What works in the treatment of people with eating disorders?  Search around a bit on the internet, or consult official treatment guidelines, and you’ll find cognitive-behavioral therapy (CBT) and interpersonal therapy (IPT) listed as the “best-supported” or “preferred” psychological approaches for bulimia, anorexia, and binge eating.

Such recommendations strongly imply such approaches contain ingredients specifically remedial to eating disorders which, when applied, result in superior outcomes.  Otherwise, why create the list in the first place?

But what does the evidence actually indicate?  While research in mental health rarely results in definitive findings, in the case of eating disorders, the story is different.  When it comes to psychotherapy, all methods work equally well.   At least, that is the conclusion of the most recent, sophisticated meta-analysis on the subject.  However, if history serves as a guide, many will find the latest results hard to swallow.

Back in 2014, an article penned by proponents of the “specific treatments for specific disorders” — aka the “empirically supported” treatments movement — appeared in The Guardian, claiming science had show that some approaches were “better for certain conditions than others,” in particular eating disorders.   Citing the tremendous cost to sufferers and the healthcare system, they urged the field to “redouble … efforts to identify … and ensure that the most effective therapies are available to all who need them.”

As I blogged about at the time, I received a ton of email when that article first appeared.  “Have you seen the Guardian?” they asked.  “What do you make of it?” others inquired.   A few messages were downright snarky, even gloating,  “Scott, research has finally proven certain approaches are more effective than others.  I knew it all along!”

I responded noting that the claims in the article were based on a single study.  One.  And yes, that one study comparing CBT to psychoanalysis found CBT resulted in superior effects in the treatment of bulimia.  Crucially, I pointed out, the authors failed to mention the existence of another, exhaustive investigation available at the time in Clinical Psychology Review—one that used the statistically rigorous method of meta-analysis to review 53 studies of psychological treatments for eating disorders, and found no differences in effect between competing therapeutic approaches.

Four-and-a-half years later, the question of “what works best” in the treatment of eating disorders is being addressed in a brand new study in the top tier journal, Psychotherapy Research.  (As of right now, you can read it for yourself for free by clicking here.  Be prepared, however, as this is not an opinion piece written in a newspaper, but rather an academically rigorous analysis of the evidence).

What did the authors find?  Confirming the results of the prior meta-analysis: (1) any treatment works better than none; (2) real treatments are more effective than sham approaches; (3) and no method works better than any other.

Similar results, have been found across a wide range concerns that bring people into treatment, including trauma, sexual abuse, alcohol abuse and dependence, depression and anxiety.

Given the evidence, the question is not whether such results can be trusted.  They can.  Indeed, they represent the “state-of-the-art” — the best research has to offer.  The real problem, then as now, is that such findings do not address the question therapists most want answered, “What can I do to better help my clients?”

To answer this question, we have to recognize a simple fact: therapists live in a fundamentally different world than researchers.  We do not deal with groups of people sharing a common diagnosis who are randomized into different treatments.  Neither are we are interested in differences in the means response of aggregate group comparisons.  We deal with individuals.  Confronted daily by their suffering, we want to know how to help the person in our office right now.  The problem comes whenever these two worlds are conflated, as advocates of particular treatment approaches are prone to do.  It’s then our pragmatic focus make us exceptionally vulnerable to anyone claiming to have discovered “a better way.”

So, what can therapists do to improve their effectiveness?

Simply put: find out if what you are doing is helping your client.  Do this by seeking feedback on a formal, session-by-session basis about their progress and experience of the therapeutic relationship–a process known as “Feedback-Informed Treatment” or FIT (you can access two, free, brief and simple-to-use scales by clicking here).  A variety of support materials, and 10,000+ clinicians and administrators are available at no cost via the International Center for Clinical Excellence website.  Importantly, evidence shows clients of therapists who have integrated FIT into their work are 2.5 times more likely to experience improvement over the course of care.

“Exposure!” a choir of professional voices sings, “its the only proven way.”

“No, no,” others insist, “You can tap yourself to emotional freedom.”

“Poppycock!” another group jumps in, “Horizontal saccadic eye movements are the ticket!”

“Beware the dominant discourse,” a few, particularly literate warn, “focusing on what was done to the person can retraumatize, help them reauthor their experiences instead.”

Meanwhile, a smaller and less vocal group shakes their heads in disbelief, saying, “There are no shortcuts.  Healing comes only from identifying and ‘working through’ painful unconscious feelings.”

Turning to the research to answer the question–what heals trauma?”–offers little clarity.  Advocates of most approaches can offer evidence that their preferred approach works–at least one study, and often more, many more–a fact all should find puzzling.  Simply put, how could all approaches work, given they offer competing and often contradictory explanations and techniques?   And yet, no consistent superiority of one particular approach over others is exactly what the latest dismantling and meta-analytic studies show (1, 2, 3, 4, 5, 6, 7).

When attempting to account for why all approaches work equally well, the most common argument made is that different models work for different people.  Said another way, what may be effective in the aggregate may not work for the individual.   “Choice is key,” advocates of this position assert.

More recently, and perhaps in response to the continuing failure to find any meaningful difference in outcome between treatment methods, it has become popular to talk of a set of “mechanistically transdiagnostic…therapeutic strategies…[targeting] the role of a given mechanism in the development and maintenance of a range of psychopathology.”   Ironically, the call for a “universal treatment protocol,” is the “go to” position of those who once advocated for the creation of officially sanctioned lists of specific treatments for specific disorders.

So, which explanation holds water?  Here again, the empirical evidence offers little clarity.  What is important, however, is that these two, diametrically opposed perspectives share a common assumption: healing results from the appropriate application of the right treatment methods.

But what if that’s not true?  What if therapeutic techniques–whether specific to a given model or shared by all–have no inherent power to heal?  Where would that leave us as a profession?  Does it mean that our methods are the therapeutic equivalent of Dumbo’s magic feather?

“A great deal changes, in terms of our ability to help and heal,” psychologist Stephen Bacon suggests, “if we embrace what the research indicates.  Psychotherapy, as a science, is not like engineering.  It operates in a different reality.”

Recently, I had a chance to interview Stephen about his work, and new, thought-provoking, and imminently practical book, Practicing Psychotherapy in a Constructed Reality: Ritual, Charisma, and Enhanced Client Outcomes.  

As you’ll see, he’s a very interesting person–six years in an ashram, a neighbor and student of Krishnamurti, a degree in religious studies, and more.  For me, the “enhanced client outcomes” referenced in the title immediately got my attention.   Wait until you have 30, uninterrupted minutes available, as the interview is one of my longer, and you will want to watch every minute.