Cohort versus cross-sectional design in large field trials: precision, sample size, and a unifying model

Efficient designs for three-sequence stepped wedge trials with continuous recruitment

BACKGROUND/AIMS

The standard approach to designing stepped wedge trials that recruit participants in a continuous stream is to divide time into periods of equal length. But the choice of design in such cases is infinitely more flexible: each cluster could cross from the control to the intervention at any point on the continuous time-scale. We consider the case of a stepped wedge design with clusters randomised to just three sequences (designs with small numbers of sequences may be preferred for their simplicity and practicality) and investigate the choice of design that minimises the variance of the treatment effect estimator under different assumptions about the intra-cluster correlation.

METHODS

We make some simplifying assumptions in order to calculate the variance: in particular that we recruit the same number of participants, m , from each cluster over the course of the trial, and that participants present at regularly spaced intervals. We consider an intra-cluster correlation that decays exponentially with separation in time between the presentation of two individuals from the same cluster, from a value of ρ for two individuals who present at the same time, to a value of ρ τ for individuals presenting at the start and end of the trial recruitment interval. We restrict attention to three-sequence designs with centrosymmetry - the property that if we reverse time and swap the intervention and control conditions then the design looks the same. We obtain an expression for the variance of the treatment effect estimator adjusted for effects of time, using methods for generalised least squares estimation, and we evaluate this expression numerically for different designs, and for different parameter values.

RESULTS

There is a two-dimensional space of possible three-sequence, centrosymmetric stepped wedge designs with continuous recruitment. The variance of the treatment effect estimator for given ρ and τ can be plotted as a contour map over this space. The shape of this variance surface depends on τ and on the parameter m ρ / ( 1 - ρ ) , but typically indicates a broad, flat region of close-to-optimal designs. The 'standard' design with equally spaced periods and 1:1:1 allocation rarely performs well, however.

CONCLUSIONS

In many different settings, a relatively simple design can be found (e.g. one based on simple fractions) that offers close-to-optimal efficiency in that setting. There may also be designs that are robustly efficient over a wide range of settings. Contour maps of the kind we illustrate can help guide this choice. If efficiency is offered as one of the justifications for using a stepped wedge design, then it is worth designing with optimal efficiency in mind.

Medical Imaging Decision And Support (MIDAS): Study protocol for a multi-centre cluster randomized trial evaluating the ESR iGuide

OBJECTIVES

Medical imaging plays an essential role in healthcare. As a diagnostic test, imaging is prone to substantial overuse and potential overdiagnosis, with dire consequences to patient outcomes and health care costs. Clinical decision support systems (CDSSs) were developed to guide referring physicians in making appropriate imaging decisions. This study will evaluate the effect of implementing a CDSS (ESR iGuide) with versus without active decision support in a physician order entry on the appropriate use of imaging tests and ordering behaviour.

METHODS

A protocol for a multi-center cluster-randomized trial with departments acting as clusters, combined with a before-after-revert design. Four university hospitals with eight participating departments each for a total of thirty-two clusters will be included in the study. All departments start in control condition with structured data entry of the clinical indication and tracking of the imaging exams requested. Initially, the CDSS is implemented and all physicians remain blinded to appropriateness scores based on the ESR imaging referral guidelines. After randomization, half of the clusters switch to the active intervention of decision support. Physicians in the active condition are made aware of the categorization of their requests as appropriate, under certain conditions appropriate, or inappropriate, and appropriate exams are suggested. Physicians may change their requests in response to feedback. In the revert condition, active decision support is removed to study the educational effect.

RESULTS/CONCLUSIONS

The main outcome is the proportion of inappropriate diagnostic imaging exams requested per cluster. Secondary outcomes are the absolute number of imaging exams, radiation from diagnostic imaging, and medical costs.

TRIAL REGISTRATION NUMBER

Approval from the Medical Ethics Review Committee was obtained under protocol numbers 20-069 (Augsburg), B 238/21 (Kiel), 20-318 (Lübeck) and 2020-15,125 (Mainz). The trial is registered in the ClinicalTrials.gov register under registration number NCT05490290.

Analysis of multiple-period group randomized trials: random coefficients model or repeated measures ANOVA?

BACKGROUND

Multiple-period parallel group randomized trials (GRTs) analyzed with linear mixed models can represent time in mean models as continuous or categorical. If time is continuous, random effects are traditionally group- and member-level deviations from condition-specific slopes and intercepts and are referred to as random coefficients (RC) analytic models. If time is categorical, random effects are traditionally group- and member-level deviations from time-specific condition means and are referred to as repeated measures ANOVA (RM-ANOVA) analytic models. Longstanding guidance recommends the use of RC over RM-ANOVA for parallel GRTs with more than two periods because RC exhibited nominal type I error rates for both time parameterizations while RM-ANOVA exhibited inflated type I error rates when applied to data generated using the RC model. However, this recommendation was developed assuming a variance components covariance matrix for the RM-ANOVA, using only cross-sectional data, and explicitly modeling time × group variation. Left unanswered were how well RM-ANOVA with an unstructured covariance would perform on data generated according to the RC mechanism, if similar patterns would be observed in cohort data, and the impact of not modeling time × group variation if such variation was present in the data-generating model.

