Dissemination of evidence‐based practices: how far we've come, and how much further we've got to go

There is ample evidence that in addiction treatment, as in most areas in health care, there remains much work to be done before the best evidence from clinical research is integrated to provide the best clinical practice. A shift in focus to clinically significant patient outcomes may be the most efficient strategy to encourage implementation of the evidence-based practices (EVPs) that are most likely to improve practice and patient outcome. When I began my training as the first pre-doctoral fellow in the Division of Substance Abuse at the Yale Department of Psychiatry in 1982, the landscape of addiction treatment in the United States was quite meager. Available medications were limited to disulfiram for alcohol, methadone/LAAM for opioids, with occasional use of clonidine for detoxification. Naltrexone for opioid dependence was a novel approach used typically only in highly specialized university-affiliated programs. Non-pharmacological approaches were largely of two forms: the first was long in-patient or residential stays, characterized by multiple daily groups and, frequently, highly confrontational. The second approach was out-patient counseling rooted in a self-help framework, where former drug- or alcohol-dependent individuals provided group counseling organized around encouraging enduring affiliation with self-help organizations. The only ‘manual’ in wide circulation was the Big Book of AA. The supervision I received on clinical cases was firmly grounded in psychoanalytical or object relations theory. There were very few outcome studies of well-defined behavioral therapies, as the landmark Woody 1 and Rounsaville 2 studies, the first randomized trials of manual guided therapies in the addictions, would not appear until 1983. Thus the ‘black box’ that was addiction treatment could not have been much blacker. Addiction treatment is much changed, largely for the better. There are now a wide range of EVPs available for a broad range of addictions, with contingency management, cognitive–behavioural (CBT), motivational interviewing (MI) and structured family therapies the most prominently cited among meta-analyses, systematic reviews and treatment guidelines 3-5. Manuals and training resources for most EVPs are available and easily accessible via the internet. Use of EVPs in larger practices and treatment systems is now a virtual mandate. Most clinicians profess that they routinely implement at least one EVP. In the United States there have been coordinated, ambitious efforts to train very large numbers of clinicians in EVPs and to implement them across the nation's largest health-care system, the Veterans Administration 6, 7. There has also been a surge in the availability of approved medications, most notably buprenorphine, naltrexone and acamprosate. Particularly exciting are promising new pharmacotherapeutic strategies (e.g. vaccines and cognitive enhancing agents) as well as novel mechanisms to promote adherence to medications where non-compliance has undermined their impact and dissemination (e.g. depot formulations, implants). These efforts are significant and admirable, but multiple lines of evidence also indicate that EVPs are far from universally available and competently implemented. For example, a recent national survey in the United States indicated that most treatment programs do not offer training or supervision in EVPs, nor do they require their newly hired clinicians to have training in EVPs 8. In our own work, we evaluated ‘treatment as usual’ as practiced in 20 sites across the United States in a recent Clinical Trials Network study of motivational approaches. Prior to implementation of the study, clinicians reported that they made extensive use of EVPs such as CBT and 12-Step facilitation (TSF) 9 However, when more than 700 audio-taped ‘treatment as usual’ sessions were evaluated by independent raters, interventions associated with those and other EVTs were virtually undetectable 10-12. In fact, therapist-initiated discussion of issues clearly unrelated to any patient problem or issue (e.g. ‘chat’) was seen more frequently than any EVP 13, 14. There is growing consensus that monitoring, supervision and feedback are needed until the clinician can demonstrate adequate fidelity and skill, as randomized training studies of clinicians have highlighted that workshops are necessary, but clearly insufficient in teaching clinicians to implement evidence-based treatments effectively 15. Regarding utilization of new medications for addiction, utilization of approved medication is growing in most countries, but remains several-fold smaller than the number of individuals who might benefit from them 16 Insurance and third-party support are uneven for new, proven treatments, but remains constant and plentiful for some therapies that have been demonstrated many times to be both expensive and ineffective (e.g. opioid detoxification without aftercare). Upon reflection, my own recommendations regarding strategies for dissemination have not changed greatly since Bruce Rounsaville and I wrote our piece on ‘A vision of the next generation of behavioral therapies research in the addictions’ in Addiction 5 years ago 17: First and most fundamentally, more research is needed that verifies the superior efficacy and cost-effectiveness of ESTs over standard practice. Second, when multiple ESTs are available, studies with adaptive designs are needed to guide the choice of frontline treatments and to identify optimal sequences of treatments to deliver when initial results are poor. Third, a new subfield of ‘training effectiveness’ research is needed to address a series of questions about treatment dissemination that parallel those articulated by Paul . . . ‘which training is needed in which therapy for which trainee to treat which type of patient?’. 18 Fourth, while behavioral treatments resist formulation and delivery with the precision of FDA-approved medications, computer-assisted therapist training and computer-delivered treatment hold considerable promise for bringing the efficiency and precision of EST dissemination to new levels. More of the type of research suggested above is beginning to appear, and there are many studies of training and implementation ongoing world-wide 15. At the same time, what may be getting lost in these efforts is attention to the original aim of dissemination of EVPs: to improve patient outcome. In my view, the point of training providers in EVPs is not to have them simply go through the motions of using motivational language or teaching coping skills but to use these techniques, and others, to produce the most successful outcomes in the largest number of cases possible. What may be needed, both in clinical practice and clinical research, is a radical shift to focus on reliable, clinically significant outcomes and performance indicators. Currently, the drug abuse field lacks consensus as to what constitutes clinically meaningful outcome; that is, clear, quantifiable indictors of a successful treatment outcome 19. This is a highly complex question, and one that has thus far eluded drug addiction. It is clear that retention alone is an imperfect performance indicator (i.e. compliance with ineffective approaches tends to yield disappointing results). Lack of widely accepted benchmarks in the addictions is in sharp distinction to emerging standards in most areas of medicine where, for example, patients contemplating a knee replacement or a heart procedure can access aggregate information regarding success rates for various procedures and clinics prior to making a decision. We are far from a system that allows individuals entering addiction treatment facilities to make informed decisions regarding probable outcomes. This endeavor will doubtless be extremely complex, given the variability in the treatment systems throughout the world and the complex, multi-problem nature of the addicted population. However challenging, development of consensus performance indicators would be likely to spur significant clinical and research innovation and novel efforts to improve treatment outcome. For example, a recent study by McLellan and colleagues on the effects of performance-based contracting on a state drug abuse treatment system indicated that systematic incentives for performance can improve patient outcomes dramatically in a wide variety of clinics 20. In that study, it was notable that the most successful clinics spontaneously made a number of changes to make treatment more accessible and inviting, and all learned at least one EVT 20. A shift in the field to a more consistent focus upon clinically significant outcomes could have could have a number of effects. Clinical trials of novel medications or behavioral therapies would have a clear benchmark, or denominator, against which to compare results. Adoption of new practices could be based on demonstrated feasibility, efficiency and cost savings. Treatment centers, or clinicians, that have not embraced specific EVPs would be free to continue current practice, provided they could demonstrate that their outcomes are commensurate with established outcome standards. With rapid changes in health care delivery world-wide, our continued failure to define, measure and report on a common set of clinically significant outcomes and thus to demonstrate that our treatments produce meaningful benefit puts our field in peril. Support was provided by National Institute on Drug Abuse grants P50-DA09241, R01-DA 015969, U10 DA13038.