We explore the potential of Bayesian hierarchical modelling for the analysis of cluster randomized trials with binary outcome data, and apply the methods to a trial randomized by general practice. An approximate relationship is derived between the intracluster correlation coefficient (ICC) and the between-cluster variance used in a hierarchical logistic regression model. By constructing an informative prior for the ICC on the basis of available information, we are thus able implicitly to specify an informative prior for the between-cluster variance. The approach also provides us with a credible interval for the ICC for binary outcome data. Several approaches to constructing informative priors from empirical ICC values are described. We investigate the sensitivity of results to the prior specified and find that the estimate of intervention effect changes very little in this data set, while its interval estimate is more sensitive. The Bayesian approach allows us to assume distributions other than normality for the random effects used to model the clustering. This enables us to gain insight into the robustness of our parameter estimates to the classical normality assumption. In a model with a more complex variance structure, Bayesian methods can provide credible intervals for a difference between two variance components, in order for example to investigate whether the effect of intervention varies across clusters. We compare our results with those obtained from classical estimation, discuss the relative merits of the Bayesian framework, and conclude that the flexibility of the Bayesian approach offers some substantial advantages, although selection of prior distributions is not straightforward.

Unintentional injury is a leading cause of mortality and disability among young and old. While evidence about the effectiveness of interventions in reducing injuries is accumulating, reviews of this evidence frequently fail to include details of implementation processes. Our research, of which the work reported here formed a part, had two main objectives: (1) to identify evidence about the implementation of interventions aimed at reducing unintentional injuries amongst children and young people; and (2) to explore methods for systematically reviewing evidence on implementation. Existing systematic reviews of the effectiveness of interventions aiming to reduce unintentional injuries in children and young people formed the starting point for the work reported here. In summary, many of the published papers we identified contained little information on implementation processes and, even when these were discussed, the extent to which authors' claims were based on research evidence was unclear. On the basis of the studies we reviewed implementation data were insufficiently strong to provide a sound evidence base for practitioners and policymakers. Notwithstanding this, we identified valuable data about the context in which such initiatives are implemented and the type of factors that might impinge on implementation. This work has implications in three areas: (1) researchers with an interest in evidence-based public health could be encouraged to consider implementation issues in the design of intervention studies; (2) funding bodies could be encouraged to prioritise intervention studies using mixed methods that will enable researchers to consider effectiveness and implementation; (3) journal editors could work towards increasing the quality of reporting on implementation issues through the development of guidelines.

Titles & abstract s Titles & abstracts Include N=54 Include N=121 Include N=89 Exclude N = 0 Exclude N = 1,024 Exclude N = 21

This study is the first known to use network meta-analysis to simultaneously evaluate the effectiveness of interventions to increase the prevalence of functioning smoke alarms in households with children. The authors identified 24 primary studies from a systematic review of reviews and of more recently published primary studies, of which 23 (17 randomized controlled trials and 6 nonrandomized comparative studies) were included in 1 of the following 2 network meta-analyses: 1) possession of a functioning alarm: interventions that were more "intensive" (i.e., included components providing equipment (with or without fitting), home inspection, or both, in addition to education) generally were more effective. The intervention containing all of the aforementioned components was identified as being the most likely to be the most effective (probability (best) = 0.66), with an odds ratio versus usual care of 7.15 (95% credible interval: 2.40, 22.73); 2) type of battery-powered alarms: ionization alarms with lithium batteries were most likely to be the best type for increasing functioning possession (probability (best) = 0.69). Smoke alarm promotion programs should ensure they provide the combination of interventions most likely to be effective.

