The alliance in motivational enhancement therapy and counseling as usual for substance use problems
Effectiveness bank home page. Opens new windowResearch analysis

This entry is our analysis of a study considered particularly relevant to improving outcomes from drug or alcohol interventions in the UK. The original study was not published by Findings; click Title to order a copy. Free reprints may be available from the authors – click prepared e-mail. The summary conveys the findings and views expressed in the study. Below is a commentary from Drug and Alcohol Findings.

Links to other documents. Hover over for notes. Click to highlight passage referred to. Unfold extra text Unfold supplementary text
Copy title and link | Comment/query |

The alliance in motivational enhancement therapy and counseling as usual for substance use problems.

Crits-Christoph P., Gallop R., Temes C.M. et al.
Journal of Consulting and Clinical Psychology: 2009, 77(6), p. 1125–1135.
Unable to obtain a copy by clicking title? Try asking the author for a reprint by adapting this prepared e-mail or by writing to Dr Crits-Christoph at You could also try this alternative source.

Rarely has counselling been so deeply analysed as in this US study of mainly alcohol and cocaine dependent patients. The far-reaching implications are that some counsellors generate relationships with clients which feed through to better outcomes – but also that the 'best' relationship builders are not on average the most effective.

Summary A working relationship or 'therapeutic alliance' between patient and counsellor is expected to be particularly important for motivational interviewing, which seeks to engender change built on empathy, listening, respect and other features of a collaborative and supportive relationship. The featured report sought to determine whether this was the case in comparison to usual US substance use counselling. It carefully teased out variation between therapists in the extent to which they fostered an alliance across all their patients, versus variation between patients seeing the same therapist. If the former was substantial it would mean some therapists were better at generating collaborative working in their patients, which might be related to how well those patients did in controlling their substance use. One reason alliance might vary is the extent to which therapists embodied the supposedly relationship-building motivational style, another issue investigated by the report. First this account summarises the earlier findings (small text) from the study which set the context for the analysis made in the featured report.

The report drew its data from a study conducted in 2001–2004 at five US treatment centres The centres were collaborators in researchpractice partnerships in the US National Institute on Drug Abuse Clinical Trials Network. offering outpatient counselling. In all 35 therapists at the centres who volunteered Most did. to join the study were allocated at random to implement two forms of individual counselling, an addition to the group counselling at the centres. Both individual approaches occupied three sessions over the first four weeks of what was on average a 10-week treatment episode. One simply replicated usual one-to-one counselling at the centres, for which the counsellors received no special training or supervision as part of the study. The other replaced this with sessions based on a manualised form of motivational interviewing, for which the counsellors had been specially trained, tested to ensure at least adequate/average delivery of the approach, and supervised via taped sessions to ensure they remained on track.

Of the 461 patients in the trial, about 60% each were assessed as problem (mainly dependent) drinkers or cocaine users. Typically they were single men in their 30s and half were in full time employment. The severity of their substance use was assessed during the 16 weeks after they started treatment.

A report on the main outcomes found that while seeing the counsellors, both sets of patients substantially reduced use of their main problem substance, but this was sustained over the next 12 weeks only after motivational counselling – an effect due to the 40% of patients with a primary alcohol problem; others did about as well regardless of the approach used for individual counselling. The report also found clinics differed in how well their patients did and how much was gained or not by implementing motivational interviewing, but after this had been taken in to account no significant differences could be attributed to individual therapists. However, this analysis included just 20 of the 35 therapists, two from each approach at each of the five centres. With so few at each clinic it seems possible that individuals at some clinics actually were more effective than those at others, but that this was associated with where they worked in such a way that its impact was obscured when both factors were included in the same analysis. Another report on the same study found that patients of motivational counsellors In respect of motivation changes, the same was true to a lesser degree of the counsellors who implemented usual counselling. who more often and more skilfully implemented the approach expressed greater increases in motivation to change during counselling sessions and were less likely to test positive for illegal drugs during the four weeks they were individually counselled. However, other measures of substance use were not related to the counsellor behaviours assessed by the study.

