The Scottish Government

« Previous | Contents | Next »

Listen

EFFECTIVE AND COST-EFFECTIVE MEASURES TO REDUCE ALCOHOL MISUSE IN SCOTLAND: AN UPDATE TO THE LITERATURE REVIEW

SECTION SEVEN: BRIEF INTERVENTIONS

SUMMARY

This section updates the evidence relating to the effectiveness and cost-effectiveness of brief interventions used to treat problem drinking and alcohol abuse. The main findings of the previous report were;

• brief interventions are effective in reducing alcohol consumption for at least 12 months in patients who are not alcohol dependent;

• pooled results from clinical trials show a 24% reduction in alcohol consumption;

• those who received the intervention were twice as likely to change their behaviour compared with control groups;

• brief interventions have fairly low costs and have been shown to be cost-effective in 3 economic studies;

• using Scottish cost data the cost per life year lies in the range £1446 - £2628 assuming no cost savings; and

• if resource savings are taken into account brief interventions may provide net benefits.

Savings from reduced future use of health care services need to be interpreted with care. It is more likely that resources will be released for alternative uses than that financial savings will be achieved.

More recently published literature supports the effectiveness and cost-effectiveness of brief interventions. Additional points of interest are:

• no gender difference was found where separate data were available;

• brief interventions in a general hospital setting have not been shown to be effective;

• effectiveness has not been shown to vary with type of provider or intensity of intervention;

• four year follow up of Project TrEAT suggests that effects are sustained over this period;

• ten year follow up of the Australian arm of the Phase II WHO collaborative project finds no significant difference between intervention and control groups.

INTRODUCTION

7.1 A brief intervention is a time-limited intervention focusing on changing patient behaviour with respect to alcohol consumption through motivational counselling. The literature reviewed previously in this area was more extensive than for many other interventions. Four recent reviews of the effectiveness literature are included in this update and 5 economic evaluations.

EFFECTIVENESS OF BRIEF INTERVENTIONS

Types of brief interventions.

7.2 Different studies use varying definitions of what is a brief intervention but there are certain characteristics that are generally followed and have been defined in the literature. Moyer et al (2002) suggest that apart from their length, brief interventions generally:

• aim to reduce or to achieve non problem drinking as opposed to abstinence;

• are delivered by general health care professionals not addiction specialists;

• are delivered to non dependent drinkers;

• address motivation to change drinking habits; and

• are self directed.

Reviews have been included if they refer to brief interventions, minimal interventions or extended brief interventions. One review of bibliotherapy, or self help manuals, has also been included (Apodaca and Miller 2003).

Effectiveness of brief interventions.

7.3 The previous review (Ludbrook et al 2002) concluded that brief interventions are effective in reducing alcohol consumption for at least 12 months in patients who are not alcohol dependent, and when compared with no intervention or usual care. Pooled results from 6 trials estimated the effect of brief intervention as a 24% reduction in alcohol consumption (95% CI; 18%-31%). Moyer et al (2002) provide a meta-analysis of a larger number of studies and consider both brief interventions versus controls in opportunistic or non-treatment seeking settings (table 7.1) and brief interventions versus extended treatment in treatment seeking populations (table 7.2).

7.4 The results for brief intervention versus control in non-treatment seeking populations are broadly similar to those reported by previous reviews. The lack of significant effect after 12 months may reflect the small number of studies with extended follow up and large confidence intervals on the results of some of the studies. The authors investigated sources of variation in the results and found that the exclusion of more alcohol dependent individuals could explain this for the >3 - 6 months follow up point. The effect of brief intervention was significantly larger when individuals with more severe alcohol problems were excluded.

7.5 The comparison between brief intervention and extended treatment found little evidence of a significant difference in effect. However, the type of brief intervention delivered in these studies tends to be different from that delivered in the non-treatment seeking population. The brief intervention is more likely to be delivered by a therapist or counsellor rather than a general health care provider and the intervention is more intensive. Drummond (2002) also comments that the extended treatments in these studies may not have been the most effective forms of treatment available.

7.6 A second review concentrated on psychosocial interventions delivered by general practitioners regardless of problem area (Huibers et al 2003). Only two studies relating to alcohol were reported. One is included in the review by Moyer et al and the other was a comparison of different providers reported below.

Table 7.1 Aggregate effect sizes: brief intervention versus control in non-treatment seeking samples

Source Moyer et al 2002

** p < 0.01

*** p < 0.001

Table 7.2 Aggregate effect sizes: brief intervention versus extended treatment in treatment seeking samples

Note: positive values for effect size indicate that extended treatment had a better outcome than brief intervention.

