Personalised digital interventions for reducing hazardous and harmful alcohol consumption in community‐dwelling populations
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
The main objective is to assess the effectiveness and cost effectiveness of digital interventions for reducing hazardous and harmful alcohol consumption and/or alcohol‐related problems in community‐dwelling populations.
We envisage two comparator groups: (1) no intervention (or minimal input) controls; and (2) another active intervention for delivering preventive advice or counselling to reduce hazardous or harmful alcohol consumption. Specifically, we will address two questions: (1) Are digital interventions superior to no intervention (or minimal input) controls? This question is important for individuals accessing interventions through their own motivation or interest. These individuals will be unlikely to experience active practitioner input and it is important to understand whether digital interventions are better than general material they might seek out on the internet or via mobile phone‐based apps etc. (2) Are digital interventions at least equally effective as face‐to‐face brief alcohol interventions? Practitioner delivered brief interventions are generally accepted to be the best alternative in secondary preventive care in health, workplace, educational or community settings. However, time constraints can impede face‐to‐face delivery of such interventions and it is important to know whether digitally provided input can yield comparable effects to interventions delivered by trained practitioners. We will also identify the most effective component behaviour change techniques of such interventions and their mechanisms of action.
Secondary objectives are as follows:
-
To assess whether outcomes differ between trials where the digital intervention targets participants attending health, social care, education or other community‐based settings and those where it is offered remotely via the internet or mobile phone platforms;
-
To develop a taxonomy of interventions according to their mode of delivery (e.g. functionality features) and assess their impact on outcomes;
-
To identify theories or models that have been used in the development and/or evaluation of the intervention – this will inform intervention development work.
Background
Description of the condition
Excessive drinking contributes significantly to physical and psychological illness, injury and death and a wide array of social harm in all age groups (WHO 2011). Contributing to over 60 types of diseases, alcohol drinking is the leading risk factor worldwide for disease burden in middle‐income countries; it is second only to tobacco use in high‐income countries, and third after childhood underweight and unsafe sex in low‐income countries. As well as the direct harms to health, 20% of deaths due to road traffic accidents, 30% of deaths caused by oesophageal and liver cancer, epilepsy and homicide, and 50% of all deaths caused by liver cirrhosis are attributable to alcohol (WHO 2009). Although drinking low amounts of alcohol has been shown to decrease the incidence of some diseases (particularly coronary heart disease in later life) and can have a positive social effect, the net effect of alcohol consumption is detrimental to health. The economic cost – including both health and social harm, such as property damage and domestic violence – relating to alcohol consumption tends to amount to more than 1% of gross domestic product in high‐ and middle‐income countries (Rehm 2009).
People drinking hazardously display a repeated pattern of drinking above recommended limits and are at risk of (but not yet experiencing) physical or psychological harm, whilst those drinking harmfully are drinking above recommended limits and currently experiencing harms (WHO 1992). Hazardous or harmful patterns of alcohol consumption can involve either regular exceeding of consumption guidelines, or more infrequent but high volume binge drinking. People exhibiting hazardous or harmful drinking are more numerous than those with alcohol dependence (e.g. McManus 2009 in the United Kingdom (UK)), and at a population level the greatest impact on alcohol‐related problems can be made by addressing interventions towards the former groups (McGovern 2013).
Description of the intervention
A proven strategy for reducing excessive alcohol consumption levels across the population is to offer a brief intervention in primary care provided by general practitioners, nurses or other generalist health professionals; a Cochrane review incorporating a meta‐analysis of 22 randomised clinical trials found that face to face brief interventions in primary care settings were consistently effective at reducing excessive drinking, producing an average reduction of 38 grams or 4 to 5 standard drink units per week (Kaner 2007). These interventions typically comprise a conversation of anywhere between 5 and 45 minutes, include an initial screening process to identify individuals who are experiencing alcohol related risk or harm, provide personalised feedback on alcohol use and harms, identify high risk situations for drinking and coping strategies, suggest strategies to increase motivation for positive behaviour change, and develop a personal plan to reduce drinking. Face to face brief interventions delivered by health professionals have been in use for decades (O'Donnell 2013), but more recently technological innovations have allowed people to interact directly via their computer, mobile device or smart phone with digital interventions designed to address problem alcohol consumption using some of the same ingredients (Khadjesari 2011).
