Adopting Evidence-Based Medically Assisted Treatments in Substance Abuse Treatment Organizations: Roles of Leadership Socialization and Funding Streams

Abstract

This paper examines the organizational adoption of medically assisted treatments (MAT) for substance use disorders (SUDs) in a representative sample of 555 US for-profit and not-for-profit treatment centers. The study examines organizational adoption of these treatments in an institutionally contested environment that traditionally has valued behavioral treatment, using sociological and resource dependence frameworks The findings indicate that socialization of leadership, measured by formal clinical education, is related to the adoption of MAT. Funding patterns also affect innovation adoption, with greater adoption associated with higher proportions of earned income from third party fees for services, and less adoption associated with funding from criminal justice sources. These findings may generalize to other social mission-oriented organizations where innovation adoption may be linked to private and public benefit values inherent in the type of socialization of leadership and different patterns of funding support.

Introduction

The treatment of substance use disorders (SUDs) has grown rapidly since the 1970s, now comprising a significant component of the U.S. health care industry with a gross expenditure on SUD treatment in 2006 estimated at over $20 billion (Mark, Levit, Vandivort-Warren, Coffey & Buck. 2007; Kimberly & McLellan, 2006). Together with the entire health care industry (Timmermans & Kolker, 2004), the sector dealing with the treatment of SUDs is under substantial pressure to adopt evidence-based practices. Among the innovative evidence based practices that can enhance outcomes of individuals' struggles with abuse of psychoactive substances is the use of several different pharmaceuticals or medication-assisted treatments (MAT). Yet, there are organizational leadership and ideological barriers to the incorporation of MAT in SUD treatment.

In this paper, we examine the effects of SUD treatment center directors' background and center funding sources on the organizational utilization of MAT in a nationally representative US sample of substance abuse treatment centers. We attempt to understand variation in the adoption of MAT across treatment centers in a treatment environment that has long been centered upon the value of behavioral interventions, abstinence, and particularly, 12-step approaches. We utilize the lens of formal educational socialization that influences professional identity (Hekman, Steensma, Bigley & Hereford, 2009), social capital (Lin, 2001), and resource dependence (Pfeffer & Salancik, 1978) to understand the extent to which they explain why some treatment centers adopt MAT while others do not. We focus on the relationship between the professional studies of treatment centers' executive directors, in terms of socialization dimensions, linking it to their organization's MAT adoption. We also examine the relationship between resource dependence through mixes of public and private benefit funding (Young, 2007) and MAT adoption.

Before presenting our hypotheses and reporting the results of our analyses, we describe the historical context of SUD treatment in which we outline “medicalized,” rather than punitive and social tolerance responses to SUD; describe what might be considered a treatment paradigm of abstinence achieved through 12-step programs and behavioral therapies; and, compare SUD and general medical treatment, setting the stage for understanding the context for organizational adoption of MAT as an institutionally contested innovation.

Substance Abuse Treatment

Medicalization of SUD

During the 20^th century, sociologists highlighted the medicalization of formerly deviant activity within American culture (Roman, 1980; Conrad & Schneider, 1980; Conrad & Leiter, 2004). Medicalization is one of three alternative societal reactions to problematic use of alcohol or other drugs. The second is generalized social tolerance and the absorption of the consequences of these behaviors into social functioning. This reaction prevailed throughout human history until the early 19^th century. The third alternative is negative sanctions and punishment, found in modest degree in reactions to the disruptive and non-productive consequences associated with public drunkenness over recorded history. Beginning in the 19^th century, these negative sanctions attracted major investment of social resources.

Drinking emerged as a social problem in the U.S. in the 1820s, and punishment first to excessive use and then to all use became the prime reaction (Clark, 1976). The U.S. eventually legislated nationwide prohibition, and this social attitude spilled over to other drug use that became prominent in the late 19^th and early 20^th centuries and persists today. Incomplete and ambivalent medicalization of illegal drug use is currently represented by sustained punitive attitudes and practices coupled with official policies advocating medicalization and treatment.

Treatment Paradigm

The social invention and refinement of Alcoholics Anonymous (AA) from 1935 to 1955 marked the emergence of a recognizable treatment for alcoholism within American culture (Roman, 1988). Its elements include self-reliance, populism (offering “membership” to anyone motivated to participate), organizational poverty (refusing money except for bare necessities), decentralization/localism (core structure is the “group”), bureaucratic rules (Twelve Steps and Twelve Traditions, unchanged for 75 years), isolation (no supportive ties with any other organizational structure) and medicalization (belief that the root of “loss of control” is a biomedical abnormality). Through these fundamentalist and cost-free elements, AA offers the private good of abstinence and “sobriety” to individuals, which produces a public good by reducing individual, familial, community, and societal costs of alcoholism.

In the mid-1940s, an affiliate non-profit organization, informally and tightly coupled with AA, was created: the National Council on Alcoholism, which had as its singular goal, “the recognition and treatment of alcoholism as a disease like any other” (Beauchamp, 1980). This “independent” collectivity strived to create a hospitable and respectable environment for the growth and institutionalization of the 12-step model. In 1970, these nascent interests gained a spectacular infusion of resources through federal legislation that provided what seemed like massive funds for treatment and research, setting the stage for organizational legitimacy (Winer, 1981).

