Individual‐, family‐, and school‐level interventions targeting multiple‐risk behaviours in young people
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
Primary research objective
To review the effects of interventions implemented up to 18 years of age for the primary or secondary prevention of multiple risk behaviours in individuals aged 8‐25 years.
In this review, multiple risk behaviours are considered as two or more behaviours, but interventions that address certain combinations of behaviours are excluded to avoid overlap with other reviews. These include those that target just two behaviours including unhealthy diet, low levels of physical activity and/or high levels of sedentary behaviour; and those that target a combination of tobacco use, alcohol consumption and/or drug use (see the types of interventions section).
Secondary research objectives
-
To explore whether the effects of the intervention differ within and between population subgroups.
-
To explore whether the effects of the intervention differ by risk behaviour and by outcomes.
-
To explore the influence of the setting of the intervention in the design, delivery and outcomes of the interventions.
-
To explore the relationship between the number and/or types of component(s) of an intervention, duration, and effects of the interventions.
-
To explore whether the impact(s) of interventions differ according to whether behaviours are addressed simultaneously or sequentially and/or whether behaviours are addressed in a particular order.
-
To explore the association between clustering of particular behaviours and the effects of the interventions.
-
To explore the cost‐effectiveness of interventions.
-
To identify the implications of the findings of the review for further research, policy and practice.
Background
Description of the condition
Evidence indicates that risk behaviours such as smoking, antisocial behaviour, alcohol consumption and unprotected sexual intercourse cluster in adolescence and that (Basen‐Engquist 1996; Burke 1997; Junger 2001; Mistry 2009; Pahl 2010; van Nieuwenhuijzen 2009), such behaviours, individually and collectively, are associated with increased risk of poor educational attainment, morbidity and premature mortality (Biglan 2004). Relatively few studies have directly examined the prevalence of multiple, rather than single, risk behaviours, particularly in low‐ and middle‐income countries. However, estimates of the prevalence of concurrent tobacco smoking, drinking of alcohol, and recent illicit drug or cannabis use, for adolescents in the United Kingdom, USA and Canada range from 6% to 13% (Connell 2009; Leatherdale 2010; McVie 2005; NHS 2008; NHS 2010). In the USA, it has been reported that 4% of adolescents engage in five or more behaviours (Duberstein 2000). In the Seychelles the estimated prevalence of concurrent tobacco, alcohol and illicit drug use is 11% among adolescents aged 15‐17 years (Faeh 2006), whilst in Lao People's Democratic Republic 15% of males aged 14‐19 years reported concurrent smoking and alcohol use, with 5.5% reporting tobacco smoking, alcohol use and sexual risk behaviour (a lack of condom use) (Sychareun 2011).
Multiple risk behaviours have been defined as "more than one behaviour directly or indirectly associated with health, well‐being and the healthy development of personality" (Hurrelmann 2006). They are important because the consequences are costly to society, young people who engage in any one risk behaviour are likely to engage in others, engagement in risk behaviours may continue from adolescence into early adulthood (Chen 1995; Mason 2010), and there may be shared biological and environmental factors which influence the development of multiple behaviours such as family influences, mental health, and a sense of positive connection with a school (Beyers 2004; Jackson 2010; Viner 2006). Prevention and treatment interventions may therefore impact on more than one outcome (Biglan 2004).
Description of the intervention
This review will assess the effectiveness of interventions that aim to prevent multiple risk behaviours in young people. The cost‐effectiveness of interventions will also be considered where such information is available. Interventions may be delivered at the level of the individual, family, nursery, preschool, or school. Interventions delivered at a community or population level, such as media campaigns, or policy, regulatory or legislative interventions, will be excluded from this review. However, they will be included in another Cochrane review (Campbell 2012).
