Motivational interviewing, widely used counselling technique, may not work well for sexual behaviour change

A meta-analysis of sexual risk behaviour reduction using the counselling technique known as motivational interviewing has found little evidence of efficacy, the Tenth AIDS Impact conference heard this week.

Researcher Rigmor Berg, of the Norwegian Knowledge Centre for Health Services in Oslo, told the conference that a meta-analysis of ten randomised controlled trials of motivational interviewing as a technique for reducing sexual risk found only one outcome measure, alcohol use, in which using the technique resulted in a statistically superior outcome when compared with what was on offer in the control arm.

Motivational interviewing produced no statistically significant difference over control for outcome measures such as unprotected anal intercourse, number of sexual partners, or condom use.

Motivational interviewing (MI) is a counselling technique developed by psychologists William R Miller and Stephen Rollnick in the 1980s (Rollnick and Miller). The first paper on MI was published in 1983 (Miller). The technique is a development of person-centred counselling in which the counsellor – while acknowledging clients' reluctance to change or fear of change – gently points out and 'develops' the discrepancies between the way the client is and the way they would like to be. The aim is to facilitate change that the client is already contemplating.

The technique has proved effective in treating problematic drinking (with a 50% decrease in drinking compared with control or placebo treatments), eating disorders, and drug misuse, although the authors of a meta-analysis of randomised controlled trials in 2003 (Burke) were already warning that they had not found evidence of efficacy in changing HIV risk behaviours.

The importance of MI is that it has become recommended as a standard behavioural intervention by bodies such as NICE in the UK, for a number of populations, such as drug-using adolescents, and is mentioned as a successful intervention in gay men at risk of HIV in the evidence supporting the NICE guideline on the Prevention of sexually transmitted infections and under 18 conceptions issued in 2007 (Downing). MI as a technique is amongst those taught to healthcare staff in HIV and STI clinics.

For the present meta-analysis, Dr Berg and her team found ten randomised controlled trials of MI designed to answer the question: “What is the effectiveness of behavioural interventions adapting the principles and techniques of MI on HIV-risk behaviours for men who have sex with men?”

Out of 155 outcome measures from the ten MI trials, she found ten outcomes that achieved statistical significance in individual trials. These were all behavioural outcomes such as frequency of unprotected sex, number of sexual partners, and condom use. However when these outcomes were combined in the meta-analysis, none retained statistical significance.

The overall improvements in outcomes over the meta-analysis included a 6% reduction in unprotected anal sex with casual partners, a 2% reduction with primary partners, and a 6% increase in condom use. There was a reduction of about one-third in sexual partners over the short term but again this was not statistically significant. The only outcome measure that remained significant was that, in the studies measuring alcohol use, MI more than halved alcohol consumption in the short term though this lost significance over long-term follow-up.

Only a few trials measured biological outcomes such as STIs and none were statistically significant. Dr Berg found one trial in which MI produced a clinically, but not statistically, significant reduction in HIV infections. This, the EXPLORE study, was one of the largest trials of a behavioural intervention to reduce HIV ever conducted, involving 4295 US gay men at risk of HIV, and used MI as one of its components in its 'ten sessions plus top-up' counselling package. It produced a 16% reduction in HIV infections in men given MI sessions compared with control, but this was not statistically significant.

Dr Berg commented that it was interesting that the only behaviour MI seemed to have a consistent effect on, at least in the short term, was drinking, which bore out its efficacy in studies in other populations.

She commented that the reason MI might not be successful in sexual risk behaviour was that sexual risk was a decision shared between two people, and the psychosocial theories underlying techniques such as MI assume that risk behaviours are under the control of the individual – as they are when the individual is trying to do something like stop drinking.

She commented: “The effectiveness of MI as a prevention strategy for unsafe sexual and substance use behaviours among men who have sex with men does not appear promising, though to dismiss it as an intervention for all HIV risk behaviours among all groups of MSM is premature."