How the intervention might work
Digital interventions for alcohol consumption include some of the same features as face to face interventions (e.g. personalised feedback, engaging the participant in creating coping strategies and goal‐based plans) to motivate the participant to reduce their alcohol consumption over time.
Face to face brief interventions have been found to be effective on average (Kaner 2007), but various differences should be considered when translating these interventions to a digital medium:
-
Setting: most of the cited evidence on face to face brief interventions (Kaner 2007) comes from primary care, although there is a growing literature on other health settings, such a general hospitals (McQueen 2011). However, screening for hazardous and harmful alcohol consumption may not reliably take place in busy healthcare settings or may not effectively identify all those with problems. Barriers to implementation of alcohol interventions (McAvoy 2001) include excessive drinkers not even attending primary care settings, and practitioners being too busy to engage in this work (Wilson 2011). Digital interventions have been proposed as a means of accessing ‘hard to reach’ groups outside health settings, and also of providing a cheaper alternative to interventions delivered within health settings (Kaner 2011).
-
Modality: digital interventions differ considerably in their modality or delivery mechanism, which may present advantages and disadvantages. Some individuals may find disclosing excessive alcohol consumption easier if they feel anonymous but it is also possible that intervention outcomes may be due, at least in part, to therapist effects; greater outcome effects have been reported for physician delivery compared to other practitioners (Sullivan 2011). It is also plausible that a smart phone app which can be used anywhere and at any time at the owner’s discretion may produce a different effect to a specific computer sited in a primary care practice. despite the actual content being very similar.
-
Timing: published evidence suggests that alcohol intervention effects may decay over time for face to face brief interventions (Moyer 2002), which may also apply to digital interventions. Nevertheless, the scope for repeated intervention may potentiate initial effects. Whereas a face to face intervention is often delivered as a one‐off event (although there can be several sessions), digital interventions may be engaged with one‐off or more frequently and regularly over an extended time period.
-
Population: differences in effectiveness may arise for different population groups due to variations in enthusiasm for (e.g. technophilia versus technophobia) or access to technology, for example by age, gender, ethnicity, or socio‐economic status.
Interventions aimed at reducing alcohol consumption are complex in that they are usually made up of several behaviour change techniques (BCTs) and may incorporate several stages. Most brief interventions incorporate a FRAMES approach which includes: giving Feedback on the person's intake, impressing the Responsibility for change onto them, offering Advice, listing a Menu of options, having Empathy, and building Self‐efficacy (Miller 1994).
In order to identify the ‘active ingredients’ within interventions, it is important to document the component BCTs using a reliable method. For example, an analysis of brief interventions, based on the trials in the aforementioned Cochrane review (Kaner 2007) and using a reliable taxonomy of BCTs, has recently identified self‐monitoring as an effective component of these health promoting approaches (Michie 2012).
Modelling work based on published studies to date has suggested that a programme of face to face brief interventions rolled out in primary care would be cost effective compared to no programme, providing additional health benefits at reduced health service cost (Purshouse 2013). Little has yet been published on the cost effectiveness of digital alcohol interventions, although one study (Blankers 2012) suggests that internet‐based therapy (including a therapist) is more cost‐effective than internet self‐help. A question remains on the relative cost effectiveness of digital versus face to face interventions.
Why it is important to do this review
A recent review of reviews (Kaner 2012) has identified a large and relatively well‐designed research literature with around 35 published trials in this field (Carey 2009; Khadjesari 2011; Rooke 2010; White 2010). This body of work included the use of technology to deliver alcohol interventions in social care, education and other community‐based settings as well as via the internet or mobile phone applications. This review will update previous reviews from a public health prevention perspective – it will focus on community‐dwelling individuals who are not seeking formal treatment for alcohol‐related problems but nonetheless are drinking at a level which may cause them risk or harm, who engage with any digitally delivered intervention designed to address alcohol consumption. We will not restrict by type of digital intervention so as to capture all interventions targeting this population, and so as to include interventions which take place on multiple platforms (for example text prompts to use smart phone apps). Interventions are an established part of public health policy (for example UK Government 2012) and this is a fast‐moving field, so it is crucial to keep the evidence base up to date.
Objectives
The main objective is to assess the effectiveness and cost effectiveness of digital interventions for reducing hazardous and harmful alcohol consumption and/or alcohol‐related problems in community‐dwelling populations.