A collection of isomorphic alcoholism treatment centers emerged supported from the mid-1970s to the late-1980s by health insurance and public grants. The absence of evaluative or cost-benefit data to justify to third-party payers the expensive month-long inpatient treatment regimen common across nearly all these private programs led to sharp cutbacks in insurance coverage and eventual limitation of coverage to outpatient care.

In the early 1990s, forty percent of the exclusively inpatient programs closed as managed care limited third party payments for treatment (Shane, Blum & Roman, 1997). While it might be expected that this form of enterprise would disappear, or perhaps survive in a form based only on public support, neither of these outcomes occurred. Instead a more heterogeneous collection, based on abstinence and 12 step programs, emerged.

Further, these changes were accompanied by shifts in public policy that rather quickly merged monies for the treatment of alcoholism, opiate addiction, and other substance abuse into what became, by the 21^st century, a single set of agencies for the treatment of all substance abuse. With this merger, the patient population came to include a significant element of socially dependent people with considerable contact with the criminal justice system whose recovery required far more assistance than simply abstinence from drugs and alcohol.

These challenges describe the entry and/or emergence of a new form of treatment center leadership, demonstrated by embracing a treatment philosophy that surrounded the basic treatment of addiction with “wraparound services” addressing needs across psychological, family, housing, employment and legal realms. This service expansion increased the support for recovery as well as provided opportunities for effective social functioning in post-treatment roles. Curiously, this foreshadowed contemporary emphases on the need to re-define primary care in a holistic manner in the form of a “medical home.” Within the SUDs specialty, a new paradigm was put forth, a distinct contrast to the earlier 12-step adage that all personal troubles would improve if the AA affiliate continued to “work the program.” In sharp contrast to the earlier rigid adherence to a single approach, the watchword of this new paradigm of SUD treatment is flexibility in service offerings and in ever-changing strategies for generating patient referrals.

Several principles of contemporary treatment with AA roots do however persist, even among those who deny adherence to that treatment ideology. These include the continued belief in abstinence from alcohol and drugs as the only acceptable criterion for treatment success, belief in alcoholism and addiction as biologically rooted, and continued openness to spirituality as a key to the recovery process.

SUD and General Medical Care

This leads to several important comparisons between general medical care and SUD treatment. First, using the outcome criterion of sustained post-treatment abstinence, SUD treatment shows spectacularly discouraging rates of “success,” leading to high rates of treatment recidivism. Together, “relapse” and recidivism support beliefs of pessimism about SUD treatment. Second, given the linkages between drug use and the legal system, a substantial number of referrals enter treatment without believing they are ill, i.e. without the self-image of being “addicted” and in need of treatment. Third, engagement in treatment is problematic, with a substantial proportion of patients entering treatment and quickly leaving, or leaving before completing the prescribed regimen. Fourth, patient problems and maintaining quality of care may be exacerbated by staff problems, with SUD treatment programs chronically underfunded and often understaffed, with relatively low wages and benefits and high rates of turnover (McLellan, Cerise & Kleber, 2003).

It is in this context of (1) a dominant abstinence paradigm and (2) striking contrasts between SUD treatment and general medicine that we focus upon the leadership of SUD treatment organizations and funding mix of SUD treatment organizations to understand the inclusion of MAT as an option across treatment centers.

Hypotheses

SUD Treatment Center Leadership

How can such organizations attract leaders? Perhaps surprisingly, our research demonstrates that leaders often stay for long periods within these organizations and many maintain high morale and involvement in what they do. If we combine the history of the type of organization with the organizational conditions that have been described, social entrepreneurship, the combining of social mission with commercial impulses, is an apt description of leadership job requirements. Leadership may be drawn to, and sustained by this work, through the modest improvement observed in some and dramatic improvement observed in a few. Further understanding may be captured by looking at the identity formation among leaders who apparently respond positively to seemingly low levels of objective rewards.

There are two approaches in the literature on professions to understanding how professional identity, an aspect of personal and social identity that develops in professional personnel as a result of their work activities, is established. The trait approach proposes that professional identity is defined by specific characteristics that describe the profession (Kultgen, 1988), while the phenomenological or “folk” approach proposes that professional identity is established by activities in which persons engage (Miller, 1998; Friedson, 1983). The activities described here in the brief history of SUD treatment not only present evidence for viewing these leaders in a context of social entrepreneurship, but also provide a “folk” approach to better understanding how these leaders are socialized and develop professional identity. One of the sources of socialization and development of professional identity, for these leaders can be found in the types of formal education they have received. Social categorization theory (Turner, Hogg, Oakes, Reicher & Wetherell, 1987) suggests that differential socialization that exposes potential organizational leaders to the paradigms that are associated with SUD treatment, to an emphasis on organizational management, to educational disciplines, or their own indigenous experiences with recovery are likely to be influenced by the norms and identities adopted by others who have shared their socialization experiences.

We next examine how this fits with adoption behavior. Much of the research on diffusion of innovation focuses on how structures and processes within organizations affect innovation dispersion through a population of potential adopters. Economic literature tends to answer this question on the basis of markets and information availability, while sociological explanations focus on conformity to environmental expectations.

