Effectiveness of brief interventions proved beyond reasonable doubt

Abstract

All three reviews (Miller & Wilbourne 2002; Moyer, Finney & Vergun 2002a; Moyer, Finney & Swearingen 2002b) are enormously valuable additions to the literature on the treatment of alcohol problems but, in the limited space available, I will restrict comment to the paper by Moyer and colleagues on brief interventions for alcohol problems (Moyer et al. 2002a). There is no question that the meta-analysis of brief interventions reported by these authors is the most comprehensive and methodologically sound of any to have appeared on this topic so far. Doubts have been expressed recently concerning the effectiveness of brief interventions but the main impact of the review by Moyer et al., to my mind, is that it has shown those doubts to be unreasonable; in other words, the effectiveness of brief interventions has been proved, to borrow the legal concept, beyond reasonable doubt. This applies both to the absolute effectiveness of brief interventions by generalists in the non-treatment-seeking population of excessive drinkers and to the relative effectiveness of brief interventions compared with more intensive treatment by specialists in a segment of the treatment-seeking population. This is not to say that no questions remain about the effectiveness of brief interventions, questions that could perhaps be the focus of future meta-analytical reviews and should certainly be addressed in future research. Perhaps the main issue here concerns the distinction between ‘efficacy’ and ‘effectiveness’ trials (Holder et al. 1999). In the trials of brief interventions in primary health care in which I have been involved (Heather et al. 1987; Richmond et al. 1995), screening for excessive drinkers has been incorporated into routine practice as opposed to recruiting research subjects by some special procedure, and brief intervention has been offered immediately upon a positive case being identified. Thus, in these terms, these studies qualify as effectiveness trials carried out under real-world conditions of general medical practice, rather than in special conditions designed to optimise internal validity. In both cases, we failed to find an effect of intervention on level of alcohol consumption. In one case (Heather et al. 1987) the study was under-powered but in the other case (Richmond et al. 1995) we found a significant reduction in alcohol problems which we interpreted as a genuine effect of intervention. The point, however, is that brief interventions delivered in naturalistic conditions are probably less effective than estimates of effect size based on efficacy trials, which make up the majority of trials reported in the literature, would suggest. This may even account for the fact that some brief interventions trials, particularly those in non-Anglo-Saxon cultures, have reported negative findings (Aalto et al. 2001). The evidence reviewed by Moyer and her colleagues shows that brief interventions among excessive drinkers can work but whether they do actually work in practice will depend on adjusting the means of implementation and delivery to the particular characteristics of the primary health care system and possibly to wider cultural factors in each society. This is precisely the aim of the ongoing Phase IV of the long-standing World Health Organization collaborative project on the detection and management of alcohol-related problems in primary health care (Monteiro & Gomel 1998). Apart from the obvious question of the longer-term effects of opportunistic brief interventions, probably the most important substantive issue concerns the optimal length and associated contents of these interventions. The work of Moyer et al. and of others shows clearly that the widespread implementation of brief, structured and personalized advice, taking no more than 5–10 minutes to deliver, would be effective in reducing alcohol-related harm in a population and would potentially be a hugely cost-effective response to that harm. But could somewhat longer interventions be even more effective? If so, to what types of excessive drinker should they be offered and who should offer them? An intriguing but neglected finding from the WHO Phase II trial (Babor & Grant 1992) was that ‘simple advice worked best for male patients who had experienced a recent alcohol-related problem, while brief counselling (15 minutes following 20 minutes’ assessment) worked better for those who did not have a recent problem’ (p. 3, parentheses added). This suggests that a longer type of brief intervention based on principles of motivational enhancement could be offered usefully to excessive drinkers who do not recognize a problem with their alcohol consumption, assuming that certain classes of health professionals have the time and inclination to deliver it. There is some evidence to support this hypothesis from a trial of brief interventions among hospital in-patients in Sydney (Heather et al. 1996), but more research is clearly needed on this key issue. Turning lastly to the class of brief interventions in specialist centres among the treatment-seeking population, I note the finding of the review that brief interventions ‘were not more effective that control conditions in studies where more severely impaired persons were not excluded’, suggesting that such interventions ‘are useful only for patients with less severe drinking problems’ (pp. 279–292). Against this, however, findings from the Project Match Research Group (1997, 1998) showed that a briefer treatment (motivational enhancement therapy given in four sessions over 12 weeks) was no less effective than two kinds of more intensive treatment across all patients taking part in the trial; the only specific contraindication found was that out-patients with a social network supportive of drinking did better with Twelve-Step facilitation than with motivational enhancement therapy. While I concur fully with the warnings in the last paragraph of the review against ‘wholesale replacement of specialist, extended approaches to treatment with considerably brief interventions’, I believe it is possible that research will confirm an applicability of briefer treatments of the kind studied in Project MATCH to a broader range of problem drinkers than indicated by the prudent policy Moyer et al. quite rightly recommend on the basis of present evidence. Findings from the UK Alcohol Treatment Trial (UKATT Research Team 2001), which will be published over the next couple of years, will be directly relevant to this possibility.