 (0.32)	0.93 (0.25)	0.75 (0.43)	0.83 (0.38)
Other services
Case management	0-1 dummy variable, = 1 if facility offers case management	0.64 (0.48)	0.67 (0.47)	0.73 (0.44)	0.60 (0.49)	0.65 (0.48)
Child care	0-1 dummy variable, = 1 if facility offers child care	0.10 (0.30)	0.06 (0.24)	0.12 (0.33)	0.10 (0.30)	0.09 (0.29)
Domestic violence	0-1 dummy variable, = 1 if facility offers domestic violence	0.35 (0.48)	0.16 (0.37)	0.23 (0.42)	0.35 (0.48)	0.38 (0.48)
HIV/AIDS education	0-1 dummy variable, = 1 if facility offers HIV/AIDS education	0.51 (0.50)	0.80 (0.40)	0.88 (0.32)	0.47 (0.50)	0.47 (0.50)
Outcome follow-up	0-1 dummy variable, = 1 if facility offers outcome follow-up	0.46 (0.50)	0.38 (0.49)	0.50 (0.50)	0.43 (0.50)	0.48 (0.50)
Transportation assist	0-1 dummy variable, = 1 if facility offers transportation assistance	0.27 (0.45)	0.24 (0.43)	0.30 (0.46)	0.25 (0.43)	0.30 (0.47)
Acupuncture	0-1 dummy variable, = 1 if facility offers acupuncture	0.06 (0.23)	0.10 (0.30)	0.16 (0.37)	0.05 (0.22)	0.05 (0.22)

¹Summary statistics are sample means with standard deviations in parentheses.

Table 2.