The interventions to be included in the review will include those which are both universal and targeted to high‐risk groups implemented in the antenatal period (i.e. before birth) and/or across nursery, preschool, primary and secondary school ages up to age 18 years. The interventions are likely to encompass those using support, visits, motivational interviewing, brief interventions, counselling, education, and training for teachers, parents, children, adolescents, peers, role models, or families. Those delivering the interventions are likely to include nurses, pre‐school staff, teachers, peers, and therapists. Examples of the focus of interventions will be those that attempt to improve parenting skills; communication between teachers or parents and children or adolescents; the behaviour of teachers or parents with children or adolescents; children's problem solving; or adolescents' decision‐making ability and resilience.
How the intervention might work
The determinants of engagement in risk behaviours during adolescence are complex and their antecedents may originate before birth or during the early years (Biglan 2004; Jessor 1991; Kuh 2003). Interventions which influence the early determinants may be more likely to impact on propensity to engage in risk behaviours than interventions which focus on reducing the behaviours or mitigating the harms once the risk behaviours have become established, as the logic model indicates (Figure 1). Interventions providing support to mothers during pregnancy may enhance maternal skills, promote healthy behaviours and promote emotional well‐being which may increase mother‐child interaction and reduce environmental stressors (Biglan 2004; Eckenrode 2010). Interventions during the preschool years, which provide training in parenting or increased pre‐school attendance, may prevent multiple risk behaviours later in life by reducing stressors within the family environment, and by enhancing maternal and child skills (Biglan 2004; Reid 1999; Tremblay 1995). School interventions with young children, which promote adults' (parents and teachers) effective and appropriate use of positive behaviour management, and which promote children's social and cognitive skills, may interrupt the potential for negative family or peer processes which can promote multiple risk behaviours (Biglan 2004; Hawkins 2005).
Logic Model: Interventions to prevent multiple risk behaviours in individuals aged 8‐25 years.
During adolescence, interventions may address multiple risk behaviour by promoting effective parenting practices and family involvement; by improving young people's decision‐making skills, resilience to peer influences, assertiveness and social and life skills; by altering existing norms around risk behaviours; and/or by enhancing teachers' behaviour management capabilities (Biglan 2004; Chen 1995; Jackson 2010; Mason 2010). Multi‐component interventions may promote a number of the family‐, skill‐ or knowledge‐based factors described above. Interactive, skill‐based programmes have been shown to be more effective than knowledge‐ or affective‐based programmes at preventing drug use, which may be because social and psychological factors are relevant in promoting the onset of drug use (Faggiano 2005; Tobler 2000).
Why it is important to do this review
Whilst many health interventions aim to prevent single behaviours (Faggiano 2005; Mytton 2009; Oringanje 2009), little is known about the effectiveness (or cost‐effectiveness) of interventions that aim to prevent multiple risk behaviours (Biglan 2004; Jackson 2010). A number of Cochrane reviews (including published reviews and those currently in development) have focused on specific types of interventions to address single behaviours (Bilby 2008; Fellmeth 2011; Fisher 2008; Fisher 2009; Livingstone 2010), and/or the impact of particular intervention approaches on a range of behavioural outcomes, high risk or clinical populations (Armelius 2007; Furlong 2010; Langford 2011; Littell 2007; Littell 2009; Petrosino 2009; Snedden 2010; Martin 2012). A recent review has also focused on the impact of interventions to address multiple risk behaviours (Jackson 2010). However, no Cochrane or non‐Cochrane systematic reviews have systematically evaluated the effectiveness of universal interventions, implemented at any stage up to age 18 years, designed to prevent a wide range of risk behaviours. This review is therefore broader with respect to the number of behaviours, the settings and the population focus.
Given the limited opportunities and resources to prevent health‐compromising behaviours, it might be more efficient if interventions targeted multiple behaviours. By reviewing evidence relating to the effectiveness and cost‐effectiveness of interventions to prevent multiple risk behaviours and the attendant harms, this review will be useful to public health policy makers and commissioners in assisting with decisions around investment or dis‐investment in particular interventions. In particular, it may provide evidence about appropriate stages at which to intervene to prevent multiple risk behaviours.