This paper discusses some of the processes for establishing a large cluster-randomized trial of a community and primary care intervention in 16 local government areas in Victoria, Australia. The development of the trial in terms of design factors such as sample size estimates and the selection and randomization of communities to intervention or comparison is described. The intervention program to be implemented in Program of Resources, Information and Support for Mothers (PRISM) was conceived as a whole community approach to improving support for all mothers in the first 12 months after birth. A cluster-randomized trial was thus the design of choice from the outset. With a limited number of communities available, a matched-pair design with eight pairs was chosen. Sample size estimates, adjusting for the cluster randomization and the pair-matched design, showed that with eight pairs, on average, 800 women from each community would need to respond to provide sufficient power to determine a 3% reduction in the prevalence of maternal depression 6 months after birth-a reduction deemed to be a worthwhile impact of the intervention to be reliably detected at 80% power. The process of selecting suitable communities and matching them into pairs required careful collection of data on numbers of births, size of the local government areas (LGAs), and an assessment of the capacity of communities to implement the intervention. Ways of dealing with boundary issues associated with potential contamination are discussed. Methods for the selection of feasible configurations of sets of pairs and the ultimate allocation to intervention or comparison are provided in detail. Ultimately, all such studies are a balancing act between selecting the minimum number of communities to detect a meaningful outcome effect of an intervention and the maximum size budget and other resources allow.

This mini-review examines whether accident prevention education reduces the incidence of home accidents among children under the age of 5 years. The Cochrane Library, Medline, CINAHL and Embase databases were searched for randomized controlled studies (RCTs) and systematic reviews of RCTs that compared parents who received accident prevention education with those that did not. Only two RCTs met the inclusion criteria for analysis in this review. The results of these studies showed that there was no statistically significant difference in accidents between the two groups but the quality of the trials was low. An RCT in process (Watson et al, 2002) looks likely to provide evidence of a high standard, so that practice decisions regarding accident prevention education can be made.

There is now widespread acknowledgement of the absence of a sound evidence base underpinning many of the decisions made in primary care. Randomised controlled trials represent the methodology of choicefor determining efficacy and effectiveness of interventions, yet researchers working in primary care have been reluctant to use intervention studies, favouring observational study designs. Unfamiliarity with the different trial designs now available, and the relative advantages and disadvantages conferred by each, may be one factor contributing to this paradox. In this paper, we consider the principal trial designs available to primary care researchers, discussing the contexts in which a particular design may prove most useful. This information will, we hope, also prove useful to primary care clinicians attempting to interpret trial findings.

There is a known association between social deprivation and risk of death from unintentional injury in childhood. In the UK context, these inequalities do not appear to be decreasing. This paper reports on the findings of a systematic review of the world literature between 1975 and 2000 on the prevention of childhood injuries, with particular reference to social deprivation. Literature was identified via electronic databases, key journals and informants. All papers were read independently by at least two reviewers and information was extracted using a standardized form. Results indicate that of 155 studies identified in the systematic review, 32 addressed the issue of social deprivation. The way social deprivation was defined in different studies varied considerably. The literature was not evenly spread across different injury types and did not reflect the burden of injury. There is a paucity of evidence relating to the prevention of child pedestrian injury. Very few studies examined the impact of interventions in different social groups. Without such evidence, it remains difficult for those involved in health promotion to know how to design and target interventions to address inequalities in child injury rates.

The objective was to assess the impact of new guidance on the conduct of narrative synthesis in systematic reviews of effectiveness, by means of a blinded comparison of guidance-led narrative synthesis against a meta-analysis of the same study data.The conclusions of the two syntheses were broadly similar. However, differences between the approaches meant that conclusions about the impact of moderators of effect appeared stronger when derived from the meta-analysis, whereas implications for future research appeared more extensive when derived from the narrative synthesis. These findings emphasize that a rigorously conducted narrative synthesis can add meaning and value to the findings of meta-analysis.The guidance framework provided a useful vehicle for structuring a narrative synthesis and increasing transparency and rigour of the process.While there may be risks with overinterpretation of study data, the framework, tools and techniques described in the guidance appear to increase the transparency and reproducibility of narrative synthesis.

