On the Working/Therapeutic Alliance and its Measurement

I recently finished reading B.L. Duncan et al, eds., The Heart and Soul of Change: Delivering What Works in Therapy, 2nd ed. (APA Press, 2010). Like all academic collections, this one is uneven, with some chapters managing the unpleasant trick of overpromising and being very thinly sourced, putting one in mind of that 1980s Wendy's commercial "Where's the beef?" 

But two chapters alone are worth the price of the book, and have changed my thinking (and soon my clinical practice!) in a couple of significant ways. These are chapters four ("The Therapeutic Relationship" by John Norcross) and eight ("Yes, It is Time for Clinicians to Routinely Monitor Treatment Outcome" by Michael Lambert). 

Before we get to them, just a few brief comments on the first part of the volume. In the Preface, the editors note that what unites this volume is that it "brings the psychotherapist back into focus as a key determinant of ultimate treatment outcome--far more important than what the therapist is doing is who the therapist is" (xxviii). This is indeed a common theme and welcome focus throughout the book, and it highlights what has often been discussed--but without as much depth--elsewhere: the personhood and personality of the clinician matters far above theoretical orientation, technical prowess, and all the rest. 

The Hideous Effects of the Market are Still Here: 

As we move into the introduction, the editors begin, in my view, to strain credulity somewhat in repeatedly asserting rather flatly that the mania for different models is over ("the fire for the novel, different, and exotic therapies has for the most part been extinguished") and thus proclaim that 'the great 'battle of the brands'" is behind us (25). I am distinctly unconvinced of this, however much I wish it were so. 

The Challenges We Face:

Some sobering facts are laid out in the introduction, some of them picked up later in the book for discussion, including the fact that those who need and want to pursue therapy are often put off by doing so not just by cost but also real doubt about its efficacy, which doubt leads "nearly half of those who begin psychotherapy [to] quit" (31) early on. This has been much commented on elsewhere, but for all that still amazes me. 

How to Account for Variable Outcomes?

Again and again this book lays out impressive evidence that "much of the variability in outcomes in therapy is due to the therapist" (31) and to the "therapeutic relationship" (37) or the therapeutic/working alliance, the power of which is attested to "in more than 1000 findings." A positive alliance is repeatedly demonstrated in the clinical literature as "one of the best predictors of outcome" (37).

Key Components of the Alliance:

The editors briefly outline that an alliance involves three things:

i) "agreement about the tasks and goals of therapy";

ii) a "cogent rationale" that offers an "adequate explanation for the presenting problems"; and

iii) a "set of procedures consistent with the rationale" that will aid towards the accomplishment of the goals in a way the patient can see and understand.

One of the valuable lessons I have heard from both of my supervisors this year is that if you set goals beyond what the patient wants to do or where to go, you will find that s/he will not follow, and so-called resistance will manifest itself. Equally I have learned from both that if you get out too far in front of the person in trying to hurry them along towards the goals, you will turn around and find yourself standing alone. Thus you need, as the editors say here, always to remember "the importance of starting where the client is" (38). 

If you find the person is not with you, or that things are generally out of alignment, then the editors recommend that you simply ask the person what their views are of the goals, alliance, and relationship. This will be the crucial theme developed in the fourth chapter. 

The Gifts in Front of You at the Very Outset: 

Before that, however, let me just posit some theses drawn from the third chapter, "Clients: The Neglected Common Factor in Psychotherapy" by A.C. Bohart and K. Tallman who note, inter alia, that we cannot overlook the phenomena of:

• self-generated change
• spontaneous recovery
• a person's existing strengths
• resilience
• post-traumatic growth
• integration and adaptation of what happens in therapy according to extant ways of thinking and living in the life of the patient.

It is important to take stock of all this from the very beginning and figure out ways you can draw on these gifts in the work that lies ahead.

These authors go on to note that the research indicates how rarely "particular techniques" are mentioned when patients are surveyed about their experiences in therapy. Instead, "studies have consistently shown that" what is emphasized and remembered includes "feeling understood," having "support" to deal with problems but also try out new behaviours, and a "safe space" (or, as I prefer, following Winnicott, a healthy holding environment) for all this to take place. 