Objectives
Primary research objective
To review the effects of interventions implemented up to 18 years of age for the primary or secondary prevention of multiple risk behaviours in individuals aged 8‐25 years.
In this review, multiple risk behaviours are considered as two or more behaviours, but interventions that address certain combinations of behaviours are excluded to avoid overlap with other reviews. These include those that target just two behaviours including unhealthy diet, low levels of physical activity and/or high levels of sedentary behaviour; and those that target a combination of tobacco use, alcohol consumption and/or drug use (see the types of interventions section).
Secondary research objectives
-
To explore whether the effects of the intervention differ within and between population subgroups.
-
To explore whether the effects of the intervention differ by risk behaviour and by outcomes.
-
To explore the influence of the setting of the intervention in the design, delivery and outcomes of the interventions.
-
To explore the relationship between the number and/or types of component(s) of an intervention, duration, and effects of the interventions.
-
To explore whether the impact(s) of interventions differ according to whether behaviours are addressed simultaneously or sequentially and/or whether behaviours are addressed in a particular order.
-
To explore the association between clustering of particular behaviours and the effects of the interventions.
-
To explore the cost‐effectiveness of interventions.
-
To identify the implications of the findings of the review for further research, policy and practice.
Methods
Criteria for considering studies for this review
Types of studies
We will include randomised controlled trials (RCTs), including clustered RCTs aimed at changing at least two risk behaviours of interest. RCTs that primarily assess effectiveness of interventions but which also report findings of a full or partial economic evaluation, or those reporting resource use or costs associated with an RCT intervention, will be included. Studies will only be included if there is a minimum follow‐up period of 6 months, to enable identification of the impact of interventions over the shorter term without excluding studies that were not able to monitor outcomes over a longer time period.
Types of participants
Participants receiving the intervention in the included studies will comprise parents, guardians, carers, teachers, peers, pre‐school, primary‐ or secondary‐school children and young people, and schools. Interventions may include participants in subgroups of the population, but interventions that are aimed at individuals with clinically diagnosed disorders will not be included.
Types of interventions
Interventions will comprise public health improvement programmes that address at least two risk behaviours, including: regular tobacco use; alcohol consumption; recent cannabis or other regular illicit drug use; risky sexual behaviours; anti‐social behaviour and offending; vehicle‐related risk behaviours; self‐harm (without suicidal intent); gambling; unhealthy diet; high levels of sedentary behaviour; and low levels of physical activity. Interventions that address just two risk behaviours including unhealthy diet, low levels of physical activity and/or high levels of sedentary behaviour, will be excluded to overlap with a previous Cochrane systematic review (Waters 2012). In addition, interventions that address two or more risk behaviours including tobacco use, alcohol consumption and/or drug use will be excluded and these will be covered in a separate review. In this way, interventions that target only healthy eating and physical activity, or only tobacco use, alcohol consumption and drug use, for example, will be excluded, but interventions that target healthy eating, physical activity and risky sexual behaviour; or alcohol use, tobacco use and antisocial behaviour will be included.
Comparison will be between those receiving the intervention and those receiving usual practice, no intervention or a placebo. The interventions may be individual, family, or school‐based interventions. They may include psychological, educational, parenting, or environmental interventions. These interventions may be provided universally, without regard to the young people's level of risk, or targeted to particular young people or families identified to be at higher risk. Interventions may start before the onset of behaviours (primary prevention), or may target those engaged in risk behaviours (secondary prevention), but will exclude clinical interventions such as cognitive behavioural therapy. The inclusion of a range of different types of intervention, the exclusion of interventions aimed solely at individuals at higher risk or with clinically diagnosed disorders, and consideration of interventions that address two or more different behaviours, prevents overlap with previously published, or ongoing, Cochrane reviews (Bilby 2008; Bjornstad 2005; Donkoh 2006; Fellmeth 2011; Fisher 2009; Langford 2011; Littell 2009; Livingstone 2010; Petrosino 2009; Ekeland 2006; Woolfenden 2009; MacDonald 2007).
