These data from nationally representative samples of American substance abuse treatment centers in the public and private sectors revealed a modest degree of medication adoption. Centers were considerably more likely to report having adopted SSRI medications than pharmacotherapies specifically approved for the treatment of substance abuse. Almost half of centers currently use SSRIs to treat patients, and centers offering this type of medication were more likely to report to have adopted the other medications. It may be the case that SSRI adoption represents a gateway for greater medication adoption in American treatment centers. Support for this argument was recently demonstrated by Fuller et al’s (2006) longitudinal analysis of naltrexone adoption, where the adoption of SSRIs was positively associated with subsequent addition of naltrexone to the center’s treatment programming.
These data suggest that the availability of a given medication is more likely when other medications are available, a finding consistent with Koch et al.’s (2006)
recent work. The correlations between treatment medications suggest that there may be value in considering patterns of availability between medications. However, there was also evidence that the organizational correlates of medication availability is, in part, shaped by the type of medication; this finding is consistent with Rogers (1995)
who argued that innovation adoption is shaped by the fit between an innovation’s characteristics and the organizational context. In the case of substance abuse treatment pharmacotherapies, there was some indication that medications may be differentiated by the extent to which they are regulated. In the US, there are additional requirements that must be met in order for treatment centers to offer agonist medications in contrast to other pharmacotherapies which any physician may prescribe. The logistic regression models of medication availability suggested similarities in organizational correlates between the three less-intensely regulated medications; a model treating these medications as a single latent variable fit the data well. For agonist medications, there were certain organizational correlates that were unique to this class of medications, suggesting the need to consider these medications as distinct from the less-intensely regulated pharmacotherapies.
The relevance of the regulatory environment comes into sharper focus in the final model that simultaneously estimated the correlates of the less-intensely regulated (LIR) medications and agonist medications. In the US, the regulation of agonist medications has shifted to an accreditation-based system. Consistent with this aspect of the regulatory environment, accredited centers were significantly more likely to offer these medications when other organizational characteristics were controlled. Although center type and organizational affiliation were associated with agonist medication availability at the bivariate level, these measures were no longer significant once accreditation and other organizational characteristics were taken into account. In contrast, accreditation was associated with the three LIR medications at the bivariate level, but was no longer significantly associated with the LIR medications once other organizational variables were controlled. Center type, which integrated data on ownership, profit status, and predominant source of funding, was associated with the LIR medications, as was the measure of organizational affiliation. This suggests that accreditation is less relevant to the adoption of LIR medications because these medications are not subject to the same regulations.
The finding that center type is associated with LIR medication availability in American treatment centers suggests that there are disparities in the quality of care received by clients based on the types of organizations from which they seek care. In particular, non-profit centers highly reliant on public funding appear to lag behind other types of treatment facilities in terms of offering these medications, even after controlling for other organizational characteristics such as medical resources. However, it is precisely this dependence on governmental funding that may serve as a lever through which policy changes can facilitate medication adoption. These organizations are accountable to state governments for the services they deliver in a way that privately funded programs are not; if funding is linked to the delivery of evidence-based substance abuse treatment services, such as medications, this difference in medication availability may begin to narrow (Capoccia, 2006
These results begin to suggest the types of American treatment organizations where medications are more likely to be used. Medical resources, as would be expected, appear to be a critical issue. Centers offering detoxification services, indicative of a greater ability to medically manage addiction, were more likely to offer medications. However, these services are not highly prevalent in American specialty substance abuse centers, particularly in the public sector. In addition, the presence of at least one staff physician was associated with both types of medications. Publicly funded non-profit organizations were the least likely to have this type of physician access. It appears that there are also substantial structural barriers to the adoption of medications in this segment of the American treatment system. Given that publicly funded nonprofits represent the largest segment of the overall treatment system (Horgan and Levine, 1998
), these structural differences are particularly significant to the overall process of moving evidence-based medications into routine practice.
Interestingly, there were mixed results for the measure of treatment philosophy. At the bivariate level, 12-step programs were more likely to offer naltrexone, but this association was no longer significant once other organizational factors were controlled. In the multivariate model of disulfiram availability, 12-step centers were less likely to offer this medication; there was a similar trend for SSRI availability. However, in the structural equation model where LIR medications were treated as a single latent variable, this measure of treatment philosophy was not significantly associated with either LIR or agonist medication availability. It appears that structural features of organizations, such as accreditation, center type, and physician resources, may be more relevant than treatment philosophy in terms of medication availability.
There are several limitations in these analyses due to the research design. First, the data are cross-sectional, which limits the ability to test causal relationships. The analyses address a limited number of treatment medications, so it is unclear whether these findings would generalize to other medications, such as acamprosate, which recently received FDA-approval in the US, or sedative hypnotics. Additionally, these analyses were only concerned with adoption of medications, and therefore, cannot speak to the issue of implementation. Consideration of how routinely these medications are prescribed and the proportion of eligible clients receiving them are important directions for future research.
In addition to limitations related to measurement, there are other limitations related to the types of organizations studied. While the samples are representative of the majority of specialty substance abuse treatment facilities operating in the US, it is unknown whether and to what extent these findings may generalize to VA settings, correctional facilities, opioid treatment programs, and clinicians in private practice, who were excluded from the research design.
It is also unclear the extent to which these findings would generalize to other national contexts. Some findings, such as the association between access to physicians and medication availability, would probably remain relevant in other countries. However, other findings are perhaps context-specific due to the influence of US policies. For example, the association between accreditation and availability of agonist medications may reflect the policy changes that have shifted the regulation of MMT to an accreditation-based system (CSAT, 1999
; Jaffe and O’Keefe, 2003). Relatedly, the organizational affiliations of treatment providers (e.g. hospitals, community mental health centers, and freestanding clinics) and the types of centers (based on funding and ownership) also reflect the evolution of a treatment system within a particular social context (White, 1998
). Finally, the medications available for adoption are dramatically shaped by national policies. For example, emergent treatment medications, such as buprenorphine and acamprosate, were available in European markets before they received approval in the US (Bridge et al., 2003
; Kranzler, 2000
). Future research should begin to document the range of treatment services available across national contexts. Such data would allow researchers to identify factors that influence innovation adoption across contexts and those factors that are nation-specific.
In summary, these data from the NTCS suggest that there are substantial differences in medication adoption between the public and private sectors within the American treatment system. These differences are partly a function of differences in organizational characteristics. These data also suggest that the presence of staff physicians is a key resource for medication adoption, and that contractual relationships with physicians may not be sufficient for the expansion of pharmacotherapy use in the specialty treatment system in the US. While centers offering detoxification services were significantly more likely to use these medications, their technology is predisposed to embrace the adoption of pharmacotherapies, and, moreover, such centers represent a minority in the treatment system. These results also suggest that future research should continue to explore which types of predictors are consistently associated with medication adoption in general and which are only associated with specific types of medication. Finally, the patterns of relationships between medications should be considered in future research on innovation adoption in substance abuse treatment settings.