The featured report assessed how much of these impacts could be attributed to the therapeutic alliance. It drew its data from 30 of the 35 therapists and 319 of the 461 patients, the ones who had attended their first counselling session and who after the second reported how they saw their alliance with their counsellor. In respect of substance use, the analysis further narrowed down to 257 patients who completed at least one of the required research interviews.

Main findings

The Helping Alliance Questionnaire: patient version

Patients indicate their agreement with each question ranging from strongly disagree to strongly agree

1. I feel I can depend upon the therapist.
2. I feel the therapist understands me.
3. I feel the therapist wants me to achieve my goals.
4. At times I distrust the therapist's judgment.
5. I feel I am working together with the therapist in a joint effort.
6. I believe we have similar ideas about the nature of my problems.
7. I generally respect the therapist's views about me.
8. The procedures used in my therapy are not well suited to my needs.
9. I like the therapist as a person.
10. In most sessions, the therapist and I find a way to work on my problems together.
11. The therapist relates to me in ways that slow up the progress of the therapy.
12. A good relationship has formed with my therapist.
13. The therapist appears to be experienced in helping people.
14. I want very much to work out my problems.
15. The therapist and I have meaningful exchanges.
16. The therapist and I sometimes have unprofitable exchanges.
17. From time to time, we both talk about the same important events in my past.
18. I believe the therapist likes me as a person.
19. At times the therapist seems distant.

Click here to download the questionnaire.

The first and surprising finding was that as rated by the patients ( figure), their working relationships were on average Also motivational and usual counsellors varied to roughly the same significant degree in the average ratings of their patients. very good with the counsellors regardless of whether they had been assigned to motivational counselling. After this, how well patients did during the rest of their time with the counsellors (ie, weeks three and four of treatment) was unrelated to these ratings, but an effect did emerge in weeks four to 16, even though patients were no longer being seen individually. Across this period the average trend in how often a counsellor's patients used their primary substance was significantly related to how far that counsellor had (as assessed up to 12 weeks earlier) generated a positive alliance across their patients. This relationship was, however, not straightforward. Best results were seen among patients seen by counsellors who scored about average (for this set of counsellors, still very high) on their patients' experience of working with them; these patients sustained the substance use reductions they had achieved while seeing the counsellors. But among other patients, on average the gains they had made eroded, and this was the case both when their counsellors had been particularly good at generating a positive alliance, and when they had been particularly poor.

In contrast to differences between counsellors, differences between patients of a particular counsellor in how positively they saw their relationship were unrelated to that patient's substance use trends; patients who felt they had an extremely good relationship did as well as those who thought it not so good. Other interesting 'negative' findings were that relationships between alliance and substance use were similar regardless of the type of counselling, and that the alliance as rated by the counsellor did not vary much on average between counsellors, and was unrelated to trends in their patients' substance use

Next the study investigated what might have caused a counsellor to generate on average better or worse alliance scores across their patients. Across all their recorded sessions, how competently each on average used motivational skills was unrelated to how their patients felt about working with them, but the more extensively they used these techniques and skills, Assessed skills were:
Open-ended questions
Reflective statements
Fostering collaboration
Motivational interviewing style
Client-centred problem discussion and feedback
Pros, cons, and ambivalence
Heightening discrepancies
Motivation for change
Change planning
the better their patients as a whole rated them, and this was particularly the case for the motivational counsellors. But this did not account for the relationship between alliance and substance use trends; no matter how often counsellors used motivational skills, their alliance scores had a similar relationship to substance trends in their patients. Something else generating the alliance, not just the use of the assessed motivational techniques, accounted for its apparent influence on outcomes.

The authors' conclusions

Unexpectedly, training and supervising counsellors in motivational interviewing did not lead to more positive relationships with their patients, despite the approach's stress on empathy, acceptance, positive regard, and clear discussion about goals, and even though these and other such interactions were indeed elevated among the motivational counsellors. One conclusion is that forming a positive alliance is not unique to motivational-style counselling. Counsellors never trained in motivational interviewing can generate very positive alliances, in this study perhaps because typically they were highly experienced and used client-centred counselling skills to a fairly high degree, enough perhaps to match the alliance-generating impact of explicitly motivational counselling.