Types of outcome measures
Primary outcomes
The primary outcome is the primary or secondary prevention of two or more risk behaviours in individuals aged 8‐25 years. Since there are currently relatively few studies that examine the epidemiology of multiple risk behaviours, the review will include behaviours that have an adverse impact on health, whether or not the behaviour involves an active desire for 'risk‐taking' or immediate gratification. Risk behaviours such as a lack of UV protection, disordered eating, disordered sleep and engagement in the choking game have been excluded from this review based on available evidence regarding prevalence, adverse impact on health, or relatedness to the behaviours included; or to avoid overlap with, or incorporation of, clinically diagnosed disorders. Consultation with the Centre for the Development and Evaluation of Complex Interventions for Public Health ImpRovement (DECIPHer)'s Public Involvement Advisory Group supported the inclusion of the range of behaviours outlined below.
The risk behaviours to be included relate to:
-
Tobacco use: regular tobacco use
-
Alcohol consumption: binge drinking (alcohol); heavy/hazardous drinking, regular or problem drinking.
-
Drug use: recent cannabis use; recent illicit drug use (other than cannabis), regular illicit drug use.
-
Antisocial behaviour and offending: murder; aggravated assault; sexual assault; violence (including domestic or sexual violence); assault with or without injury; gang fights; hitting a teacher, parent or student; racist abuse; criminal damage; robbery; burglary/breaking and entering; vehicle related theft; prostitution; selling drugs; joy‐riding; carrying a weapon; petty theft; other theft; pan‐handling (begging); buying stolen goods; being noisy and rude; disorderly conduct; nuisance to neighbours; graffiti (Biglan 2004; Hales 2010).
-
Self‐harm: self‐harm without suicidal intent.
-
Gambling: gambling; regular/uncontrolled gambling.
-
Vehicle‐related risk behaviours: cycling without a helmet; not using a car seatbelt, driving under the influence of alcohol, cannabis or illicit substances.
-
Risky sexual behaviours: unprotected sexual intercourse, early debut experience.
-
Activity levels: low levels of physical activity, high levels of sedentary behaviour.
-
Unhealthy diet: low levels of fruit and vegetable consumption, low fibre diet, high fat diet, high sugar diet.
Secondary outcomes
The secondary outcomes include the medium and longer term outcomes which the interventions are aiming to effect.
-
Education and employment: educational qualifications; truancy and school exclusion; employment; not being in education, employment or training (NEET).
-
Crime: Criminal record/offending; re‐offending.
-
Long‐term addictive behaviours: smoking, alcohol, drugs, gambling.
-
Health outcomes: teenage pregnancy or parenthood; sexually transmitted infections; injuries; morbidity (e.g. Hepatitis C, HIV, anxiety and depression, obesity, type II diabetes, fatty liver disease, liver cirrhosis); suicide/self‐harm; premature mortality.
-
Harm associated with the process or the outcomes of the intervention; for instance if the extent of engagement in risk behaviours or adverse health outcomes increase as a result of the intervention
-
Cost effectiveness of the intervention: measures of resource use; costs; or cost‐effectiveness of the intervention (e.g. incremental cost‐effectiveness ratios (ICERs), incremental cost‐per quality adjusted life year (QALY); cost‐benefit ratio)
Limitations and decisions made in this review
In this review, we have included randomised controlled trials delivered at the individual‐, family‐, or school‐level and those studies conducted at the community or population level, such as media campaigns, or policy, regulatory or legislative interventions, will be included within another Cochrane review (Campbell 2012) owing to their distinct study design.
In addition, whilst this review is focused on multiple risk behaviours, defined in this review as two or more behaviours, we have excluded interventions that address certain combinations of these multiple behaviours. First, to avoid overlap with another Cochrane systematic review (Waters 2012), we have excluded interventions that address just two risk behaviours including unhealthy diet, low levels of physical activity and/or high levels of sedentary behaviour. Second, we have excluded those interventions that target only a combination of tobacco use, alcohol consumption and/or drug use and these interventions will be covered in a separate review.
