Skip to main content

Rooting Out Treatment Failures

January 23, 2020

Therapists are aware of the endless hours spent acquiring the skills associated with clinical models like cognitive behavioral therapy or psychodynamic therapy. As anyone trained in both models will tell you, they are very different clinical approaches. Yet decades of psychotherapy research have found that no established model is more effective than another. A great misunderstanding exists about outcomes.

I view mental health conditions, substance use disorders, and health behaviors as critically important domains under the umbrella of consolidated behavioral health. There are valuable therapeutic interventions within each domain. My goal here is to call attention to key findings about psychotherapy that can inform substance use disorder (SUD) and wellness treatment approaches.

The psychotherapy research literature has much to teach us, but my focus will be how to root out treatment failures. Attention will also be paid to treatment factors that contribute to success. That we can indeed isolate failure is not well known, and it is an important public health issue. A substantial body of research exists for mental health, and those findings will be applied to the other domains.

Alliance is one of the “common factors” across therapy models. The research literature, well summarized by Wampold in “The Great Psychotherapy Debate,” shows therapeutic alliance may be more important to clinical outcome than allegiance to clinical protocols. We will take a brief look at this literature, examine therapist effects, and conclude with the value of feedback in the treatment process.

The therapeutic alliance

While psychotherapy research has not found one model to be superior, it has identified factors such as therapeutic alliance and empathy as critically important in achieving a positive outcome. Research into therapeutic approaches for alcohol use disorders has similarly found no superior treatment model. However, before looking deeper into this research, let’s be clear about different forms of treatment.

Our North Star in SUD treatment should be a biopsychosocial orientation. This includes everything from medication-assisted treatment to motivational interviewing to peer support and vocational rehabilitation. The focus on psychological approaches here is not intended to minimize the importance of the other treatment dimensions.

How should we understand a therapeutic alliance? Research suggests that it is characterized by the therapist and patient agreeing on the goals and tasks for therapy. This factor is noteworthy for stressing how much people matter. Clinical protocols may tell clinicians what to do, but the results of treatment are less likely to be positive if the people engaged in it don’t share expectations.

SUD treatment is ripe for research in this area. Treatment professionals historically insisted abstinence is the only acceptable goal for treatment. Disagreement was confronted as resistance. The industry witnessed a shift in framework with the introduction of motivational interviewing (MI). MI clients were not told what goals to adopt, and conflicts about changing were explored rather than confronted.

Both 12 Step-oriented and MI approaches help people. We don’t know why distinctly different approaches are beneficial, and too much debate focuses on the techniques. It is possible some clients establish an alliance due to an easy acceptance of goals and tasks dictated by the therapist. Some may experience an alliance only when these issues are actively discussed and agreed upon together.

Therapist effects

Results vary by therapist more than by clinical protocol. As Wampold describes it, “the essence of therapy is embodied in the therapist.” This finding applies to both clinical trials and naturalistic settings. Research in psychotherapy implies we should advise loved ones to find a therapy relationship that feels right rather than selecting a model of therapy.

We have little understanding why this robust finding is so. There is some evidence that more effective therapists can establish alliances with a broad range of patients, but much more work is needed in this area. The main point is that a pool of therapists can be measured on their clinical results, and they can then be ranked from most to least effective.

The good news is that most clinicians have results with modest to large effect sizes. The most striking differences are between therapists in the upper quartile of outcomes versus those in the bottom quartile. What is one to conclude? My recommendation is to focus on clinicians in the bottom 10% of outcomes. If you want to root out treatment failures, mentor or remove those in the bottom 10%.

We currently have few programs following this protocol. However, this fact is more a matter of will than the required science or technology. We could know which therapists, SUD counselors and health coaches get the best results. Robust outcomes monitoring systems are being implemented in some quarters, and they will be able to achieve superior outcomes by rooting out many treatment failures.

Clinical feedback

When clinical outcomes are measured during treatment, it is possible for the actual change achieved by each patient to be compared with the expected amount of change based on a benchmark database. Feedback about the extent to which treatment is on track can be provided to the therapist. Psychotherapy programs like this have been in place for decades.

Michael Lambert’s program at Brigham Young University was an early example of this, and his work has defined both the science and practice in this area. He found that therapists consistently do not identify likely treatment failures. His system for providing feedback was designed to prevent people from either dropping out of care or getting poor results. His research suggests this should be the standard of care.

Measurement and feedback systems can also be established for SUD programs and for health behavior programs addressing diet, exercise, smoking and so on. Psychotherapy findings show that up to 10% of patients are at risk for a poor outcome and feedback can prevent this. This data-driven focus for each patient would supplement the usual program-level focus for SUD and wellness services.

Measurement systems can be designed to emphasize the most relevant clinical issues, and a combination of validated scales for tracking change and unique items to alert potential risk (e.g., suicidal thinking, cravings) should be considered for each program. Studies find clinicians are optimistic about their work, and a clinical feedback system is the best way to root out unforeseen treatment failures.


Psychotherapy research suggests focusing more on the people in each episode of care and less on the clinical protocols. Is there a good personal fit? Is there mutual agreement about goals and tasks? Is improvement as expected or well below norms? Is the clinician getting feedback about clients who are at risk for a poor outcome? Are clinicians with outcomes in the bottom 10% being monitored and mentored?

Consolidated behavioral health is a clinical model that combines and strengthens three currently separate domains of health: our level of emotional health, our use of addictive substances and our health-promoting behaviors. In the spirit of each domain learning from the other, this discussion has taken lessons from the well-researched domain of psychotherapy and applied them to the other two.

Ed Jones, PhD, is senior vice president for the Institute for Health and Productivity Management.

Back to Top