The objective was to assess the impact of new guidance on the conduct of narrative synthesis in systematic reviews of effectiveness, by means of a blinded comparison of guidance-led narrative synthesis against a meta-analysis of the same study data. The conclusions of the two syntheses were broadly similar. However, differences between the approaches meant that conclusions about the impact of moderators of effect appeared stronger when derived from the meta-analysis, whereas implications for future research appeared more extensive when derived from the narrative synthesis. These fi ndings emphasize that a rigorously conducted narrative synthesis can add meaning and value to the fi ndings of meta-analysis. The guidance framework provided a useful vehicle for structuring a narrative synthesis and increasing transparency and rigour of the process. While there may be risks with overinterpretation of study data, the framework, tools and Evaluation Copyright © 2009 SAGE Publications (Los Angeles, London, New Delhi and Singapore) DOI: 10.1177/1356389008097871 Vol 15(1): 047–071

The objective of this study was to assess the receipt of a variety of injury- prevention interventions by socio-demographic characteristics. The families of children aged 3-12 months registered with 18 general practices across Nottingham comprised the intervention groups for the trial (n = 1100). The package of interventions included: standardised safety advice at child-health surveillance consultations (CHS) at 6-9, 12-15 and 18-24 months; provision of low-cost stair-gates, fireguards, smoke alarms and cupboard locks to families in receipt of means-tested benefits; standardised home-safety checks and free first-aid training. A cross-sectional questionnaire survey of socio-demographic characteristics was followed by assessment of whether interventions had been received during a two-year period. Seventy-eight per cent (n = 857) of families received at least one intervention. Safety advice at CHS was less likely to be received by families without access to a car and those belonging to an ethnic group other than white. Home safety checks were less likely to be received by families in receipt of benefits and families whose children had suffered an injury, First-aid training was less likely to be received by families living in rented accommodation. Low-cost safety equipment was more likely to be received by families living in rented accommodation and those receiving benefits. Injury prevention interventions are not accessed equally by all socio- demographic groups, and the provision of low-cost safety equipment may help to reduce this inequality, Inequality in receipt of safety advice at child health surveillance by ethnic group has been demonstrated and further work is required to address the reasons for this. Targeting injury prevention towards this group may be required.

The methodology described here was developed for a systematic review and individual participant-level meta-analysis of home safety education and the provision of safety equipment for the prevention of childhood accidents. This review had a particular emphasis on exploring whether effectiveness was related to socio-demographic characteristics previously shown to be associated with injury risk. Individual participant data were only made available to us for a proportion of the included studies. This resulted in the need for developing a new methodology to combine the available data most efficiently. Our objective was to develop a (random effects) meta-analysis model that could synthesize both individual-level and aggregate-level binary outcome data while exploring the effects of binary covariates also available in a combination of individual participant and aggregate level data. To add further complication, the studies to be combined were a mixture of cluster and individual participant-allocated designs.A Bayesian model using Markov chain Monte Carlo methods to estimate parameters is described which efficiently synthesizes the data by allowing different models to be fitted to the different study design and data format combinations available. Initially we describe a model to estimate mean effects ignoring the influence of the covariates, and then extend it to include a binary covariate. The application of the method is illustrated by application to one outcome from the motivating home safety meta-analysis for illustration. Using the same general approach, it would be possible to develop further 'tailor made' evidence synthesis models to synthesize all available evidence most effectively.

Objective—The aim of this study was to investigate the validity of self reported safety behaviours as a proxy for injuries in unintentional injuries research. Setting—The study population comprised parents and guardians of children aged 3–12 months in 18 general practices throughout Nottingham (n=764) who responded to a questionnaire on safety practices. Method—Injury data were collected by searching the primary and secondary care records of each child in the study. Safety behaviour was measured by computing a safety practices score from self reported safety practices for each respondent to a postal baseline questionnaire survey of safety behaviours. The score was used to classify families into low, medium, and high risk of injury occurrence. Two further scores were calculated, firstly for those safety practices which required obtaining an item of safety equipment, and secondly those safety practices requiring behavioural change without cost implications. Results—High risk families were no more likely than low risk families to sustain an injury (odds ratio (OR) 1.08; 95% confidence interval 0.65 to 1.79). Medium risk families were also no more likely than low risk families to have a medically attended injury (OR 1.09; 0.73 to 1.61) suggesting no association between safety score and future medically attended injury. Similarly, compared to low risk families, medium risk (OR 0.93; 0.33 to 2.61) and high risk (OR 0.46; 0.08 to 2.43) families were no more likely to have a child admitted to hospital with an injury. There was no correlation between the total number of injuries sustained during the study period and the baseline safety practices score (Spearman's rho = 0.004; p=0.917). Subgroup analyses for safety behaviours requiring passive and active safety measures did not reveal significant associations with injury outcomes. Conclusion—Self reported safety behaviours do not appear to be good predictors of childhood unintentional injuries. Further research is required to ascertain valid proxy outcome measures for injury research.