For any given counsellor, how positively each of their patients felt about them was unrelated to that patient's substance use. However, patients seen by counsellors who typically generated (relative to the average in this study) relatively high or low alliance scores were less likely to sustain their initial substance use reductions, and this was the case for both motivational and treatment-as-usual counsellors. Similar findings have also emerged in general psychotherapy/counselling.

These findings suggest that in so far as the alliance influences desired improvements in the patients, it is not to do with how well an individual therapist and patient get along. Instead therapists differ perhaps in skill levels, interpersonal styles, abilities to learn and implement alliance-fostering techniques, or to identify and repair alliance ruptures, differences which mean they tend to form relatively good or poor relationships with their patients, in turn affecting whether they make the desired changes in their lives.

This pattern of findings justifies efforts to train therapists and counsellors to enhance alliance with their patients, and/or to select therapists with the right personalities and skills, which may to a degree not be teachable. Training does not it seems have to explicitly be based on motivational principles.

The findings also seem to confirm the truism that patients can often initiate abstinence on entering treatment, and that where the quality of treatment has an impact is in sustaining those gains and dealing with lapse and relapse. This was the phase during which an earlier report on the study found a motivational approach more effective than usual counselling, and during which in the featured report each counsellor's ability to generate an alliance among their patients appeared to exert an influence. However, generating especially positive alliances was not necessarily a good thing; as with counsellors who generated relatively poor alliances, patients of these counsellors had poorer outcomes.

One way counsellors apparently enhanced their patients' alliance ratings was by using more of the skills and techniques which characterise motivational interviewing. However, these were not the driving force which made a high alliance counterproductive. Therapists can engage in these types of interactions without fearing 'over-use' will damage their patients' prospects.

It should be remembered that these implications derive from a just a few studies. In the featured study, results might have been different had the counselling programme been a typical length rather than just three sessions, if the alliance had been assessed later in therapy when the relationship had developed further, or if it had been rated from sessions recordings. The same counsellors led group therapy sessions which might have included their patients, and many remained in treatment during all or part of the period during which their substance use was measured, meaning the results are complicated by the impact of continuing treatment. Just 56% of the patients could be included in the outcome analysis; had more been able to be included, results might have differed. Only the motivational therapists were specially trained and supervised, so the impact of the motivational content is confounded with any impact from receiving extra training and supervision, whatever the content. Patients were not randomly assigned to counsellors, so the possibility cannot be excluded that the clinics allocated certain types of patients to certain counsellors.

Findings logo commentary The featured study seems to confirm the intuitive assumption that how well a counsellor relates to their clients has an impact on how well the client does – that in the words of a Findings hot topic entry, treatment staff matter.

This assumption is supported by evidence from psychotherapy trials in general, in respect of which a comprehensive meta-analytic review commissioned by the American Psychological Association concluded that whether seen from the patient's or the therapist's perspective, the more solid the working relationship, the better the outcomes. Though it accounts for a relatively modest proportion of variance in treatment outcomes, the reviewers saw the alliance-outcome relationship as one of the strongest predictors of treatment success. As in the featured study, they also found evidence that therapists varied in the degree to which they foster an alliance with their clients, suggesting that alliance development is a skill and/or capacity therapists can develop.

In substance use treatment the picture is less clear. A review published in 2005 found that therapeutic alliance early in treatment was more consistently related to engagement and retention than to substance use outcomes, especially when those outcomes were assessed at times distant from the assessment of the alliance.

A study of motivational interviewing training sheds further light on what it is about counsellors which generates retention/outcome-enhancing alliances. Many studies recruit skilled therapists and then train and supervise them to ensure competence, partly eliminating the normal variability in competence, but this study randomly allocated an unusually diverse (in terms of initial proficiency) set of addiction counsellors and clinicians to different motivational interviewing training regimens and then tested their performance with clients. It found that client engagement was unrelated to the frequency with which the therapist made statements compatible (such as open questions) or incompatible (such as warning) with the specific techniques recommended in motivational interviewing, but was strongly related to embodying the overall spirit of the approach and to more general social skills including empathy, warmth, supporting the client's autonomy, and coming across as 'genuine', an amalgam of seeming open, honest and trustworthy. In the featured study too, the skills and techniques exhibited by motivational and non-motivational counsellors partly determined client engagement, and to the same degree in both sets of therapists.