We envisage two comparator groups: (1) no intervention (or minimal input) controls; and (2) another active intervention for delivering preventive advice or counselling to reduce hazardous or harmful alcohol consumption. Specifically, we will address two questions: (1) Are digital interventions superior to no intervention (or minimal input) controls? This question is important for individuals accessing interventions through their own motivation or interest. These individuals will be unlikely to experience active practitioner input and it is important to understand whether digital interventions are better than general material they might seek out on the internet or via mobile phone‐based apps etc. (2) Are digital interventions at least equally effective as face‐to‐face brief alcohol interventions? Practitioner delivered brief interventions are generally accepted to be the best alternative in secondary preventive care in health, workplace, educational or community settings. However, time constraints can impede face‐to‐face delivery of such interventions and it is important to know whether digitally provided input can yield comparable effects to interventions delivered by trained practitioners. We will also identify the most effective component behaviour change techniques of such interventions and their mechanisms of action.
Secondary objectives are as follows:
-
To assess whether outcomes differ between trials where the digital intervention targets participants attending health, social care, education or other community‐based settings and those where it is offered remotely via the internet or mobile phone platforms;
-
To develop a taxonomy of interventions according to their mode of delivery (e.g. functionality features) and assess their impact on outcomes;
-
To identify theories or models that have been used in the development and/or evaluation of the intervention – this will inform intervention development work.
Methods
Criteria for considering studies for this review
Types of studies
We will include randomised controlled trials with individual, cluster, stepped wedge, and n‐of‐1 designs; initial scoping activity has identified a relatively large number of randomised controlled trials in this area.
Types of participants
Participants must be community‐dwelling individuals who have personally sought out or been directed towards any digital intervention including web‐based, mobile phone text messaging, smart phone apps, social networking, or ‘stand alone’ computer‐based technologies (including CD‐ROMs). Participants may be recruited in a range of settings, including primary health care (including emergency departments), social care, educational, workplace or community, and there is no restriction on where participants may interact with the intervention, given that it may be delivered through mobile devices. Recipients of interventions will have been identified by themselves, significant others or via a screening process as hazardous or harmful drinkers and/or have experienced problems as a result of their drinking behaviour.
Studies will be excluded if they are directed mainly towards people who are seeking specialist health or social care treatment for their alcohol consumption, or if they deliver the intervention in a secondary or tertiary care setting.
Types of interventions
-
The intervention must be digital, defined as being delivered primarily through a programmable computer or mobile device (laptop, phone, or tablet), and must respond to user input and generate personalised content which aims to change the participants’ alcohol‐related behaviours. Interventions which do not generate feedback or other output based on the personal characteristics of the user will not be included (for example, generic educational interventions). Interventions are not restricted to those accessible online.
-
The comparator condition may be no intervention, usual care (in a health or social care setting), or other digital or face to face brief intervention to reduce alcohol consumption or harm.
Types of outcome measures
Listed here are outcomes of interest; if a study contains none of these outcomes, it will be excluded. We will assess outcome on the basis of the behaviour change techniques (BCTs) incorporated in the interventions, their theoretical underpinning, and mechanisms of action as reported elsewhere (Webb 2010).
Primary outcomes
Many types of outcome measures are available in the alcohol literature. Our primary outcome will be quantity of alcohol consumed, which may be reported in standard drinks, alcohol units or similar, and which we will convert into grams of alcohol. We will consider trials reporting outcomes at 1 month or more, but we will separate trials according to follow‐up time: less than 6 months, 6 to 12 months, and more than 12 months.
Secondary outcomes
-
Other measures of consumption (e.g. number of binge episodes, frequency of drinking occasions, number of participants exceeding limits as defined by study authors);
-
Indices of alcohol‐related harm or social problems to the drinkers or affected others;
-
Cost effectiveness;
-
Any reported adverse effects.
Search methods for identification of studies
The following sources of information will be used to capture studies for the review. The search will not be limited by publication status, language or date (some digital interventions, such as CD‐ROMs, could go back decades).
Electronic searches
We will search the following databases. An example search strategy is given in Appendix 1.