OBJECTIVE To provide information concerning the magnitude of the intraclass correlation coefficient (ICC) for cluster-based studies set in primary care. STUDY DESIGN AND SETTING Reanalysis of data from 31 cluster-based studies in primary care to estimate intraclass correlation coefficients from random effects models using maximum likelihood estimation. RESULTS ICCs were estimated for 1,039 variables. The median ICC was 0.010 (interquartile range [IQR] 0 to 0.032, range 0 to 0.840). After adjusting for individual- and cluster-level characteristics, the median ICC was 0.005 (IQR 0 to 0.021). A given measure showed widely varying ICC estimates in different datasets. In six datasets, the ICCs for SF-36 physical functioning scale ranged from 0.001 to 0.055 and for SF-36 general health from 0 to 0.072. In four datasets, the ICC for systolic blood pressure ranged from 0 to 0.052 and for diastolic blood pressure from 0 to 0.108. CONCLUSION The precise magnitude of between-cluster variation for a given measure can rarely be estimated in advance. Studies should be designed with reference to the overall distribution of ICCs and with attention to features that increase efficiency.

OBJECTIVE To present the results of an intervention trial to enhance parents' home-safety practices through pediatric safety counseling, home visits, and an on-site children's safety center where parents receive personalized education and can purchase reduced-cost products. DESIGN Pediatricians were randomized to a standard- or an enhanced-intervention group. Parents of their patients were enrolled when the patient was 6 months or younger and observed until 12 to 18 months of age. SETTING A hospital-based pediatric resident continuity clinic that serves families living in low-income, inner-city neighborhoods. PARTICIPANTS First- and second-year pediatric residents and their patient-parent dyads. INTERVENTIONS Parents in the standard-intervention group received safety counseling and referral to the children's safety center from their pediatrician. Parents in the enhanced-intervention group received the standard services plus a home-safety visit by a community health worker. OUTCOMES Home observers assessed the following safety practices: reduction of hot-water temperature, poison storage, and presence of smoke alarms, safety gates for stairs, and ipecac syrup. RESULTS The prevalence of safety practices ranged from 11% of parents who stored poisons safely to 82% who had a working smoke alarm. No significant differences in safety practices were found between study groups. However, families who visited the children's safety center compared with those who did not had a significantly greater number of safety practices (34% vs 17% had > or 3). CONCLUSIONS Home visiting was not effective in improving parents' safety practices. Counseling coupled with convenient access to reduced-cost products appears to be an effective strategy for promoting children's home safety.

Most randomised trials allocate individual participants to different treatments. However, cluster randomised trials in which groups of subjects are allocated to different treatments are becoming increasingly popular.1 Cluster randomisation is often advocated to minimise treatment “contamination” between intervention and control participants. For example, in a trial of dietary change, people in the control group might learn about the experimental diet and adopt it themselves. Contamination of control participants has two related effects. It reduces the point estimate of an intervention's effectiveness and this apparent reduction may lead to a type II error—that is, rejection of an effective intervention as ineffective because the observed effect size was neither statistically nor clinically significant. Although the threat of contamination is an issue in some controlled trials, it may be not be of much practical importance in many. Trialists should use individual randomisation if possible because of the drawbacks of cluster allocation. Cluster trials are associated with problems of recruitment bias and the need for larger samples than would be required in similar, individually randomised trials. In recruitment bias, different sorts of participants are selected into the various arms of the trial, thereby defeating the objective of randomisation, while a larger sample size may increase the cost of a trial, its length, or its complexity. This paper describes the difficulties of cluster trials and argues that the problem of contamination can often be dealt with by individual randomisation. #### Summary points Cluster trials are often used to prevent “contamination” between intervention and control groups Cluster trials are usually very much larger than individually randomised trials and can be susceptible to recruitment bias The problem of contamination can often be overcome by increasing the sample size In terms of total sample size, cluster trials are only more efficient where contamination exceeds 30% Members of clusters cannot be …