Ansari, Fiss & Zajac (2010) provide an integrative model that conceptualizes the pattern of innovation diffusion as an outcome of the technical, cultural and political fit between the innovation itself and characteristics of the adopters, with the demand for innovation being associated with the organization's ability to recognize the value of an innovation associated with organizational factors such as the background and experience of decision makers (Wejnert, 2002). Diffusion may be enhanced by the content of the socialization of organizations' leaders. Professional identities are constructed through the course of training and residency programs (Pratt, Rockmann & Kaufmann, 2006), professional identities have implications for performance (Hekman, Steensma, Bigley & Hereford, 2009, Hekman, Bigley, Steensma & Hereford, 2009), and social and cognitive boundaries of professional groups can retard diffusion in communities of practice, which tend to be highly institutionalized (Ferlie, Fitzgerald, Wood & Hawkins, 2005).

Some of the careers and educational socialization of the executive directors of SUD treatment centers are typified by engagement in various types of formal clinical education that impart training for diagnosing and treating a wide range of behavior disorders. Other careers are typified by participation in the “folk” history of the field, with educational experiences that either did not include broad counseling preparation or advanced degrees. Included in this category are those who engaged in narrower certification programs for exclusive counseling of SUD patients.

Through treatment-related formal education, center directors may have developed a network of contacts different than those without this educational socialization, including those who have clinical degrees but do not work in SUD treatment. These leaders are more likely to share professional identities with others who engage in treatment of clients with disorders other than SUD, and who routinely utilize medications as part of treatment interventions. Leaders without clinical degrees would be less likely to develop these diverse ties (Ferlie, Fitzgerald, Wood & Hawkins, 2005). In addition to potentially greater influence from clinicians outside SUD treatment, leaders who engaged in various types of formal clinical education also have the potential to develop a network of distant ties, and therefore are more likely to be exposed to the value in adopting new and novel ideas (Perry-Smith, 2006; Perry-Smith & Shalley, 2003).

This is further explicated through the concept of social capital (Lin, Cook & Burt, 2001), which, in the context of a SUD treatment organization, is greater for those leaders who pursued clinical degrees than those who did not. Networks of social relationships stemming from this capital investment have embedded resources associated with increased information, influence, social credentials, and reinforcements (Lin, 2001) that in turn could enhance the adoption of new combinations of ideas and resources. Specifically, the increase in information flow can provide access to documented benefits of innovations and evidence based practices. The increase in influence and social credentials from social ties could provide greater accessibility to valued resources and provide legitimacy and credibility beyond personal capital. Social relationships where information and influence are shared reinforce identity and underline its value through the recognition that is received. Thus, through these mechanisms facilitated by social capital, relevant formal education among administrators of SUD treatment centers can have an effect on adoption of MAT in SUD treatment.

Adopting institutionally contested (Fiss & Zajac, 2004; Sanders & Tuschke, 2007) MAT innovations for treatment of SUD is related to the social capital of educational socialization and associated professional identity (Ferlie, Fitzgerald, Wood & Hawkins, 2005), with executive directors who have formal educational experiences in clinically relevant disciplines having “mindlines” that increase the likelihood of adoption of MAT in SUD treatment organizations. Therefore, we hypothesize:

H1: Across SUD treatment organizations, there will be a relationship between the extent of the adoption of MAT and the educational experience of the organizational leaders such that there will be greater adoption among those with clinical education in relevant disciplines.

Environmental Context of Resource Dependence

In their definition of social entrepreneurship, Murphy and Coombes (2009) emphasize the importance of context. They describe environmental context in terms of mobilization or large scale public support of the social entrepreneurial endeavor. In their description, support is actualized as a convergence of social, economic and environmental resources to facilitate identification of innovative opportunities. The source of funding to sustain the organization's core technology is one dimension of the environmental context that influences managerial strategies such as the extent of adoption of innovations.

Resource dependence theory argues that organizations need to take in resources from entities external to the organization, and organizations that control resources have power over those that need these resources (Pfeffer & Salancik, 1978). Thus the source of a health care organization's resources may be more important in understanding whether new practices are adopted than the legal status or ownership of the organization. Whether socially entrepreneurial organizations are privately or publicly owned, they must acquire resources to realize their mission.

Bozeman and Straussman (1983) argue that an understanding of resource dependence helps us understand the degree of control exerted by organizational sources of funding. Managerial strategies used to reach organizational objectives are influenced by resource acquisition processes regardless of organizational ownership - government, non-profit or for-profit; or their business model - classic non-profit, for-profit or hybrid (Elkington & Hartigan, 2008). The resource dependence of SUD treatment centers may trump the “cui bono” (Blau and Scott, 1962), or who benefits, approach to understanding organizational behavior.

In an undercapitalized social entrepreneurial context, an organization depending primarily on public funds that are scarce and highly competitive may be reluctant to embrace new and different values that may accompany an innovative way of addressing a social need. As Bozeman and Slusher (1979) argue, resource scarcity can result in less willingness to consider new opportunities. In both undercapitalized for-profit and non-profit organizations, where resources are unpredictable, adoption of innovation could be stifled in favor of efficiency norms, with an emphasis on risk avoidance strategies focused on operating at full capacity by delivering traditional services, at the lowest unit price, to as many clients as possible (Bozeman and Straussman, 1983).

