Re: Effectiveness of screening and brief alcohol intervention in primary care (SIPS trial): pragmatic cluster randomised controlled trial
It is gratifying that our article has stimulated further discussion and debate and we are pleased to be able to respond to the principal issues raised.
Firstly, it is helpful to see additional research from a different cultural setting (such as that presented by Professor Assanangkornchai and colleagues) which appears to show similar findings to the SIPS trial. We agree that more research is needed on alcohol interventions, both in countries where alcohol is an overt part of many people’s day-to-day life and where drinking is less socially acceptable (as identified by Professor Sheriff).
In contrast to Professor Saitz, we feel that the brief intervention evidence-base to date has indicated a superiority of brief advice over brief counselling and a growing preponderance of effectiveness rather than efficacy trials. In 2007, a meta-analysis of 22 brief intervention trials (of 29 identified) in primary care reported a statistically significant reduction in average weekly drinking compared to control conditions but no evidence of significant benefit of longer versus shorter input or extended (typically counselling) versus simpler (generally advice-based) intervention. (1) Furthermore, the majority of studies in this field were judged to be clinically relevant effectiveness trials (with high external validity) rather than ideal world efficacy trials (with high internal generalizability). In a field that has evolved for over 25 years, it is to be expected that evaluations have increasingly reflected the variability and constraints of real world primary care. (2)
The key aim of the SIPS trial was to evaluate the impact of screening and brief alcohol intervention on patients’ drinking outcomes rather than assess the implementation (service-delivery) outcomes of different brief intervention strategies (3). Such an implementation trial is currently being conducted in five countries as part of the ODHIN trial (4). Similarly, van Beurden et al. (5) considered the impact of an intensive multi-faceted improvement programme on GPs’ management of alcohol problems and the primary outcome of this study was the number of patients screened and intervened with. There is a need for caution in reaching conclusions about the implementability of the different approaches compared in the SIPS trial, since recruitment in each practice ended as soon as 31 patients were enrolled. Moreover, the high satisfaction rates reported by patients across the SIPS trial seem to be at odds with the notion of brief interventions done poorly as suggested by Professor Saitz. Nevertheless, intervention fidelity is being explored in a qualitative process evaluation of the SIPS trial. Furthermore, the use of technology to facilitate brief intervention delivery (raised by Dr Winstock) is receiving attention in this field, as exemplified by the reference cited. (6)
In our paper, (3) we clearly reported a lack of statistically significant differences between the three arms of the trial. We also explained that to unequivocally demonstrate that providing feedback plus a patient information leaflet led to reduced rates of hazardous and harmful drinking, the SIPS trial would have had an additional no intervention control to compare the possible effects of feedback/leaflet against screening/assessment only. However, since we introduced a screening procedure into routine consultations, our control condition was deemed to be the minimal level of input required for patients with alcohol-related risk or harm for ethical reasons. (7) In considering the reasons for a lack of statistically significant differences between the conditions, it was appropriate to include ‘regression to the mean’ and this was specifically raised by the papers’ referees during the peer review process.
Dr Braillon has suggested that the 10 item AUDIT questionnaire may not have been a suitable tool to detect differences between the trial conditions. However, AUDIT has been shown to be responsive to changes in drinking behaviour after intervention in a primary care population. (8) In addition, since assessment reactivity is a ‘live issue’ in this field, (9) the brevity of AUDIT and its coverage of both alcohol intake and drinking problems were felt to provide an advantage over a more precise but much longer consumption measure, such as the Time-line Follow-back method. (10) We examined a range of secondary drinking outcomes, none of which showed a statistically significant difference between interventions. This included a composite measure of the consumption items of AUDIT (1-3) expressed as average drinks per day. Although not reported in the paper for the sake of brevity, this measure showed no significant differences between interventions at either 6 or 12 months.