The featured study's supposition that to some extent these personal attributes cannot be trained is supported by another study of motivational interviewing training, the findings of which suggested that when it comes to choosing addiction and mental health therapists, choosing the 'right' people who have not been trained in motivational interviewing would be better than choosing the 'wrong' people who have been trained; the former not only started at a higher level, but were more able to benefit from and retain training.

Some puzzling findings

Of the thought-provoking findings thrown up by the featured study, most notable was that the patients of counsellors who generated unusually high alliance ratings did less well than those of average counsellors. An earlier psychotherapy study found a similar relationship. The authors surmised that patients with an unrealistic idealisation of their therapists may later feel disappointed, or that a very high alliance rating signifies an excessive need for attachment indicative of a poor prognosis. Neither explanation fits the data from the featured study, because under both scenarios there should have been a relationship between outcomes and the alliance levels of each individual patient seeing the same therapist.

Another possibility is that such counsellors also generated feelings of being reliant on their support; when individual sessions ended, reliance may have rebounded to a feeling of now being devoid of needed support, with a consequent rebound in substance use. At the other end, counsellors with relatively poor alliance scores from their patients may have developed relationships too weak to support a determination not to relapse and strategies to avoid it.

A further possibility is that rather than what might be thought as the perfect stable-normal personality profile for a therapist, workers whose imperfections and deviancies to an extent match those of their drugtaking clients are most able to help them, even if this means their clients see them as less than ideal. These were the implications of a study in England set in a drug service for marginalised clients which found that drug workers who prioritised stimulation, self-direction and hedonism experienced more positive client outcomes than those prioritising security, conformity, benevolence, and tradition. A similar picture emerged from a US study of ex-addict methadone counsellors published in 1974, which found that "deviant" personalities who shared the insecurities and edginess of their patients and had a suspicious outlook on life had patients who engaged better and used drugs less.

The featured study's findings may have been dependent on the generally high alliance-generating competence of the counsellors. Had there been a greater range and on average lower competence, it seems possible that there would have been a simpler finding: that the more a therapist generated good alliances, the better the outcomes for their patients.

Another puzzling finding was that though over the following 12 weeks motivational counselling promoted desistance from the main drug the client was using, it did not do so to any greater extent the more competently and skilfully it had been delivered, as would have been expected if motivational interviewing was the active ingredient. It seems more likely that factors common both to motivational approaches and effective substance use counselling were the main ingredients. These were perhaps present to a greater degree among the motivational therapists, possibly due to their specific training, but also possibly due to the opportunities they had to learn from feedback on their performance with patients and the optimism and morale-boost associated with learning a new approach and receiving extensive attention from trainers and supervisors.

Also it seems contradictory that some therapists were characterised by alliance levels (ie, about average) among their patients which were associated with the best substance use outcomes, yet there was no significant relationship between which counsellor a patient saw and how well they fared on the same measure. On the face of it one would expect the impact of a counsellor's alliance-generating abilities to be reflected in the progress of their caseload.

Thanks for their comments on this entry to Paul Crits-Christoph of the Center for Psychotherapy Research at the University of Pennsylvania Medical School in the USA. Commentators bear no responsibility for the text including the interpretations and any remaining errors.

Last revised 22 May 2013. First uploaded 10 May 2013

Open Effectiveness Bank home page

Top 10 most closely related documents on this site. For more try a subject or free text search

STUDY 2011 Therapist effectiveness: implications for accountability and patient care

DOCUMENT 2013 Sometimes best to break the rules

STUDY 2005 How does motivational interviewing work? Therapist skill predicts client involvement within motivational interviewing sessions

ABSTRACT 2011 Evidence-based therapy relationships: research conclusions and clinical practices

REVIEW 2006 My way or yours?

REVIEW 2005 The motivational hallo

STUDY 2012 A preliminary study of the effects of individual patient-level feedback in outpatient substance abuse treatment programs

STUDY 2010 A randomized controlled study of a web-based performance improvement system for substance abuse treatment providers

REVIEW 2018 Alliance rupture repair: a meta-analysis

REVIEW 2018 Congruence/genuineness: a meta-analysis