One of the dilemmas faced by leaders in a resource constrained context when deciding whether to adopt an innovation is aligning the values of the new innovation with values held by those providing funding. Resource dependency theory suggests the importance of such alignment for organizational survival. Incompatible new values may be introduced to the status quo, such as introducing private good-type innovations in an environmental context that values public goods. For example, when funders value a particular treatment for SUD patients that has been perceived historically as providing a public good, such as a 12-step approach versus new MAT perceived as providing a private good, innovation adoption may be repressed.

In a normative theory of nonprofit finance, the nature of goods and services provided is associated with benefits that influence who will pay for them (Young, 2007), with the organizational income portfolio associated with goods and services provided mediated by strategic adjustments of organizations. In a study of nonprofit organizations in subcategories of the National Taxonomy of Exempt Entities (NTEE) of the arts, health, and human services the proportion of revenues derived from earned income or fees for services to specific individuals is lowest for services deemed public benefits, highest for those deemed private, and in the middle for those that have mixed benefits (Fischer, Wilsker &Young, 2010).

Consequently, it may be that SUD treatment organizations receiving greater portions of their funding from sources that ordinarily provide private benefits, as fees for services to specific individuals, are less hesitant and more willing to embrace change than those that provide block resources as part of its mission for societal good. Therefore,

Hypothesis 2: Across SUD organizations, there will be a positive relationship between the extent of the adoption of MAT and proportion of funding providing private, fee for services, benefits to specific individuals.

Methods

Hypothesis 1 examines the education of the executive director and its relationship to an organizational level dependent variable, adoption of evidence based MAT, reflecting diffusion of these innovations throughout the SUD treatment field. Hypothesis 2 examines the relationship between an organizational level dependent variable, the percent of earned income from fees and services that provide private benefits to specific individuals. The hypotheses are tested using data from 555 SUD treatment centers collected over an 18-month period from late 2002 through early 2004 in the United States (US) as part of a larger National Treatment Center Study. Even though there are several waves of data collected in the larger project, different questions were asked in the different waves. The fourth wave of data is used for these analyses because it includes variables on the range of center ownership and funding sources needed to test our hypotheses.

The SUD treatment centers used in the study were drawn from a two-stage stratified random sample of geographic areas throughout the U.S. Counties in the U.S. were allocated to strata based on population, then sampled within strata. The total population of treatment centers for the selected counties was compiled from a variety of sources including federal and state directories, yellow page listings, employee assistance program referral directories, survey sample call lists, and other available sources. These centers are facilities offering treatment for alcohol and drug problems with a level of care at least equivalent to structured outpatient programming as defined by the American Society of Addictive Medicine (ASAM). Since our program population is defined as community based treatment programs where all members of the community are eligible to seek services, our sample does not generalize to Veterans Administration programs that limit their access to veterans (Harris, Humphreys, Bowe & Kivlahan, 2009) or criminal justice programs that limit access to prisoners (Taxman, Perdoni & Harrison, 2007) that are outside the scope of our sampling frame.

On-site interviews were conducted with 714 individuals leading treatment centers (84% response rate). The analysis of hypotheses was conducted using complete responses for the variables identified below, providing a minimum sample of 555 centers (77% of those interviewed, and 68% of organizations in the sampling frame). Logistic regression analysis of whether a center was excluded from the analysis because of list wise deletion of partially missing information, or included because all study variables were complete was used to assess whether the linear combination of independent and control variables included in our analyses affected our results. None of the variables was significantly related to inclusion or exclusion, suggesting that the cases that are missing are missing at random and are not biasing our results (Goodman and Blum, 1996).

Measures

Adoption of Medication-Assisted Treatment

This variable measures whether each of 4 MATs is used as a treatment option at the treatment center. The respondents were asked if each of the 4 “evidence based” medications – Disulfiram (widely known as Antabuse), Naltrexone, Buprenorphrine, and Selective Serotonin Reuptake Inhibitors, were used to treat clients in their center. These medications have been shown in research studies to be effective and have been accepted as evidence based practices, but have not crossed the chasm to be adopted by a large number of physicians or SUD treatment programs (Knudsen, Abraham & Roman, 2011; Roman, Abraham, & Knudsen, 2011). If the center used the treatment for patients in the categories included in the approval provided by the Food and Drug Administration (FDA), the response was coded as a “1.” If the center did not use the treatment, the response was coded as “0.” When factor-analyzed, the four treatment variables for the centers in this study loaded onto one factor which had an eigenvalue of 1.92. All four treatment variables were also significantly correlated with each other at the significance level of .01. We calculated a MAT technological adoption index for each center by summing the response codes (0 or 1) for the four treatments for each center. The range of the extent of adoption is 0-4, with relatively low extensiveness of adoption indicated by a mean of 1.07. Forty-two percent of the centers used none of the MATs, while 28.6% used 1, 15.9% adopted 2, 11.2% adopted 3, and 1.4% adopted all 4 of them.

The FDA approved Disulfiram for use in the treatment of alcoholism in 1951. Despite the relatively early date of its approval for use, the nature of Disulfiram's action has never become widely accepted by treatment specialists nor has it been accepted within the normative culture of SUD treatment organizations, although it is an evidence-based practice and recommended by the National Institute for Drug Abuse (NIDA).