Summary Statistics of Regressors¹

			Modality
		Full Sample	Methadone		Drug Free
Variable Name	Definition	N = 6,013	MC = 0 N = 378	MC = 1 N = 232	MC = 0 N = 2,198	MC = 1 N = 3,205
MC	0-1 dummy variable, = 1 if facility has agreements or contracts with managed care organizations to provide substance abuse treatment services	0.57 (0.49)	0 (0)	1 (0)	0 (0)	1 (0)
Methadone	0-1 dummy variable, = 1 if facility dispenses methadone	0.10 (0.30)	1 (0)	1 (0)	0 (0)	0 (0)
Control: Private & for profit	0-1 dummy variable, = 1 if facility is operated by a private for-profit organization	0.35 (0.48)	0.56 (0.50)	0.35 (0.48)	0.37 (0.48)	0.31 (0.46)
Control: Private & nonprofit	0-1 dummy variable, = 1 if facility is operated by a private nonprofit organization	0.54 (0.50)	0.33 (0.47)	0.54 (0.50)	0.48 (0.50)	0.59 (0.49)
Control: Public	0-1 dummy variable, = 1 if facility is operated by a public organization	0.11 (0.32)	0.11 (0.31)	0.11 (0.31)	0.15 (0.36)	0.09 (0.29)
Focus: substance abuse	0-1 dummy variable, = 1 if primary focus of facility is substance abuse treatment services	0.67 (0.47)	0.96 (0.19)	0.92 (0.27)	0.71 (0.45)	0.58 (0.49)
Focus: mental health &   substance abuse	0-1 dummy variable, = 1 if focus of facility is a mix of mental health and substance abuse treatment services	0.33 (0.47)	0.04 (0.19)	0.08 (0.27)	0.29 (0.45)	0.42 (0.49)
Hospital	0-1 dummy variable, = 1 if facility is located in, or operated by, a hospital	0.09 (0.29)	0.15 (0.36)	0.16 (0.36)	0.03 (0.18)	0.12 (0.33)
Solo practice	0-1 dummy variable, = 1 if facility is a private solo practice	0.08 (0.27)	0.12 (0.33)	0.01 (0.11)	0.12 (0.32)	0.05 (0.23)
Halfway house	0-1 dummy variable, = 1 if facility operates a halfway house for substance abuse clients	0.06 (0.24)	0.01 (0.05)	0.04 (0.19)	0.04 (0.20)	0.08 (0.27)
Regfda2	0-1 dummy variable, = 1 if facility operates a narcotic treatment program regulated by the FDA	0.97 (0.17)	0.97 (0.17)	0.97 (0.17)	N/A	N/A
Accredited	0-1 dummy variable, = 1 if facility is accredited by JCAHO, CARF or NCQA	0.28 (0.45)	0.30 (0.46)	0.42 (0.49)	0.15 (0.36)	0.35 (0.48)
Licensed by state	0-1 dummy variable, = 1 if facility is licensed by a state substance abuse agency	0.93 (0.26)	0.97 (0.18)	0.97 (0.16)	0.88 (0.32)	0.95 (0.23)
Admissions (log)	Size of facility, measured by log of annual admissions	4.90 (1.15)	5.04 (1.06)	5.18 (1.05)	4.75 (1.19)	4.96 (1.13)
Accepts cash or self-payment	0-1 dummy variable, = 1 if facility accepts cash or self- payment for substance abuse treatment	0.96 (0.21)	0.99 (0.11)	0.99 (0.09)	0.90 (0.30)	0.99 (0.11)
Accepts private health insurance	0-1 dummy variable, = 1 if facility accepts private health insurance for substance abuse treatment	0.74 (0.44)	0.39 (0.49)	0.72 (0.45)	0.53 (0.50)	0.93 (0.25)
Accepts Medicaid	0-1 dummy variable, = 1 if facility accepts Medicaid for substance abuse treatment	0.57 (0.49)	0.52 (0.50)	0.89 (0.32)	0.37 (0.48)	0.70 (0.46)
Accepts Medicare	0-1 dummy variable, = 1 if facility accepts Medicare for substance abuse treatment	0.36 (0.48)	0.21 (0.41)	0.43 (0.50)	0.22 (0.41)	0.48 (0.50)
Accepts state-financed health insurance	0-1 dummy variable, = 1 if facility accepts state- financed health insurance for substance abuse treatment	0.39 (0.49)	0.10 (0.30)	0.41 (0.49)	0.22 (0.42)	0.53 (0.50)
Receives public funds (not Medicare, Medicaid)	0-1 dummy variable, = 1 if facility receives public funds (not Medicare or Medicaid) for substance abuse treatment	0.65 (0.48)	0.48 (0.50)	0.72 (0.45)	0.64 (0.48)	0.67 (0.47)
Offers payment assistance	0-1 dummy variable, = 1 if facility offers payment assistance for clients receiving substance abuse treatment	0.84 (0.37)	0.62 (0.49)	0.83 (0.38)	0.81 (0.40)	0.89 (0.31)
Region: Northeast	0-1 dummy variable, = 1 if facility is located in CT, MA, ME, NH, NJ, NY, PA, RI, or VT	0.24 (0.42)	0.25 (0.43)	0.49 (0.50)	0.13 (0.33)	0.29 (0.45)
Region: South	0-1 dummy variable, = 1 if facility is located in AL, AR, DE, DC, FL, GA, KY, LA, MD, MS, NC, OK, SC, TN, TX, VA, or WV	0.27 (0.45)	0.33 (0.47)	0.21 (0.41)	0.34 (0.48)	0.22 (0.42)
Region: Midwest	0-1 dummy variable, = 1 if facility is located in IL, IN, IA, KS, MI, MN, MO, NB, ND, OH, SD, or WI	0.26 (0.44)	0.15 (0.36)	0.15 (0.36)	0.22 (0.41)	0.30 (0.46)
Region: West	0-1 dummy variable, = 1 if facility is located in AK, AZ, CA, CO, HI, ID, MT, NV, NM, OR, UT, WA, or WY	0.23 (0.42)	0.27 (0.44)	0.15 (0.35)	0.31 (0.46)	0.18 (0.39)
Metropolitan area	0-1 dummy variable, = 1 if facility is located within a metropolitan area	0.75 (0.43)	0.98 (0.15)	0.94 (0.23)	0.77 (0.42)	0.70 (0.46)
Population density – county (log)	Log of the number of people in the county per square mile	6.01 (2.00)	7.30 (1.90)	7.40 (1.64)	5.96 (2.00)	5.79 (1.94)
Population – county (log)	Log of the county population	12.52 (1.67)	13.67 (1.22)	13.49 (1.04)	12.57 (1.72)	12.29 (1.64)
Competition – county	Herfindahl Index in the county	0.28 (0.29)	0.14 (0.15)	0.12 (0.14)	0.28 (0.29)	0.31 (0.31)
Medicare MC enrollees – county (log)	Log of Medicare managed care enrollees in the county	7.73 (3.30)	9.50 (2.46)	9.69 (1.81)	7.78 (3.37)	7.34 (3.29)
HMO enrollees – county (log)	Log of HMO enrollees in the county	10.60 (3.09)	12.18 (2.12)	12.36 (1.42)	10.53 (3.34)	10.33 (3.00)

¹Summary statistics are sample means with standard deviations in parentheses.