Search methods for identification of studies
Electronic searches
The following databases will be searched:
-
ASSIA
-
Australian Education Index
-
Bibliomap ‐ Database of health promotion research
-
British Education Index
-
Campbell Library
-
CINAHL
-
Clinicaltrials.gov
-
Cochrane Central Register of Controlled Trials (CENTRAL)
-
Dissertation Express ‐ cutdown version of Dissertation abstracts
-
Database of Promoting Health Effectiveness Reviews (DoPHER)
-
Embase
-
ERIC
-
EThOS – British Library electronic theses on line
-
International Bibliography of the Social Sciences ‐ politics & economics
-
MEDLINE
-
PsychINFO
-
Sociological Abstracts
The search strategy that will be used to search MEDLINE can be found in Appendix 1. It will be modified where necessary for the other databases listed. There will be no date or language restrictions. Health economics databases will not be included since only studies that incorporate economic evaluation alongside assessment of the effectiveness of interventions will be included in this review.
Searching other resources
Hand searches of reference lists of relevant articles will be searched to identify additional relevant studies and experts in the field will be contacted to identify ongoing research. Citation searches will be carried out for key studies identified. Websites of organisations actively involved in the prevention of risk behaviour will also be searched, including:
-
World Health Organization
-
UNICEF; United Nations
-
World Bank
-
Centres for Disease Control and Prevention
-
National Institutes of Health
-
National Youth Agency
-
Foundations: Joseph Rowntree, Nuffield Trust.
-
National Criminal Justice Reference Service
-
Policy organisations ‐ Evidence for Policy and Practice Information and Co‐ordinating Centre (EPPI Centre), National Institute for Health and Clinical Excellence (NICE), Scottish Intercollegiate Guidelines Network (SIGN), Department of Health, University of York Centre for Reviews and Dissemination, The King's Fund, Institute for Public Policy Research
Data collection and analysis
Selection of studies
References obtained from literature databases, website searches, and hand searching of reference lists, will be downloaded into a reference management software package and duplicates removed. Papers will be screened according to the titles and abstracts (where available) by a single author, using specific inclusion and exclusion criteria.
The initial screening process will be carried out by two authors for the first 500 publications in the list to assure quality and accuracy of the process. A further 10% of studies selected at random will be double‐screened to ensure that the screening process is consistent and accurate throughout. The full text of selected articles will be obtained and multiple publications from one particular study will be grouped together. Full text articles will also be obtained if additional information is required, to assess eligibility for inclusion.
Full‐text papers will be screened by two authors using a pre‐specified set of criteria for inclusion. Disagreement relating to the inclusion of particular studies will be resolved by discussion or, where disagreements persist, by a third author, to enable a consensus to be reached.
Data extraction and management
A data extraction form will be used independently by two authors to extract data from studies that meet the inclusion criteria for this review. The data extraction form will be piloted by two review authors to assure that it captures study data and assesses study quality effectively. Data to be extracted will include:
-
Lead author, review title or unique identifier and date
-
Eligibility for inclusion
-
Reasons for exclusion
-
Study aim(s)
-
Study design
-
Study location
-
Study setting
-
Theoretical underpinning
-
Context
-
Implementation factors
-
Equity (using PROGRESS Plus (see below for details))
-
Sustainability
-
Intervention (content and activities, number/type of behaviours addressed and whether the behaviours addressed include those motivated by an active desire for risk‐taking or whether they are related to a greater extent to the environmental context, duration of intervention, and details of any intervention offered to the control group)
-
Participants of intervention (including the number randomised and the number in each intervention group; age at start of intervention; and demographic data where possible e.g. ethnicity, gender, socio‐economic status)
-
Scope of the intervention (universal or targeted to high risk or vulnerable group)
-
Proximal or distal nature of the intervention in relation to the behaviours examined
-
Method of measurement of risk behaviour (self‐report or objective measure)
-
Duration of follow up(s)
-
Attrition rate
-
Randomisation
-
Allocation concealment
-
Outcome measures post‐intervention at each stage of follow up (including unit of measurement)
-
Effect size and precision (e.g. 95% confidence interval)
-
Whether clustering was taken into account in cluster RCTs and intra‐cluster correlation coefficient (ICC)
-
Method of analysis
-
Estimates of resource use
-
Source of resource use
-
Estimates of unit costs
-
Source of unit costs
-
Price year
-
Currency
-
Incremental resource use and costs
-
Point estimate and measure of uncertainty for incremental resource use, costs and cost effectiveness
-
Economic analytic viewpoint
-
Time horizon for costs and effects and discount rate
-
Year of study
-
Any other comments
Disagreements around data extraction between the two authors will be resolved by discussion, or by a third author if a consensus cannot be reached by discussion alone.