Naltrexone has been found effective in the treatment of alcohol dependence and for opiate dependence, and has been approved for use by the FDA since 1994. Selective Serotonin Reuptake Inhibitors (SSRIs) have been approved by the FDA for the treatment of depression since the late 1980s. Their use in SUD treatment is for co-occurring psychiatric problems. Buprenorphine was approved by the FDA for use in the treatment of opioid dependence in 2002. The data for this study were collected relatively soon after the drug's FDA approval. Disulfiram was used in 25.7% of the centers, while naltrexone was used in 21.7%, SSRIs were used in 48.4% and buprenorphine was used in 7% of the treatment organizations.

Clinically Relevant Education

This characteristic of the SUD treatment center leader is an indicator of one form of formal socialization. It is measured as a combination of the highest degree held and the field of study. Respondents were given 5 categories from which to choose. The five categories provided were coded as follows: “1”- less than Bachelor's; “2” - Bachelor's; “3”-Master's; “4” – Doctorate; “5” - M.D. Five educational category variables were created to distinguish the types as well as the level of education. These five variables are labeled 1) Bachelor's or less, 2) Masters in Business, 3) Master's in Social Work and Master's in Psychology/Counseling, 4) Master's in other fields, and 5) Ph.D. or M.D. For this analysis, the Master's in Social Work/ Masters in Psychology/Counseling, nursing, and MD or Ph.D., typically degrees that include formal education and internships, are considered treatment relevant degrees and were coded “1” and the variable was named “Clinical Degree”, while the remaining categories were coded “0”. Those with Bachelor's degrees or less, or advanced degrees in a field unrelated to clinical practice are coded “0”. Forty-five percent of the executive directors have clinically relevant socialization.

Public/Private Funding

To measure the organizational context of resource dependency, we examined the funding sources for these SUD treatment centers. We measured the proportion of public funding received by the SUD treatment center as follows: Funding from the public criminal justice system; funding from federal block grants; other funding from each of federal, state and county sources. The proportion of funding that provides fees for services for individuals, considered private benefit funding, include: Funding from private or commercial insurers; payments for individuals from Medicare and from Medicaid; payments by Health Maintenance Organizations (HMOs); and client payments. This information was provided in the on-site interview and verified by coders. The proportion for each category was calculated with the total organizational revenues for the previous year as the denominator, and the revenue attributed to the category as the numerator. On average, 93% of the center revenues come from the sources listed, with 48% coming from public sources and 52% coming from private or “voucher” based, fees for services, sources.

Eight variables are used as controls in the analysis.

Hospital Subsidiary

Because organizations that are parts of larger organizations can benefit from resource exchanges with the parent organization, this variable was used as a control. A cushion of resources could have a positive effect on innovative activity (Cyert and March, 1963). Therefore, if the treatment center is a part of a hospital or located on a hospital campus, it is coded as 1, while organizations that are not owned by hospitals or located in hospitals are coded as 0.

Size

Because the size of an organization may explain the absence or presence of innovative activity (Baldridge and Burnham, 1975; Bantel and Jackson, 1989), this variable is used as a control. Organization size was measured by the number of full-time equivalent employees the year before the on-site interview was conducted, and was transformed by its logarithm.

Director Experience

Experience has been found to have both positive and negative relationships with innovation. The experience of the substance abuse center director is measured as number of years in the field. Data were gathered from responses to a query regarding the total number of years the respondent worked in the substance abuse treatment field, and was transformed by its logarithm.

Opiate Treatment

Two of the medications have been used for those who have addictions to opiates. To control for the use of these medications for this purpose, we created a dummy variable coding “1” for centers that treat clients with opiate addictions and “0” for centers that do not treat clients with opiate addictions.

Dual Diagnosis Treatment

Clients diagnosed with an addiction to alcohol or drugs and depression can be treated with Selective Serotonin Reuptake Inhibitors (SSRIs). To control for the use of this medication for this specific purpose, we created a dummy variable coding “1” for centers that handle patients with dual diagnoses and “0” for centers that do not treat patients with dual diagnoses.

Physician Staff/Contract

To control for access to prescriptions for medications being an explanation for greater use of MATs, we created a dummy variable coding “1” for centers with availability of a staff or contract physician, and “0” for those centers that do not have access to prescribing professionals.

Ownership

To control for ownership being an explanation for greater use of MAT, we created two dummy variables. Government-owned centers are coded “1”and “0” is coded for those that have other ownership. Centers operating for profit were coded “1” and others “0”. The residual, comparison, category is that of organizations that are non-profit organizations.

Analysis and Results

Descriptive statistics of the data tell us that the average leader of a SUD treatment center in the data set has obtained either a Bachelor's or Master's degree (analyzed prior to creating the education category variables) and has worked in the healthcare field for 19 years. He or she has a staff of approximately 34 full time employees and at the time of the interview averages one technological innovation or MAT adoption out of a possible four queried (see Table 1). Among the predictor and outcome variables, correlation analysis indicates significant positive correlations between leaders having clinical degrees and adoption of MAT. There is also a positive and significant correlation between the proportion of revenues from sources of payments for treatment of specific individuals and MAT innovation adoption at the organizational level.

TABLE 1. Means, Standard Deviations and Correlations.