²Applicable only to those facilities that dispense methadone.

The outcome variables (Table 1) include the total number of services offered at OSAT facilities as well as binary indicators of each specific service. The 26 individual services are grouped by NSSATS into five categories: assessment, therapy/counseling, medical testing, transitional, and other.

The regressors (Table 2) fall into two categories: facility characteristics and county-specific characteristics. Our measure of managed care is whether or not the facility has a relationship with a managed care company, MC, set equal to 1 if a facility has at least one agreement or contract with a managed care organization to provide substance abuse treatment services, and 0 otherwise. Other facility characteristics include organizational control, setting, size, accreditation, and sources of revenue. NSSATS does not collect data on client characteristics, thus we cannot directly measure client mix. However, we control for types of payment accepted and ownership status, both of which proxy for client mix to some degree. County-specific characteristics include population size, metropolitan area, region, competition among facilities, and managed care activity.

Results

Number of Services

The raw data displayed in Table 1 show that, for each modality, facilities with MC offer more total services on average. Specifically, MC is associated with 2.7 more services in MMs and 1.5 more services in DFs. However, as shown in Table 2, facilities with and without MC differ systematically within each modality in other important ways as well. For example, for both MM and DF facilities, those with MC are more likely than their non-MC counterparts to be private and nonprofit, focus on both SAT and mental health, and be accredited. These factors could be associated with the availability of more services thus explaining the positive correlation between MC and services. This indicates a need to control for relevant covariates when estimating the impact of MC on service offerings. Moreover, MC is likely to be endogenously determined; that is, there may be unobserved factors (e.g., administrative sophistication) that affect both the availability of services and the presence of MC contracts. Thus, we turn to the IV models.

When we control for observed covariates and use IVs to account for possible endogeneity, we find that MC results in, on average, 4.2 additional services offered at MMs and 2.2 fewer services offered at DFs. Both effects are significant. Tables 3 and 4 show these results. Thus, it appears that the impact of MC on OSAT service offerings depends importantly on modality. F-statistics for both models are above 10.0, indicating that there is unlikely to be any bias due to weak instruments in these models (Staiger & Stock, 1997). Moreover, both models pass the standard test of overidentification (Davidson & MacKinnon, 1993; Kennedy, 1998; Greene, 2003), adding reassurance that equation (1) is specified correctly and that both instruments are valid.⁴

Table 3.

Total Services: Structural Equation Results (Equation #1)

	Methadone N = 610		Drug Free N = 5,402
Variable Name	Coef	p value	Coef	p value
Managed care	4.188	0.036	−2.223	0.014
Control: Private & nonprofit*	1.765	0.003	0.397	0.037
Control: Public*	2.811	0.000	0.847	0.014
Focus: substance abuse†	−1.459	0.074	−0.922	0.000
Hospital	0.296	0.700	0.401	0.131
Solo practice	−0.887	0.235	−0.830	0.001
Halfway house	−0.525	0.744	1.062	0.001
Regfda	1.588	0.201
Accredited	0.046	0.908	0.786	0.000
Licensed by state	0.947	0.403	0.867	0.000
Admissions (log)	0.405	0.009	0.462	0.000
Accepts cash or self-payment	−2.675	0.002	−1.701	0.000
Accepts private health insurance	0.692	0.165	1.516	0.000
Accepts Medicaid	−1.540	0.014	1.167	0.000
Accepts Medicare	1.261	0.009	0.718	0.000
Accepts state-financed health insurance	−0.429	0.536	0.981	0.000
Receives public funds (not Medicare, Medicaid)	0.004	0.993	0.827	0.000
Offers payment assistance	0.643	0.161	1.092	0.000
Region: South**	−0.359	0.555	−0.395	0.183
Region: Midwest**	−0.018	0.978	−0.994	0.000
Region: West**	−0.512	0.604	0.218	0.411
Metropolitan area	−0.912	0.331	−0.652	0.005
Population density – county (log)	0.228	0.226	0.403	0.000
Population – county (log)	−0.069	0.786	0.039	0.724
Competition – county	−0.194	0.912	−0.073	0.845
Constant	12.986	0.003	6.433	0.000
F-statistic for instruments (statistic, p_value)	16.2	0.000	24.6	0.000
Test of overidentification (statistic, p_value)	1.24	0.265	2.00	0.157

^*excluded category is “private & for profit”

^†excluded category is “mental health & substance abuse”

^**excluded category is “Northeast”

Table 4.