Mixed treatment comparisons (MTC) extend the traditional pair‐wise meta‐analytic framework to synthesize information on more than two interventions. Although most MTCs use aggregate data (AD), a proportion of the evidence base might be available at the individual level (IPD). We develop a series of novel Bayesian statistical MTC models to allow for the simultaneous synthesis of IPD and AD, potentially incorporating study and individual level covariates. The effectiveness of different interventions to increase the provision of functioning smoke alarms in households with children was used as a motivating dataset. This included 20 studies (11 AD and 9 IPD), including 11 500 participants. Incorporating the IPD into the network allowed the inclusion of information on subject level covariates, which produced markedly more accurate treatment–covariate interaction estimates than an analysis solely on the AD from all studies. Including evidence at the IPD level in the MTC is desirable when exploring participant level covariates; even when IPD is available only for a fraction of the studies. Such modelling may not only reduce inconsistencies within networks of trials but also assist the estimation of intervention subgroup effects to guide more individualised treatment decisions. Copyright © 2012 John Wiley & Sons, Ltd.

Many of the difficulties encountered in the design, organization and analysis of cluster randomized trials arise from the dual nature of such trials; that is, they focus on both the cluster and the individual. A trial now in progress to compare three methods of promoting secondary prevention of coronary heart disease in primary care includes only 21 general practices, but 2142 patients, and thus contains the problems of both small and large samples. With only seven practices in each arm, the trial demanded carefully restricted randomization, may be difficult to analyse, and risks loss of power if one practice should drop out. At the same time, the large number of patients makes for an expensive and administratively complex study. The simultaneous demands of clarity and thoroughness point to an analysis at both cluster and individual level. With two different approaches, however, there may be difficulties of presentation, even if the results agree, and additional problems of interpretation if they do not. Finally, practical considerations may conflict with theoretical demands. Since the trial contained a service element, all patients with heart disease had to be included, even though it would have been more efficient to take only a sample of patients from some practices.

In the last 10 years there has been considerable progress in the science and art of injury prevention. The scientific study of what works in different fields of health care and health promotion has expanded, and evidence-led policy development has dominated health planning. We have collected evidence on evaluated intervention studies related to childhood injury prevention since 1992, and published reviews in 1993 and 1996.1,2 This paper updates this evidence. We attempt to answer three questions: 1. Have there been any changes in the evidence relating to the effectiveness of childhood injury prevention? 2. What additions have been made to the literature, relating to the target groups and implementation strategies of interventions? 3. What additions have been made to the literature, relating to the ways interventions have been evaluated? A database of primary studies has been built up over the years in the Department of Child Health in the University of Newcastle. We identified the relevant literature by a search of computerised databases (for example, BIDS, Medline, Excerpta Medica, the DHSS database, the Social Science Research Index, Web of Science, Transport Research Laboratory databases). This was supplemented by …

Elsewhere1 we have described the rationale for carrying out cluster randomised controlled trials (CRCTs) in injury prevention and discussed key issues relating to the design and ethical conduct of such studies. In this companion paper we focus on sample size calculations for cluster randomised trials and on the methods for statistical analyses of these studies. As previously reported in our companion paper,1 a major disadvantage of CRCTs is that they generally require a larger sample size than do individually randomised trials. This increase in sample size can be quite substantial, depending on the size of the clusters being randomised and the degree of similarity of outcomes among members of the same cluster. The key measure of this similarity is called the intracluster or intraclass correlation coefficient (ICC), often denoted as ρ . This measure reflects the correlation between outcome values in members of the same cluster. If all members of the same cluster have identical values of the outcome measure, the ICC is equal to 1. An ICC of 0 would indicate that there is no correlation in outcome values between members of the same cluster, such that a member of any particular cluster is likely to have values that are no more similar to those of another member of the same cluster than they are to a member of a different cluster. Negative values of the ICC would occur where outcomes for members in the same cluster are less alike than they are for members in different clusters, but this is unlikely to be the case in CRCTs, although estimates of the ICC may be negative due to sampling error. Assuming that the ICC has only …