Where there are multiple reports from the same study, separate data collection forms will be completed for each and information collated from the forms afterwards. The impact of interventions on equity will be identified across a number of categories using PROGRESS plus, an acronym for the following parameters: place of residence; race/ethnicity; occupation; gender; religion; education; social capital; and socio‐economic status; with plus representing the additional categories: age; disability; and sexual orientation. PROGRESS plus factors reported at baseline and at the endpoint will be collected. Data will be entered into Review Manager 2011 and checked by a second author.
Characteristics of studies will be summarised in a table.
Assessment of risk of bias in included studies
Risk of bias will be assessed in included studies using the Cochrane Collaboration's Risk of Bias tool (Higgins 2008). The tool includes assessment of risk of bias in relation to: random sequence generation; allocation concealment; blinding of participants and personnel; blinding of outcome assessment; incomplete outcome data; selective reporting; and other sources of bias. Studies will be scored as at 'high risk', 'low risk' or 'unclear risk' of bias. Risk of bias will be assessed by two review authors. Disagreements between these review authors will be addressed by discussion, and where necessary, a third review will independently assess the study and remaining disagreements resolved by consensus. For economic evaluation the Evers checklist will be used to assess methodological quality (Evers 2005).
Measures of treatment effect
Binary data will be summarised using odds ratios (OR) with 95% confidence intervals (CIs). Continuous outcomes will be handled as the difference in mean values with 95% CIs. However, where continuous outcomes are measured using different measures or scales, standardised mean difference will be calculated. Ordinal outcomes will be reported as binary or continuous outcomes as appropriate.
Unit of analysis issues
The review will include interventions implemented at individual, family, nursery, preschool, or school level. As such, study outcomes may be reported at the group and/or individual level. Review authors will determine whether analysis has taken the effect of clustering into account, for instance if analyses are based on a multilevel model, variance components analysis or generalised estimating equations. Where clustering has not been taken into account, approximately correct analyses will be conducted if data are available regarding: the number of clusters or groups randomised to each intervention group or the average size of each cluster; the outcome data for the total number of individuals ignoring the cluster design; and an estimate of the ICC which provides a measure of the relative variability within and between clusters, as outlined in section 16.3.4 of the Cochrane Handbook (Higgins 2008). Missing data will be requested from study authors wherever possible to enable re‐analysis. Data that has been re‐analysed will be marked as such in the review.
Where there are multiple measures of behaviours e.g. number of sexual partners and condom use or body‐mass index (BMI) and weight, all data will be extracted.
Where there are multiple repeated measurements or recurring events in studies with a long follow‐up period, or the outcome is measured at multiple points in time, data relating to outcome(s) measured at all of the time points included in the study will be extracted.
Where more than two intervention groups are included in the intervention, groups will be combined to create a single pair‐wise comparison, as recommended in section 16.5.4 of the Cochrane Handbook (Higgins 2008).