-
MEDLINE (Ovid) 1946 to present
-
The Cochrane Library (Wiley) – including Drugs & Alcohol Group Specialised Register, CENTRAL (Cochrane Central Register of Controlled Trials), DARE (systematic reviews), HTA (health technology assessments)
-
CINAHL (EBSCO) 1981 to present
-
PsycINFO (Ovid) 1967 to present
-
ERIC (EBSCO) 1966 to present
-
SCI (Science Citation Index via Web of Knowledge) 1970 to present
-
CPCI‐S (Conference Proceedings via Web of Knowledge) 1990 to present
-
Index to Theses
-
Clinicaltrials.gov
-
WHO International Clinical Trials Registry Platform (ICTRP)
-
Google Scholar
We will also search relevant websites which are likely to contain evaluations of digital brief interventions, such as:
-
International Alcohol Information Database (IAID) http://www.drinksresearch.org/
-
Beacon 2.0 https://beacon.anu.edu.au/
-
SAMHSA (Substance Abuse and Mental Health Services Administration)
-
NREPP (National Registry of Evidence‐based Programs and Practices) http://nrepp.samhsa.gov/Index.aspx
-
Drug and Alcohol Findings http://findings.org.uk/
Searching other resources
We will check reference lists of all included studies and relevant reviews, carry out citation searches for included studies, and consult experts to confirm nothing has been missed.
Data collection and analysis
Selection of studies
Two researchers will independently screen all titles and abstracts identified, using Endnote to ensure consistency in screening approach. The full research papers of any studies identified as being potentially eligible will be reviewed by two researchers independently. Any discrepancies will be resolved by discussion and by consulting a third researcher if necessary to reach consensus. A kappa statistic will be calculated at each stage to assess agreement between researchers.
Data extraction and management
A standardised data extraction form will be developed and piloted. This will be used by two researchers independently to carry out data extraction of all included studies. Discrepancies between researchers will be resolved by a third researcher. Relevant data pertaining to patient and intervention characteristics (including mode of delivery and costs), sample sizes, outcome measures (including standard deviations or related measures of variability), and trial characteristics which allow quality assessment will be extracted from included studies onto the piloted data extraction form.
In order to identify the ‘active ingredients’ within interventions, we will code all interventions in terms of their component BCTs using a reliable taxonomy developed for specifying the content of brief interventions for excessive alcohol use (Michie 2012).
The mechanisms of action by which interventions have their effect will be investigated by documenting theories cited by authors as informing the interventions. Both the name of the theory and the extent to which it has been applied in designing or evaluating the intervention will be documented, and the latter will be investigated using the 19‐item Theory Coding Scheme (Michie 2010). This specifies theory use in six areas: reference to underpinning theory, targeting of relevant theoretical constructs, using theory to select recipients or tailor interventions, measurement of constructs, testing of mediation effects and refining theory. This will not only illustrate the extent to which theory is applied but also associations between type of theory and theory use and the effectiveness of the intervention (its usefulness has been demonstrated in, for example, a meta‐analysis investigating this in interventions aimed at increasing physical activity and healthy eating, which found weak relationships between theory type and use and effectiveness (Prestwich 2013)). The findings from this analysis can be used to improve future interventions through a better understanding of the mechanisms of action and theoretical frameworks used in effective interventions.
Assessment of risk of bias in included studies
Methodological quality will be assessed independently by two researchers using the criteria recommended in the Cochrane Handbook (Higgins 2011). The recommended approach for assessing risk of bias in studies included in Cochrane reviews is a two‐part tool, addressing seven specific domains, namely sequence generation and allocation concealment (selection bias), blinding of participants and providers (performance bias), blinding of outcome assessor (detection bias), incomplete outcome data (attrition bias), selective outcome reporting (reporting bias), and other sources of bias. The first part of the tool involves describing what was reported to have happened in the study. The second part of the tool involves assigning a judgement relating to the risk of bias for that entry, in terms of low, high or unclear risk. To make these judgments we will use the criteria indicated by the Handbook adapted to the addiction field (see Appendix 2 for details).
The domains of sequence generation and allocation concealment (avoidance of selection bias) will be addressed in the tool by a single entry for each study. Blinding of participants, personnel and outcome assessor (avoidance of performance bias and detection bias) will be considered separately for objective outcomes (e.g. drop out, use of substance of abuse measured by urine analysis, subjects relapsed at the end of follow‐up, subjects engaged in further treatments) and subjective outcomes (e.g. duration and severity of signs and symptoms of withdrawal, patient self‐reported use of substance, side effects, social functioning as integration at school or at work, family relationship). Incomplete outcome data (avoidance of attrition bias) will be considered for all outcomes except for the drop out from the treatment, which is very often the primary outcome measure in trials on addiction.
'Risk of bias' assessments will be used to carry out sensitivity analyses (see Sensitivity analysis).