Total Services: Selection Equation Results (Equation #2)

	Methadone N = 610		Drug Free N = 5,402
Variable Name	Coef	p value	Coef	p value
Medicare MC enrollments (log)	0.198	0.001	0.044	0.006
HMO enrollments (log)	0.246	0.142	0.048	0.002
Control: Private & nonprofit*	−0.282	0.117	−0.223	0.001
Control: Public*	−0.658	0.027	−0.696	0.000
Focus: substance abuse†	−0.530	0.034	−0.146	0.004
Hospital	−0.580	0.005	0.222	0.034
Solo practice	−0.767	0.029	−0.349	0.000
Halfway house	1.198	0.024	0.347	0.000
Regfda	0.303	0.413
Accredited	0.229	0.180	0.279	0.000
Licensed by state	−0.665	0.025	0.213	0.019
Admissions (log)	−0.013	0.850	0.073	0.000
Accepts cash or self-payment	−0.356	0.596	0.357	0.025
Accepts private health insurance	0.462	0.002	1.162	0.000
Accepts Medicaid	0.755	0.001	0.319	0.000
Accepts Medicare	0.233	0.167	0.088	0.098
Accepts state-financed health insurance	0.659	0.000	0.331	0.000
Receives public funds (not Medicare, Medicaid)	0.244	0.213	−0.074	0.252
Offers payment assistance	0.217	0.227	−0.081	0.222
Region: South**	−0.182	0.507	−0.587	0.000
Region: Midwest**	−0.127	0.609	−0.367	0.000
Region: West**	−1.155	0.000	−0.519	0.000
Metropolitan area	−0.755	0.070	−0.177	0.045
Population density – county (log)	−0.103	0.152	−0.022	0.422
Population – county (log)	−0.622	0.002	−0.123	0.015
Competition – county	−0.987	0.184	−0.136	0.311
Constant	5.309	0.001	−0.160	0.771

^*excluded category is “private & for profit”

^†excluded category is “mental health & substance abuse”

^**excluded category is “Northeast”

Specific Services

Raw data comparisons in Table 1 show that facilities with MC are more likely to offer each specific service in the vast majority of cases. This is true for each modality. However, as noted previously, facilities with and without MC differ systematically within each modality on important covariates that probably affect service offerings, and MC is likely to be endogenously determined. Thus, we turn again to the IV models. Table 5 displays the results for the IV models corresponding to each modality. We report only the coefficients on MC and suppress the coefficients on the control variables to focus on the key results.

Table 5.