Dealing with missing data
Study authors will be contacted electronically where there are missing or unclear data (for instance, relating to the primary outcome; or attrition rate). Missing data will be reported in the data extraction form and in the risk of bias table. Study authors will also be contacted if insufficient data are provided to permit intention‐to‐treat analyses. Studies for which insufficient data are available (for instance where missing data cannot be obtained) will be excluded from the review and reasons for exclusion listed in the 'Characteristics of excluded studies' table.
Assessment of heterogeneity
It is anticipated that the studies identified in this review will be heterogeneous with respect to settings, participants, interventions, the risk behaviours addressed, and outcomes analysed. If it is appropriate to combine studies, we will undertake meta‐analyses. Heterogeneity will be examined via inspection of the forest plot and by a chi‐squared test to demonstrate whether the observed differences in results are compatible with chance alone. The I2 statistic will be calculated to examine the percentage of variability that is due to heterogeneity rather than to sampling error. Subgroup analyses will be carried out to investigate heterogeneous results, if the data are available. We will not pool data relating to resource use or cost outcomes across trials but will report such findings in the qualitative synthesis.
Assessment of reporting biases
Funnel plots will be used to plot the study effect size against sample size to assess publication bias, if sufficient studies are identified (a minimum of 10). Publication bias is one of several possible explanations for asymmetry, and these explanations will be discussed in the review.
Data synthesis
Since we anticipate heterogeneity between studies, we will include a structured description and summary of the findings of included studies in the review. The qualitative synthesis will be grouped using categories to be determined when the studies have been identified. The groupings may be type of intervention, length of intervention or type of outcome. Findings relating to the cost‐effectiveness of interventions (e.g. incremental resource use and resource costs; ICERs, incremental cost‐per QALY; or cost‐benefit ratio) will be reported in the qualitative synthesis. We will use the five GRADE considerations (study limitation, consistency of effect, imprecision, indirectness and publication bias) to assess the quality of the body of the evidence and when drawing conclusions. If appropriate to the studies included in the review, we will include a Sumary of Findings Table, including the number of participants and studies for each outcome, the intervention effect and measure of the quality of the body of evidence. The table may include the primary and secondary outcomes.
Where studies are sufficiently similar, we shall undertake meta‐analyses. We will carry out random‐effects meta‐analyses for given outcomes but will compare summary estimates with those from fixed‐effects meta‐analysis. The appropriate method of meta‐analysis will be used depending on the nature of outcome data (dichotomous, ordinal, continuous, time‐to‐event etc) as outlined in chapter 9 of the Cochrane Handbook (Higgins 2008). If outcome data in some studies to be combined are dichotomous and continuous in others, the dichotomous data will be re‐expressed as standardised mean difference if the underlying continuous measurements in each intervention group follow a normal or logistic distribution, enabling data to be pooled. Outcome data combined from different scales will be standardised. Ordinal data will be converted to continuous or dichotomous data as appropriate. Data from cluster randomised trials will only be included in meta‐analyses if clustering has been taken into account or if approximately correct analyses can be undertaken (as outlined above and in sections 16.3.3 and 16.3.4 of the Cochrane Handbook (Higgins 2008)).
Subgroup analysis and investigation of heterogeneity
If there is evidence of heterogeneity amongst the studies we will explore reasons for this. Where sufficient data are available, we will perform subgroup analyses to compare outcomes by:
-
Age group at start of intervention
-
Gender
-
Participants (individual, infant, child, adolescent, parent, guardian, carer, grandparent, teacher, nurse)
-
Number of behaviours targeted
-
Duration of intervention
-
Type of intervention (pre‐ or antenatal, family, pre‐school, school, friendship group; and whether the intervention was universal or targeted to a high‐risk group(s))
-
High‐income or low‐ and middle‐income country
Sensitivity analysis
For meta‐analysis, we will use sensitivity analysis to determine to what extent the overall intervention effect is changed by the inclusion of studies at high or unclear risk of bias. We will also assess whether meta‐analyses are subject to small study bias by carrying out fixed and random‐effects meta‐analyses.
Logic Model: Interventions to prevent multiple risk behaviours in individuals aged 8‐25 years.