Source Moyer et al 2002

** p < 0.01

7.7 Finally, a review of brief interventions adapted from motivational interviewing included 15 studies related to substance abuse (Dunn et al 2001). Some of the studies relate to alcohol and others to illicit drug use but results are not summarised separately. The studies represent a mixture of motivational interviewing versus no treatment, motivational interviewing versus other treatments and motivational interviewing as an addition to usual treatment. Ten of the 15 studies had significant results in favour of motivational interviewing with effect sizes ranging from 0.30 to 0.95. A similar group of studies, but with three more recent additions, was reviewed by Burke et al (2003). Results were positive and significant in comparison with no intervention but not significant when compared to other active treatments. Both of these reviews included some studies with more dependent populations of drinkers.

7.8 The review of effectiveness of self-help manuals (Apodaca and Miller 2003) found some effect, compared with no intervention, with self referred drinkers but variable results for those identified through opportunistic screening. When compared with any more extensive intervention effect sizes were close to zero.

Population groups.

7.9 For a small number of studies, Moyer et al were able to disaggregate effects by gender and found no significant differences. A review of brief intervention studies involving adolescents found a small but significant effect (Tait and Hulse 2003). All of the studies identified were carried out in the US. Some of the numbers are quite small and the settings range across schools, universities and health care settings.

Settings.

7.10 No reviews of alternative settings were identified. One systematic review related to brief interventions in a general hospital setting (Emmen et al 2004). Of eight studies that met the inclusion criteria, only one found a significant effect. However, in the other studies, drinking was reduced in both intervention and control groups. This suggests that the assessment process may be having an influence. A review by D'Onofrio and Degutis (2002) focussed on the emergency department (ED) setting but found only a small number of studies that had recruited from EDs. In some of these, enrolment into other programmes was the main outcome.

Providers.

7.11 No reviews of alternative settings were identified. One RCT, included in a wider review, compared the effects of a brief cognitive behavioural intervention (CBI) delivered by GPs or a nurse practitioner and found no significant difference between the groups in quantity or frequency of drinking (McIntosh 1997; cited in Huibers et al 2003).

Intensity of intervention.

7.12 McIntosh also compared a two-session intervention with a one-session intervention, both delivered by GPs, and found no significant difference (McIntosh 1997; cited in Huibers et al 2003). Again, there were no reviews on this topic.

COST-EFFECTIVENESS

7.13 The previous literature review concluded that brief interventions for alcohol treatment had been shown to be cost-effective in published research. This update identified 5 studies that provided cost-effectiveness results. These studies varied in settings and provided more mixed results.

7.14 Fleming et al (2002) provided a longer-term (48 months) update of the results from Project TrEAT. Results at 12 months were reported previously. The no treatment group received a booklet on general health issues. The treatment group received the booklet and were scheduled to see their GP for the brief intervention, including a workbook of current health behaviour, a review of the prevalence of problem drinking, a list of the adverse effects of alcohol, a worksheet on drinking cues, a drinking agreement in the form of a prescription and drinking diary cards. The intervention was based on MRC protocols and consisted of two 15-minute appointments a month apart. Patients received a follow up call from the clinic nurse two weeks after each GP meeting. The cost per patient was $205 (£140) (1993 prices). Details of the costing are provided in Ludbrook et al 2002.

7.15 Health outcomes (self-reported alcohol consumption) at six and twelve months had shown that the average number of drinks in the past seven days declined by 39.5% at six months and 40.0% at 12 months. A significantly smaller decrease was reported by the control group (p<0.05). At 48 months the reduction in drinking by the brief intervention group was maintained at a similar level. A further reduction in drinking by the control group resulted in an insignificant difference at 48 months, although the difference had been significant at 24 and 36 months. The effect across the whole period was also significant (p=0.0018).

7.16 Binge drinking reduced by 49.1% (6 months) and 45.7% (12 months), which was significantly greater than the reduction in the control group (p<0.05). In contrast with the results for drinking levels, there was a subsequent increase in binge drinking by the intervention group although it remained below pre intervention levels. The difference between the control and interventions groups was still significant at 48 months and the effect across the whole time period was significant (p=0.0002).

7.17 Savings in emergency room visits and hospital use were $712 per patient at 48 months (p=0.02). Hospitalisations and emergency department visits were both significantly reduced. In terms of crime costs, the difference in the costs of legal events (substance abuse, theft, forgery, assault) was $102 per patient but this was not statistically significant. The saving of $7171 per patient for motor vehicle events was significant (p=0.03). Total savings of $7,985 per patient were also significant (p=0.007).

