Can stand-alone computer-based interventions reduce alcohol consumption? Zarnie Khadjesari (PhD student) Elizabeth Murray (Director e-Health Unit) e-Health.

Can stand-alone computer-based interventions reduce alcohol consumption?

Zarnie Khadjesari (PhD student)Elizabeth Murray (Director e-Health Unit)

e-Health Unit, University College London

Christine Godfrey (Head of Department)Catherine Hewitt (Research Fellow)

Dept. Health Sciences, University of York

Suzanne Hartley (Senior Trial Coordinator)CTRU, University of Leeds

Background

•Alcohol misuse is a major public health concern

•Gap between need and access •Internet interventions

– Convenient, confidential, and comparatively low cost– Scalability and personalised approach

•Recent reviews– Elliot 2008 (computer-based interventions for college drinkers)– Bewick 2008 (Internet interventions)– Riper 2009 (personalised feedback interventions – any modality)

Why conduct this review?

•All designs of computer-based intervention

•All computer-based (on- and off-line)

•All adult populations

•Meta-analysis

Aim

• To determine the effectiveness of computer-based interventions aimed at reducing alcohol consumption

• Computer-based interventions compared with either:

i. Minimally active comparator (e.g. assessment-only, information-only website)

ii. Active comparator (e.g. face-to-face motivational interview)

Inclusion criteria

• Study design: RCT• Population: Adults (excl. dependent drinkers)• Intervention:

– Computer-based interventions aimed at reducing alcohol intake– Definition: behavioural interventions, adapted for computer– Stand-alone: no expert facilitation

• Outcome: Alcohol consumption– Grams per week– Frequency of binges / week

Search results

Databases searched from inception – end 2008

MedlineEmbaseWeb of ScienceCochrane LibraryPsycINFOCinahlERICISI ProceedingsIBSSIndex to Theses

10 databases searched 8,084 references

Excluded7,930

Full paper ordered

154

Excluded119

Included publications

36

Individual studies

23

Characteristics of included studies (1)

Year published

1997 = 1; 2004 = 4; 2005 = 3; 2006 = 3; 2007 = 7; 2008 = 5.

Country US = 17; NZ = 3; Netherlands = 1; Germany = 1; UK = 1.

Population •Students = 17 •Problem drinkers from general population = 3•Workplace employees = 2•Emergency department attendees = 1

Characteristics of included studies (2)

Screening •At-risk drinkers = 10•Any drinkers = 6•No-screen = 7

Intervention approach

•Personalised feedback•Harm-prevention / skills training•Expectancy challenge•Self-control / CBT / motivational enhancement

Characteristics of included studies (3)

Comparator Minimally active comparator = 21 Active comparator = 3

Outcome Grams per week = 18Frequency of binges (days or occasions / week) = 8

Results

Comparison: minimally active comparator (n=2,425)Outcome: g/wk

Sub-group analysis - population

Results

Comparison: minimally active comparator (n=848)Outcome: binge frequency / wk

Results

Comparison: active comparator (n=457)Outcome: g/wk

Further analyses

• Skewed data

• Baseline risk

• Loss to follow-up

Summary of findings

• Computer-based interventions appear:– more effective than minimally active comparator– as effective as alternative treatment approaches

• Findings support continued development and evaluation of computer-based interventions for reducing alcohol intake

Limitations of this review

• Restricted to stand-alone interventions

• Different types of computer-based interventions

• Two measures of alcohol consumption

• Mediators of drinking outcomes, s/a motivation, normative perceptions.

• Dose response

Gaps in the literature

• Few comparisons with conventional approaches

• Few studies in non-student adult populations

• Few studies outside the US

• Few studies measuring long-term effectiveness

Thank you for listeningQuestions, comments, suggestions?