	M	SD	1	2	3	4	5	6	7	8	9	10
#MATs	1.00	1.08	1
Hospital	.21	.41	.27	1
Size (lg)	2.83	1.16	.22	.04	1
Director Experience eBHC(lg)	2.83	.54	.15	.05	.10	1
Physician Staff/Contract	.79	.41	.19	.04	.30	.01	1
Opiate Treatment	.88	.33	.15	.09	.15	.01	.16	1
Dual Diagnosis Treatment	.74	.44	.20	.08	.14	.00	.16	.04	1
Government-Owned	.14	.34	-.03	-.10	.10	.05	.10	.10	.00	1
For-Profit	.15	.36	.07	-.03	-.10	-.01	-.10	-.05	-.03	-.15	1
Clinical Degree	.48	.50	.19	.02	.10	.17	.07	-.03	.10	.07	-.04	1
Proportion Private Funds: Fees for Individual	.54	.39	.32	.17	.11	.05	.09	-.05	.08	-.12	.01	.02

N=555 r>.09<.11=p<.05; r>.11=p<.01

Ordinal regression, using the Polytomous Universal Model (PLUM), with a link Negative Log-Log, was utilized to analyze the data and test the hypotheses. Table 2, column 1, shows that having a clinical degree is significantly related to MAT with a regression estimate of .502 and standard error of .12, p< .01. Therefore, Hypothesis 1 is supported.

TABLE 2. Ordinal Regression-Polytomous Universal Model (PLUM): Link FunctionNegative Log-Log. DependentVariable=Index Count (0-4) Medication-Assisted Treatment (MAT).

	Col 1	Col 2	Col 3	Col 4
Control Variables	Estimate	Estimate	Estimate	Estimate
Size	239**	.249**	.233**	.251**
Hospital	0.273	0.05	0.26	0.032
Director Tenure BHC	.270*	.285*	0.206	0.237
Opiate Treatment	417*	0.284	.486*	0.325
Dual Diagnosis Treatment	.544**	.546**	.404*	.530**
Physician Staff/Contract	.419*	.347**	0.181	0.327
Government-Owned	0.341	.382*	0.198	0.196
For-Profit	0.026	0.198	-0.025	0.216
Independent Variables
Clinical Degree	.502**	.415**	.538**	.406**
Proportion Private Funds:
Fees for Service	1.103**	---	---	---
Components of Private Funds
Private Insurance		1.945**		2.190**
Medicaid		.655*		0.851
Medicare		1.683**		1.874*
HMO		1.973**		2.194**
Client Pay		.693*		1.064
Components of Public Funds
Criminal Justice			-3.936**	-2.254*
Federal Block			- .681*	0.707
Federal Funds			-1.377**	-0.041
State			-.751*	0.673
County			-1.652**	-0.288
Cox Snell Pseudo R-sq	0.26	0.3	0.29	0.32
Hosmer-Lemeshow GOF	19.8 (p=.71)	12.89 (p=.97)	17.31 (p=.84)	12.89 (p=.97)

^*p<.05

^**p>.05<.01

The analysis described in Table 2, column 1, also shows that there is a significant and positive relationship between private benefit funding and adoption of MAT with an ordered logistic regression coefficient estimate of 1.103 and a standard error of .21, p < .01. When analyzing the relationship between specific private benefit sources of funding and adoption of MAT, we found that private or commercial insurance, Medicaid, Medicare, and Health Maintenance Organization (HMO) funds for specific individuals, and client payments had positive and significant associations with MAT adoption (column 2).

There was a converse negative and significant relationship between centers receiving higher percentages from public funds and the extent of adoption of medications, with the analysis of the specific sources of funds coming from the public criminal justice system, federal block grants, and federal, state and county funds (column 3) all negatively related to MAT adoption. However, when specific sources of public funds were analyzed along with specific sources of private benefit, funds from the criminal justice system were negatively and significantly related to innovation adoption (column 4), while the remaining sources of public funding were not significant. The private and commercial insurance, Medicare and HMO funding remained positively associated with MAT adoption at the organizational level. Therefore, Hypothesis 2 is supported.

Discussion

A convergence of data across the for-profit, not-for-profit and government sectors (Gowdy, Hildebrand, LaPiana & Campos, 2009: Young, Salamon & Grinsfelder, 2012) is evident in that ownership is not significantly related to adoption of MAT innovations in SUD treatment centers. It is the source of funding and not the legal form of organizational ownership that appears to make a difference. Further, analyses disaggregated by type of education show that each clinically relevant education category is significantly different from the omitted categories of BA/BS Degrees or less and advanced degrees in a non-clinical discipline.

In addition, sensitivity analyses were performed using logistic regression analysis for each type of MAT as a dependent variable. The results show the same pattern of relationships for clinical education and public funding for each of the adoption of SSRIs, Naltrexone and Disulfiram. They are more likely to be adopted in SUD treatment centers that have executive directors with clinically relevant educational socialization, and more likely to be adopted in SUD treatment centers that have higher proportions of their funding coming from private market based (fees for service) benefit sources. The relationship for the adoption of Buprenorphrine, the newest evidence based MAT with the lowest base rate of adoption at the time of the data collection, differs from the other three. Buprenorphine reaches statistical significance with clinically relevant education, but its relationship with adoption is not explained by funding source or by type of ownership. As with the multinomial regression models that tested our hypotheses, the Hosmer-Lemeshow goodness-of-fit statistics for each of the binary logistic regressions support our models.