What Kind of Therapist Not to Be:

Later in the chapter the authors outline problematic behaviours from therapists that will almost certainly contribute to treatment failure or collapse. These include therapists who:

• are authoritarian
• are prone to hurtful remarks
• are not good listeners
• are aloof, distant, unresponsive
• are too dissimilar in personality from the patient
• are unwilling to offer ideas or practical exercises or advice.

To the above I would want to add something I have learned from reading Jonathan Shedler: therapists who do not challenge their patients to change and grow are also not helping them, and the relationship, if it consists only of affirming and supporting, will prove less than fruitful.

That having been said, everything turns on how the challenges are posed. These authors note later in the chapter that "controlled research trials, particularly in the addictions field, consistently find a confrontational style to be ineffective" (130). Instead, the therapist needs to manifest empathy, be able to deal with resistance, support self-efficacy, and bring discrepancies to the fore; all these, they say, are characteristic of motivational interviewing. 

The Therapeutic Relationship:

We come now to Norcross' chapter of the above name. It repeats some of the findings from earlier in the book about the centrality of the therapeutic relationship over and above theory and technique. 

But it goes on to make a point that I have only once, and briefly, encountered before in a more informal manner, and that was from reading Yalom's The Gift of Therapy, where he will regularly ask patients "How are we--you and I--doing?" as a way of gauging the therapeutic relationship and any possible ruptures to it. When I read that book earlier this year I filed Yalom's question away and thought I might bring it out on occasion. But Norcross, in this chapter, has thoroughly convinced me that it needs to be brought out systematically and regularly, not least because he says (in a finding I have seen repeated elsewhere), "psychotherapists are comparatively poor at gauging their client's experiences of their empathy and the alliance" (117). In fact, therapists very regularly, and by an alarmingly wide margin, overestimate the strength of the alliance and underestimate the problems in it and the patient's willingness to walk away from it with little or no notice. 

To avoid these blind spots and biases, therapists, Norcross says, must regularly ask for feedback from patients to see how the working alliance is, and to repair it when that becomes necessary, as it regularly will. The methods and challenges of asking for such feedback are taken up by Lambert in the eighth chapter, which has some invaluable references to published research, some examples of which you can, happily and helpfully, find online in places such as this. 

Measuring the Working Alliance:

Lambert and others (including Scott Miller, linked above) have been pioneers in developing short, practical, accessible, and easy methods of soliciting regular feedback from patients, not least because without it "practitioners grossly underestimate negative outcomes" (240). 

Before going on, let me record here my previous unreflective snobbery for modern psychology's fetish for measurement. Much of that is still merited (especially towards the rest of the social sciences), but on this issue Lambert's chapter has totally changed my mind. I don't want to sound like a convert, because all converts to anything are a pest whom one should regard with deep suspicion and keep well clear of, but let me record my gratitude for having been forced take account of the evidence that Lambert and others have amassed in an impressive and convincing manner. 

What evidence is that, you ask? There is evidence, first--as noted above--that therapists are not good at discerning the strength of the alliance, or anticipating breaches. There is evidence, second--again noted above--that half of people who begin in therapy never finish and drop out very often because of deficiencies in the relationship, therapist, or both. And finally, and most important here, there is increasing evidence that therapists who solicit and respond to regular feedback "about client progress" see much improved outcomes for those predicted to be at risk of deteriorating or dropping out entirely. 

Looking at studies in both Europe and North America, Lambert has noted dramatic decreases in drop-out and deterioration rates by patients whose therapists regularly sought and used their feedback. He goes on to review a variety of recent mechanisms to solicit this feedback, noting strengths and weaknesses of each. He does all this "because the empirical literature has shown that the quality of the therapeutic alliance is consistently related to outcome" (247). 

Additional benefits of using such feedback include the ease with which such mechanisms fit in to a diversity of approaches: one can use them "regardless of theoretical orientation." For them to work, of course, they need to be brief and easy to follow, and to my mind the best one is the four-question model illustrated in the Miller, Duncan et al article linked above and here. It is certainly something I want to start using. 

If you are like I was, and still skeptical about this, then you very much want to review this chapter in detail, and the very considerable bibliography it provides of additional studies and evidence. Other chapters in the book may be of value to you as well.