Measures of treatment effect
In the outcome assessment, for continuous variable outcomes (e.g. quantity of alcohol consumed) we will compare mean differences, and for dichotomous outcomes (e.g. participants classified as binge drinker, or drinking over set limits) we will compare proportions using relative risks. Where outcomes have been assessed at more than one time, data for each time point will be extracted. Attention is likely to be focused on outcomes at 6 months and 12 months post‐intervention, although this may depend on the number of trials that have reported data at these times.
Unit of analysis issues
For trials with more than one – and very similar ‐ control arms, the results for these arms will be combined in the meta‐analysis. The same approach will be used for very similar treatment arms.
If all arms in a multi‐arm trial are to be included in the meta‐analysis and one treatment arm is to be included more than once in some comparisons, then we will divide the number of events and the number of participants in that arm by the number of treatment comparisons made. This method avoids the multiple use of participants in the pooled estimate of treatment effect while retaining information from each arm of the trial. It compromises the precision of the pooled estimate slightly.
Cluster randomised trials will be eligible for inclusion in the meta‐analysis. If the analysis in a trial report accounts for the cluster design, we will assign imputed standard deviations to the treatment and control groups such that the standard error of the treatment effect estimated by the weighted mean difference method in RevMan is the same as the standard error of the treatment effect as reported in analysis which allowed for clustering. If the analysis in a trial report does not account for the cluster design, we will add an external estimate of the intra‐cluster coefficient (ICC) to estimate a design effect, thus inflating the variance of the effect estimate. Then we can enter the data into RevMan and combine the cluster randomised trials with individually randomised trials in the same meta‐analysis.
Dealing with missing data
We will contact authors to try to obtain missing data. Where this is impossible, we will attempt to estimate primary outcome measures using secondary outcome measures; for example, estimating quantity of alcohol consumed using frequency and intensity of consumption. Trials with missing standard deviations will be excluded from the main analysis for the associated continuous measure, but may be included in a sensitivity analysis, using imputed values for the standard deviations.
Assessment of heterogeneity
The magnitude of heterogeneity will be assessed using the I2 statistic, and the statistical significance of the heterogeneity will be assessed using P values derived from Chi2 tests (Deeks 2001). Heterogeneity will be explored both narratively and using subgroup and sensitivity analyses. Clinical heterogeneity is likely, due not only to the variation in delivery methods but to aspects of content.
Assessment of reporting biases
We will note whether studies appear to have incomplete reporting bias. We will have made every effort to minimise publication bias by searching a wide range of databases and sources of grey literature and not restricting by language or publication status, but we will use funnel plots to assess the potential for bias related to the size of the trials, which may indicate publication bias.
Data synthesis
If studies are sufficiently homogeneous to enable meta‐analysis, we will pool the data for each outcome using a random‐effects model in a meta‐analysis that compares intervention and control arms, using mean differences for continuous variables and relative risks for dichotomous outcomes. The meta‐analysis will be performed using RevMan. If meta‐analysis is not feasible we will carry out a narrative summary of studies.
Where possible, analysis will consider key population groups such as men versus women, older versus younger, and different socio‐economic groups.
We will estimate long‐term cost‐effectiveness of strategies for the use of internet, mobile phone text messaging, smart phone app interventions or computer‐based technologies if data allow, by adapting the current Sheffield Alcohol Policy Model (SAPM) analysis of screening and brief interventions, which was developed to inform National Institute for Health and Care Excellence (NICE) public health guidance for England.
If there are sufficient data for analysis we can identify effective BCTs using meta‐regression and theoretical combinations of BCTs using Classification and Regression Trees. This will help to identify the mechanisms of action of effective interventions to inform future development of interventions.
Subgroup analysis and investigation of heterogeneity
If there are sufficient studies, subgroup analyses will be carried out based on:
-
Intervention modality (functionality and setting): to capture potential differences caused by different delivery mechanisms and settings for the intervention outside of the actual content of the intervention;
-
Timing of outcomes (intermediate versus delayed): to investigate possible delay over time;
-
Component BCTs (Michie 2012) as a comparison for face to face brief interventions;
-
Theoretical basis of the interventions;
-
Key population subgroups, such as age, gender, ethnicity, and socio‐economic status.
Sensitivity analysis
We will conduct sensitivity analyses by investigating the effect of omitting studies with a high risk of bias.