This analysis was designed to test hypotheses concerning the extensiveness of MAT adoption in a social entrepreneurship context, proposing that extensiveness of innovation is affected by the fit between the socialization of leadership and the innovation, as well as by the fit between environmental contingencies (such as funding source) and MAT.

This study examined the professional socialization of leaders of SUD treatment centers by focusing on the relevance of formal education and its relationship to a proclivity to adopt MAT for patients. We proposed a hypothesis that posited positive relationships between the observable characteristics of SUD treatment center leaders and adoption, at the organizational level, of MAT innovations. While these characteristics serve as proxies for values derived during socialization processes, professional identity and social relations, these assumptions are supported by social categorization theory where identity construction is an outcome of cognitive processes (Oakes, Haslam & Turner, 1994).

For Hypothesis 1, our analysis shows that obtaining a clinical degree is significantly and positively related to the organizational level decision to adopt medications. Those organizations where leaders received treatment-relevant professional education and socialization were more likely to utilize MAT. Thus, those organizational leaders who experience these types of professional socialization may be more able to recognize the value of maximizing evidence based treatments for their clients, thereby more easily implementing institutionally contested pharmaceutical treatment into SUD treatment protocols. The complexity of the terrain is demonstrated by the non-exclusive reasons provided by the non-adopters of a particular MAT: 64% of those who do not use Disulfiram/Antabuse claim that it is inconsistent with their center's philosophy, and 27% of them claim they do not use it because of its cost. For Buprenorphrine, the comparative percentages are 47% and 25%, for Naltrexone, 50% and 33%, and for SSRIs, 41% and 21%. Because pharmaceutical treatment can be expensive and is not necessarily reimbursed by insurers, adoption can result in treating fewer clients with available resources than traditional behavioral treatments, or treating patients differentially based on access to individualized payments. However, it could be a paradigm shift in treatment effectiveness for some clients. This is likely to be especially true in fields like SUD treatment where behavioral change is very difficult to sustain and relapse is a prominent concern.

It is also important to note that advanced degrees in business or other disciplines without formal socialization in a counseling discipline were not statistically significant predictors of decisions reflective of adoption of MAT, suggesting the importance of the content of professional experiences and the values such content can impart. Thus, it appears that the disciplinary area of study, rather than the level of education, is associated with social capital that drive our results. The support for Hypotheses 1 is evidence that values inherent in socialization experiences of these leaders of SUD treatment organizations can predict managerially strategic adjustments, including the risky organizational level decision-making behavior concerning the adoption of institutionally contested medications. These results are consistent with the proposition that alignment between organizational characteristics such as leader socialization and the innovation itself can lead to more extensive diffusion of innovation (Ansari, et al. 2010) as well as the links between learning and identity which are built through participation and social contact in communities of professional practice (Ferlie, Fitzgerald, Wood & Hawkins, 2005). The fidelity of these adoptions and the extensiveness of these may change over time, as the institutional arena becomes less contested, but social and cognitive barriers that mutually reinforce each other would have to be overcome through social interaction, trust and motivation.

Hypothesis 2 posits that the adoption of innovative medications will be more likely in organizations where a funding model based on fees for service to individuals is more prevalent relative to other sources. The significant and positive relationship found between the proportion of private source funds received by SUD treatment centers and adoption of medications is also evidence for support of Murphy and Coombes' (2009) argument that the values of the environmental context are important considerations in the process of social sector opportunity discovery when it includes the decision to act upon the opportunity. MAT availability appears to be associated with funding sources that represent public, private and mixed benefits.

This finding could explain why more centers have not acted upon the opportunity to adopt evidence based MAT. It may well be that in organizations where higher portions of funding derive from public sources the value of medications is perceived differently from those receiving greater portions from private benefit, fees for service, funding. While public goods benefit everyone and are funded by public monies, private goods are consumable and are funded through private sources and may be perceived to benefit only members of a specific group (Kingma, 1997). Also, this finding supports the theory that there is a relationship between the type of services provided and type of revenue stream (Fischer, Wilsker & Young, 2010), with “private” services such as the provision of individual medications to clients being more positively related to the private benefit proportion of funding than to the public proportion. Furthermore, the apparent lack of fit between centers with higher proportions of public funding sources and MAT supports the proposition that lack of fit is related to low innovation diffusion (Ansari, et al. 2010).

It is important to notice the consistent negative relationship of adoption with funding from criminal justice sources across the several equations. The finding of a negative association with this source of funding is not novel. Data from an older national sample of SUD treatment programs reported reliance on criminal justice referrals to be negatively associated with availability of psychotropic medications (Knudsen, Ducharme & Roman, 2007) in SUD treatment programs. Medication use in SUD treatment is relatively low within services delivered as part of services provided within correctional institutions (Smith-Rohrberg, Bruce, & Altice, 2004; Rich, Boutwell, Shield et al., 2005) as well as in treatments that are associated with criminal justice referrals. Kubiak, Arfken, and Gibson (2009) found that less than one-third of Departments of Corrections' administrators reported that medications are provided to offenders receiving community-based SUD treatment services.