Finally, when assessing the SIPS findings in relation to the wider evidence-base, we raised the possibility that our control condition (consisting of personalised verbal feedback and relevant written information) may have contained ‘active ingredients’ of behaviour change which could have resulted in reduced levels of drinking. The issue of potential control group effects has been noted in the wider health psychology field (11) and discussed in three systematic reviews reporting consistently reduced drinking in the control groups of brief intervention studies. (12-14) Overall, this wider literature points to the conclusion that even relatively minimal input in the context of opportunistic alcohol screening could have beneficial findings. What the SIPS trial adds to the existing evidence base is that we were unable to demonstrate in a typical primary care setting that more intensive interventions provide additional benefit over screening-feedback plus written information. This is consistent with the findings of the Cochrane review (1).
1. Kaner E, Beyer F, Dickinson H, Pienaar E, Campbell F, Schlesinger C, et al. Effectiveness of brief alcohol interventions in primary care populations. Cochrane Database of Systematic Reviews. 2007;Issue 2:CD004148.DOI:10.1002/14651858.CD4148.pub3.
2. Flay BR. Efficacy and effectiveness trials (and other phases of research) in the development of health promotion programs. Preventive Medicine. 1986;15:451-74.
3. Kaner E, Bland M, Cassidy P, Coulton S, Dale V, P. D, et al. Effectiveness of screening and brief alcohol intervention in primary care (SIPS trial): Pragmatic cluster randomised controlled trial. BMJ. 2013;346:e8501. Doi:10.1136/bmj.e8501.
4. Keurhorst M, Anderson P, Spak F, Bendtsen P, Segura L, Colom J, et al. Implementing training and support, financial reimbursement, and referral to an internet-based brief advice program to improve the early identification of hazardous and harmful alcohol consumption in primary care (ODHIN): study protocol for a cluster randomized factorial trial. Implementation Science. 2013;8(11):http://www.implementationscience.com/content/8/1/11.
5. van Beurden I, Anderson P, Akkermans R, Grol R, Wensing M, Laurant M. Involvement of general practitioners in managing alcohol problems: a randomized controlled trial of a tailored improvement programme. Addiction. 2012;107:1601-11.
6. Wallace P, Murray E, McCambridge J, Khadjesari Z, White I, Thompson S, et al. On-line Randomized Controlled Trial of an Internet Based Psychologically Enhanced Intervention for People with Hazardous Alcohol Consumption. PLoS ONE 6(3). 2011;6(3):e14740. doi:10.1371/journal.pone.0014740.
7. Sackett D, Holland W. Controversy in the detection of disease. The Lancet. 1975;306(7930):357-9.
8. Bradley KA, McDonell MB, Bush K, Kivlahan DR, Diehr P, Fihn SD. The AUDIT alcohol consumption questions: reliability, validity, and responsiveness to change in older male primary care patients. Alcoholism: Clinical & Experimental Research. 1998 Nov;22(8):1842-9. PubMed PMID: 9835306.
9. Kypri K, Langley JD, Saunders JB, Cashell-Smith ML. Assessment may conceal therapeutic benefit: findings from a randomised controlled trial for hazardous drinking. Addiction. 2007;102:62-70.
10. Sobell LC, Sobell B. Time-Line Follow Back: A technique for assessing self-reported alcohol consumption, in Measuring Alcohol Consumption: Psychosocial and Biological Methods. In: R. Litten Z, Allen P, editors. Totowa NJ: Human Press; 1992. p. 41-72.
11. Michie S, Prestwich A, deBruin M. Importance of the nature of comparison conditions for testing theory-based interventions: reply. Health Psychology. 2010;29(5):468-70.
12. Bernstein J, Bernstein E, Heeren T. Mechanisms of change in control group drinking in clinical trials of brief alcohol intervention: implications for bias towards the null. Drug and Alcohol Review. 2010;29:498-507.
13. Jenkins R, McAlaney J, McCambridge J. Change in alcohol consumption in control groups in brief intervention studies: a systematic review. Submitted to Addiction. 2008 January:1-30.
14. McCambridge J, Kypri K. Can Simply Answering Research Questions Change Behaviour? Systematic Review and Meta Analyses of Brief Alcohol Intervention Trials. PLOS ONE. 2011;6(10):e23748. doi:10.1371/journal.pone.0023748.
Competing interests: The team co-authored the original article on which this response to comments is based. The same conflicts of interests statment applies as per the published paper.