7.18 Combining the total economic costs and benefits, the benefit-cost ratio is 39:1 (95% CI [5.4, 72.5]) with a net benefit of $7780 (95% CI [$894, $14,668]). Considering only the perspective of the health care system, the benefit-cost ratio is 4.3:1 (95% CI [0.6,8.0]) with a net benefit of $546 (95% CI [$-71, $1164]). These are essentially cost offset results. In terms of health benefits, data are presented on mortality. There were 7 deaths in the control group and 3 in the treatment group but the difference was not statistically significant. No other health benefits were estimated and there was no attempt to combine the health results with the economic data.

7.19 Storer (2003) reports a higher benefit:cost ratio for reduced hospital care costs following brief intervention. However, this is based on an unmatched comparison of those who did and did not receive a brief intervention and the results may be subject to selection bias.

7.20 Freeborn et al (2000) report different results, based on a randomised, controlled trial in an HMO setting. Significant results for drinking behaviour at 6 months had eroded at 12 months. Only number of drinking days per week continued to show a significant difference. There was no significant difference in resource use between the intervention and control groups. The authors point out that this could be considered an effectiveness study as opposed to an efficacy study, in the sense that the study did not rely on volunteer clinics or clinicians. This may have reduced the effectiveness of the intervention, with consequences for resource use.

7.21 Hulse and Tait (2003) compared brief interventions and information only in an inpatient psychiatric setting. A 5 year follow up of administrative data found no difference in health service utilisation between the two groups. However, both groups used significantly fewer services than a matched group of patients who had received neither intervention. No formal cost-effectiveness was carried out; savings were implied from less hospital use. A second Australian study (Shakeshaft et al 2002) compared brief interventions and cognitive behavioural therapy in an outpatient community based drugs and alcohol setting. The interventions were similar in effectiveness but brief intervention had lower costs ($32.84 compared with $76.53) resulting in significantly better cost-effectiveness for brief interventions. Clearly, a more intensive intervention does not provide additional benefit in a group that responds to brief intervention. An interesting point is that the least responsive group were those referred from a legal source. Sobell et al (2002) compared a mailed intervention using either drinking guidelines or personalised feedback. Both interventions were low cost and effective in reducing drinking. However, the subjects were recruited by advertisement and must be assumed to have been well motivated to change their behaviour before responding.

QUALITY AND RELEVANCE OF THE EVIDENCE.

7.22 The trials included in the reviews of evidence come from a range of countries including the UK. The results are fairly consistent in their support for brief interventions and should be applicable to Scotland.

7.23 What is less clear is the generalisability of the results. The practitioners delivering interventions in research settings will be highly motivated and the effects may not be completely replicated in routine practice. Some groups are clearly excluded from the scope of the intervention but the process of consent for research studies means that the participants in the trial are self-selecting. A single study, included in the cost-effectiveness review, has reported effectiveness eroding at 12 months in a study that did not rely on volunteers (Freeborn et al 2000) but there are no systematic reviews in this area.

7.24 None of the cost-effectiveness studies has been conducted in a UK context and the cost-offsets suggested for medical care may be based on higher treatment costs in other countries. There remains little evidence in the economic evaluations of attempts to include health benefits in the analysis. The one study which fails to find any cost saving, (Freeborn et al 2000), provides a cautionary note about the possibility of reduced effectiveness of brief interventions in routine practice.

FURTHER RESEARCH

7.25 The areas identified previously for further research were:

• the effectiveness of brief interventions when used routinely outwith a study setting;

• longer term effectiveness;

• incremental effectiveness when combined with medication;

• costs and effectiveness of implementation strategies for a brief intervention programme; and

• improved economic modelling.

7.26 In the additional literature reviewed, there is one US study addressing the first of these issues (Freeborn et al 2000) and 2 studies have published longer-term follow up data (Fleming et al 2002; Wutzke et al 2002). The former is reported in the cost-effectiveness section above and suggests that effects are maintained at 4 years. The latter study is a 10 year follow up which finds no difference between intervention and controls. This is a result of reduced alcohol consumption in the control group and may relate to diminishing alcohol consumption with age. This study was not strictly eligible for this review, as it was a single effectiveness study, and the results are reported here for noting as the only follow up of this length identified in the literature. Another single effectiveness study of naltrexone in combination with brief intervention showed a reduction in frequency of heavy drinking but not in drinking days (Kranzler et al 2003). Most of the patients were defined as having mild alcohol dependence.

7.27 Overall, the need for further research in all these areas remains, particularly with respect to UK settings.

« Previous | Contents | Next »