Here an explanation may rest on a collision of different paradigms, those governing the delivery of criminal justice and those governing the delivery of medical care. While many professionals and other actors in the criminal justice system have become convinced of the value of “diversion” to treatment as an alternative to imprisonment for non-violent offenders, it is likely that this support is within the context of their guiding paradigm. Specifically, the traditional SUD treatment centered on 12-step philosophy has long been attractive to many in the criminal justice community, likely because it rests upon self-investment, self-denial, conformity to a strict set of rules, and passage through a clearly described set of numbered “steps.” This contrasts with the image of passive and perhaps even indulgent treatment of the patient through the administration of medications to curb his/her desires or “craving” for the use of psychoactive substances. This inconsistency may explain the distinctive diminution of adoption of MAT by SUD treatment programs that receive greater portions of their funding from criminal justice sources.

The results also demonstrate that the form of organizational ownership is not a predictor of MAT adoption - an indicator of convergence across sectors, when resource funding is considered. There appears to be variation within different types of ownership or organizational forms. As suggested by Bozeman and Straussman (1983: p 86), it appears that innovation can be easily stifled when effectiveness is sacrificed to narrower efficiency norms. When funding is unpredictable and inadequate for meeting the demand for services, early adoption is more risky and less likely. In resource constrained contexts, efficiency of traditional treatment methods that allow for treatment of larger numbers of people, perhaps less effectively, may be preferred. However, effectiveness is subordinated to efficiency goals, a lack of adoption of evidence based treatments could be dysfunctional for both the organization and those served. The control that is perceived to be exerted by funding sources can influence resource strategies and appears to be a better predictor of the adoption of novel treatment combinations in SUD treatment centers than the legal ownership status of the organizations.

Limitations

This study has several limitations. One limitation of this study is missing data. Because of missing data on some variables, cases were eliminated from analysis through list wise case deletion. However, given the large number of centers sampled, there were enough cases to have adequate power for this study. Also, logistic regression analysis of whether a case was observed or missing was not predicted by the linear combination of independent and control variables included in our analyses, suggesting that the cases that are missing at random are not biasing our results (Goodman and Blum, 1996).

Since the sample was drawn from a population of community treatment programs where all members of the community are eligible for services, the findings are limited in their generalizability beyond these parameters. Importantly the SUD treatment programs that are part of the Department of Veterans Administration are not included in the sampling frame. These organizations have very different treatment and policy environments than the community based treatment organizations in our sample, our inability to include these organizations in our analyses may affect the generalizability of our findings and could therefore be considered a limitation.

We used education level and field of study as an indicator of formal socialization. Educational socialization is a process through which professional identity, an aspect of personal and social identity that develops in professional personnel as a result of their work activities, and social capital, productive benefits derived from social relations, emerge. While the operationalization is consistent with social categorization theory (Oakes, Haslam & Turner, 1994), and the processes documented to be associated with identity development (Pratt, Rockmann & Kaumann, 2006), more direct measures of the variation in professional identity and social capital can be used in future studies to provide more nuanced understanding of their impact on decision making, adoption of new practices, and organizational adaptation.

The data for this analysis were collected before parity legislation for behavioral health with physical health care was implemented. It is beyond the scope of our analysis to consider the variation across states in parity legislation or the potential effects of The Affordable Care Act. However, based on the experience of Massachusetts, which implemented its own universal insurance coverage in 2007 incorporating substance abuse services into benefits to be provided to all residents, coverage alone is insufficient to increase treatment use (Capoccia, Grazier, Toal, Ford II, & Gustason, 2012).

Implications

These findings contribute to the literature by offering evidence for the importance of identifying the environmental context of social purpose organizations to help explain organizational decisions concerning the adoption of innovations that challenge a prevailing ideology. It also suggests the importance of the socialization of leaders of SUD enterprises whose educational experiences are related to whether their social enterprises act upon opportunities to address social needs through adoption of ideologically contested MAT. In addition, MAT adoption in the resource constrained environments of SUD treatment organizations appear to be more limited in organizational contexts with greater dependence on public funding sources as compared to those who have larger portions of their revenues derived from commercial (fees for service) sources. It appears that centers that depend more on public block sources of funding, rather than fees for services for individuals, avoid, consciously or unconsciously, risky behaviors, thus reducing the adoption of innovations that can change treatment paradigms and effectiveness. These findings support propositions that sources of income for social mission-oriented organizations tend to be aligned with the perceived nature of benefits provided to clients (Young, 2007). That is, sources of private (fees for services) funding are more aligned with benefits for individuals and sources of public funding are more aligned with collective type of benefits. These findings can also assist government and policy making organizations to address the social and economic factors that influence the adoption of innovations across SUD treatment organizations across other social mission sectors. The competitive and legitimacy challenges faced by treatment centers, in addition to the fiscal changes on the horizon, create uncertainty and the imperative for adaptation. While treatment organizations appear to be responding to market pressures, there appears to be patterned differences among organizations (Young, Salamon & Grinsfelder, 2012) in their adoption of MATs. Uncovering patterns across organizations can help identify the risks and opportunities in the diverse health care sector with its many subfields or specialty organizations. Even when there is legislation for universal coverage, variation will likely exist among treatment organizations that will need to be understood to effectively serve behavioral health care needs.