Effects of Managed Care on Individual Services^†

	Methadone		Drug Free
	Coef	Marginal*	Coef	Marginal*
Assessment
Substance abuse assessment	0.164 (0.926)	0.015 (0.928)	0.384 (0.170)	0.013 (0.220)
Mental health assessment	0.880 (0.062)	0.303 (0.049)	0.021 (0.969)	0.008 (0.969)
Therapy/Counseling
Family counseling	0.799 (0.185)	0.264 (0.196)	0.189 (0.566)	0.034 (0.572)
Group therapy	1.439 (<0.001)	0.374 (0.011)	0.309 (0.181)	0.027 (0.237)
Individual therapy	2.049 (<0.001)	0.000 (0.796)	−0.674 (0.025)	−0.045 (0.125)
Pharmaco therapy	0.745 (0.058)	0.197 (0.109)	−0.250 (0.593)	−0.078 (0.599)
Relapse prevention groups	1.238 (0.001)	0.389 (0.004)	0.307 (0.373)	0.096 (0.379)
Aftercare counseling	1.269 (<0.001)	0.420 (<0.001)	−0.084 (0.919)	−0.019 (0.920)
Medical Testing
Blood alcohol test	1.441 (0.006)	0.522 (0.002)	−0.307 (0.657)	−0.110 (0.660)
Drug/alcohol urine screen	0.997 (0.028)	0.003 (0.803)	0.316 (0.601)	0.081 (0.608)
Hepatitis	−0.020 (0.981)	−0.006 (0.981)	−0.540 (0.056)	−0.072 (0.114)
HIV test	−0.807 (0.291)	−0.096 (0.375)	−1.159 (<0.001)	−0.373 (<0.001)
STD test	−0.108 (0.860)	−0.040 (0.861)	−0.755 (0.024)	−0.117 (0.082)
TB screen	−0.454 (0.204)	−0.032 (0.228)	−1.050 (<0.001)	−0.292 (<0.001)
Transitional
Assist in obtaining social services	−0.137 (0.839)	−0.039 (0.837)	−0.776 (0.005)	−0.299 (0.004)
Discharge planning	1.179 (0.003)	0.316 (0.017)	0.550 (0.128)	0.166 (0.139)
Employment training	−1.341 (0.068)	−0.496 (0.030)	0.088 (0.824)	0.026 (0.823)
Housing assistance	−0.581 (0.483)	−0.202 (0.496)	0.145 (0.676)	0.040 (0.674)
Referral to other transitional services	0.906 (0.552)	0.207 (0.875)	−0.283 (0.349)	−0.077 (0.352)
Other services
Case management	0.088 (0.877)	0.029 (0.875)	0.069 (0.813)	0.023 (0.813)
Child care	1.285 (0.022)	0.253 (0.096)	−0.543 (0.416)	−0.102 (0.478)
Domestic violence	0.976 (0.006)	0.226 (0.026)	0.153 (0.686)	0.053 (0.684)
HIV/AIDS education	1.159 (0.041)	0.157 (0.372)	−0.995 (<0.001)	−0.364 (<0.001)
Outcome follow-up	0.486 (0.675)	0.185 (0.666)	−0.815 (0.012)	−0.314 (0.008)
Transportation assist	−0.957 (0.051)	−0.366 (0.037)	0.267 (0.557)	0.088 (0.551)
Acupuncture	−0.377 (0.722)	−0.061 (0.736)	−0.630 (0.312)	−0.053 (0.481)

^†Boldface entries are statistically significant (p value < 0.100).

^*Marginal effects measure the change in the probability that a given service is offered for a discrete change in MC from 0 to 1. Marginal effects are calculated for a typical methadone OSAT facility comprising the following characteristics: private & nonprofit, primary focus is substance abuse, not affiliated with a hospital, not a solo practice, not a halfway house, not accredited, licensed by the state, accepts cash or self-payment, accepts private insurance, accepts Medicaid, does not accept Medicare, does not accept state-financed health insurance, receives public funds (not Medicaid or Medicare), offers payment assistance, and located in a metropolitan area in the Northeast. Marginal effects are calculated for typical drug-free OSAT facilities comprising characteristics identical to methadone facilities except that they are located in the South.

MM Facilities

The large positive effect of managed care on MM service offerings is concentrated primarily in increasing the likelihood of counseling/therapy services (family, group, individual, pharmaco, relapse prevention, and aftercare). The IV bivariate probit coefficients are positive and significant for all of these services except family counseling. MC also increases the likelihood that MM facilities offer blood alcohol tests and drug/urine screens, which is consistent with the idea that MC favors monitoring and oversight. MC has no significant effect on the likelihood that MM facilities offer tests for diseases (e.g., hepatitis, HIV, STDs, TB). The impact is mixed on transitional and other services; although MC increases the likelihood that MM facilities offer child care and discharge planning, it decreases the likelihood that MM facilities offer employment training and transportation assistance.

DF Facilities

The negative effect of MC on DF service offerings is concentrated primarily in medical tests for diseases. That is, the IV bivariate probit coefficients are negative and significant for all of the tests of diseases in DF facilities (i.e., hepatitis, HIV, STD, and TB). Thus, managed care decreases the likelihood that DF facilities offer tests for infectious diseases. In contrast to its effect on counseling services in MM facilities, managed care has no systematic impact on the likelihood that DF facilities offer counseling services. Specifically, half of the counseling services have positive coefficients (the other half have negative coefficients) and only one is significant (individual therapy). It appears that MC has a somewhat negative impact on wraparound services offered by DF facilities, though the results are mixed. Although the IV coefficients for DF facilities are negative and significant for assistance in obtaining social services, HIV/AIDS education, and outcome follow-up, no other wraparound coefficients are significant for DF facilities and several are positive (e.g., employment training, housing assistance).

Discussion

We find that MC has significant and contrasting effects on the number and type of services offered at MM as compared to DF facilities. The finding that managed care increases the availability of counseling/therapy services in MM facilities is consistent with several viewpoints. One is that a role of MC is to ensure the provision of a certain standard of care. Methadone maintenance was originally defined to include methadone together with counseling services (Dole & Nyswander, 1965; Institute of Medicine [IOM], 1990), and methadone plus counseling has been shown to be both more effective (Ball & Ross, 1991; McLellan et al., 1993; Center for Substance Abuse Treatment [CSAT], 1995) and more cost-effective (Kraft, Rothbard, Hadley, McLellan, & Asch 1997) than methadone alone. However, counseling resources have been substantially eroded or limited in recent years due to fiscal constraints (IOM, 1990; CSAT, 1995). Thus, through a combination of standards and reimbursements, MC may be both requiring and enabling MM programs to increase their counseling/therapy service offerings. MC may set standards in other ways as well. For example, to the extent that DFs are more likely than their MM counterparts to offer counseling services, managed care may simply be trying to standardize care across modalities by requiring more counseling in MM programs.

MC’s positive impact on blood alcohol and drug/urine screens in MMs is consistent with MC use of utilization review and assessments prior to approval of additional care. It is not surprising, however, that MC does not impact the likelihood of substance abuse assessment inasmuch as this service is already offered by the vast majority of OSAT facilities.

The finding that MC decreases the likelihood that DFs offer tests for diseases is consistent with our previous work (Olmstead et al., 2004). We interpret this as MC being less interested in public health issues such as the spread of disease to others and/or that MC considers tests for diseases to be medical services that should be obtained at facilities that are more strictly medical.⁵ In any case, this finding has potentially important public health implications inasmuch as onsite availability increases the use of medical testing services (Umbrict-Schneiter, Ginn, Pabst, & Bigelow, 1994; National Institute on Drug Abuse [NIDA], 1999; D’Aunno, 1997; Friedmann, D’Aunno, Jin, & Alexander, 2000; Friedmann, Lemon, Stein, Etheridge, & D’Aunno, 2001), which, in turn, could reduce the spread of these infectious diseases.

Limitations

While we believe that we have added to the literature in several ways, including the use of a large data set that contains the universe of OSAT facilities and the ability to estimate causal effects, there are nonetheless several limitations to our study. NSSATS does not collect data on client characteristics, so we cannot control directly for client mix. However, we attempt to mitigate potential omitted variable bias by using several proxies for client mix, including types of payment accepted and ownership status. Another limitation of NSSATS is that it records managed care as a binary variable. Thus, we do not know the intensity, strength, types, and mix of MC mechanisms at each facility. Similarly, we know only if a service is offered, but not if the service has been received, who receives it, or anything about the intensity or quality of the service. However, using binary indicators of MC and also of specific services is probably conservative inasmuch as their relatively blunt nature could bias against finding significant results.

Conclusion

This paper provides useful insights into the causal effect of MC on service offerings in both methadone maintenance and drug-free OSAT facilities. These findings are consistent with the dual roles of MC: to restrict the growth of costs and to ensure at least a minimum standard of care. We conclude that, on average, MC significantly increases the total number of services offered at OSAT methadone maintenance facilities. In contrast, we find that MC significantly decreases the total number of services offered at OSAT drug-free facilities. These findings are derived using an IV approach in which we control for relevant factors and adjust for potential endogeneity between MC and service offerings. As such, the results can be interpreted as causal (as opposed to associations), subject to the limitations described above.

When we separately examine individual services offered, we find that MC significantly increases the likelihood that MMs offer a variety of counseling/therapy services, including group therapy, pharmaco therapy, relapse prevention groups, and aftercare counseling, as well as tests for drugs and alcohol. In contrast, we find that MC significantly decreases the likelihood that DFs offer tests for diseases, including hepatitis, STD, HIV, and TB.

The contrasting impacts of managed care on DFs as compared to MMs demonstrate the multiple roles that MC may play in affecting service offerings. Further research is needed to understand the effect of these impacts on patient outcomes.