METHODS

Continuous outcomes for cohort and cross-sectional parallel GRT data were simulated according to RM-ANOVA and RC mechanisms at five total time periods. All simulations assumed time × group variation. We varied the number of groups, group size, and intra-cluster correlation. Analytic models using RC, RM-ANOVA, RM-ANOVA with unstructured covariance, and a Saturated random effects structure were applied to the data. All analytic models specified time × group random effects. The analytic models were then reapplied without specifying random effects for time × group.

RESULTS

Results indicated the RC and saturated analytic models maintained the nominal type I error rate in all data sets, RM-ANOVA with an unstructured covariance did not avoid type I error rate inflation when applied to cohort RC data, and analytic models omitting time-varying group random effects when such variation exists in the data were prone to substantial type I error inflation unless the residual error variance is high relative to the time × group variance.

CONCLUSION

The time × group RC and saturated analytic models are recommended as the default for multiple period parallel GRTs.

Power analysis for stepped wedge trials with multiple interventions

Stepped wedge design (SWD) trials are cluster randomized trials that feature staggered, unidirectional cross-over between treatment conditions. Existing literature on power for SWDs focuses primarily on designs with two conditions, typically a control and an intervention condition. However, SWDs with more than one treatment condition are being proposed and conducted. We present a linear mixed model for SWDs with two or more interventions, including both multiarm and factorial designs. We derive standard errors of the intervention effect coefficients, and present power calculation methods. We consider both repeated cross-sectional and cohort designs. Design features, with a focus on treatment allocations, are examined to determine their impact on power.

A flexible sample size solution for longitudinal and crossover cluster randomized trials with continuous outcomes

Longitudinal cluster randomized trial (LCRT) and crossover cluster randomized trial (CCRT) are two variants of cluster randomized trials. In LCRTs, clusters of subjects are randomly assigned to different treatment groups and each subject has repeated measurements over the study period. In CCRTs, clusters of subjects are randomly assigned to different sequences. Within each sequence, clusters receive all treatments in a particular order. Both LCRTs and CCRTs lead to complicated correlation structures that involve longitudinal and intracluster correlations. Generalized linear mixed model (GLMM) and generalized estimating equation (GEE) approaches have been frequently employed in data analysis and sample size estimation. In this study we propose closed-form sample size and power formulas for LCRTs and CCRTs based on the GEE approach. These formulas are flexible to incorporate unbalanced randomization, different missing patterns, arbitrary correlation structures, and randomly varying cluster sizes, providing a practical yet robust sample size solution. Simulation studies show that the proposed methods achieve good performance with empirical powers and type I errors close to their nominal values.

Intra-cluster correlations from the CLustered OUtcome Dataset bank to inform the design of longitudinal cluster trials

BACKGROUND

Sample size calculations for longitudinal cluster randomised trials, such as crossover and stepped-wedge trials, require estimates of the assumed correlation structure. This includes both within-period intra-cluster correlations, which importantly differ from conventional intra-cluster correlations by their dependence on period, and also cluster autocorrelation coefficients to model correlation decay. There are limited resources to inform these estimates. In this article, we provide a repository of correlation estimates from a bank of real-world clustered datasets. These are provided under several assumed correlation structures, namely exchangeable, block-exchangeable and discrete-time decay correlation structures.

METHODS

Longitudinal studies with clustered outcomes were collected to form the CLustered OUtcome Dataset bank. Forty-four available continuous outcomes from 29 datasets were obtained and analysed using each correlation structure. Patterns of within-period intra-cluster correlation coefficient and cluster autocorrelation coefficients were explored by study characteristics.

RESULTS

The median within-period intra-cluster correlation coefficient for the discrete-time decay model was 0.05 (interquartile range: 0.02-0.09) with a median cluster autocorrelation of 0.73 (interquartile range: 0.19-0.91). The within-period intra-cluster correlation coefficients were similar for the exchangeable, block-exchangeable and discrete-time decay correlation structures. Within-period intra-cluster correlation coefficients and cluster autocorrelations were found to vary with the number of participants per cluster-period, the period-length, type of cluster (primary care, secondary care, community or school) and country income status (high-income country or low- and middle-income country). The within-period intra-cluster correlation coefficients tended to decrease with increasing period-length and slightly decrease with increasing cluster-period sizes, while the cluster autocorrelations tended to move closer to 1 with increasing cluster-period size. Using the CLustered OUtcome Dataset bank, an RShiny app has been developed for determining plausible values of correlation coefficients for use in sample size calculations.

DISCUSSION

This study provides a repository of intra-cluster correlations and cluster autocorrelations for longitudinal cluster trials. This can help inform sample size calculations for future longitudinal cluster randomised trials.

Mixed-effects models for the design and analysis of stepped wedge cluster randomized trials: An overview

The stepped wedge cluster randomized design has received increasing attention in pragmatic clinical trials and implementation science research. The key feature of the design is the unidirectional crossover of clusters from the control to intervention conditions on a staggered schedule, which induces confounding of the intervention effect by time. The stepped wedge design first appeared in the Gambia hepatitis study in the 1980s. However, the statistical model used for the design and analysis was not formally introduced until 2007 in an article by Hussey and Hughes. Since then, a variety of mixed-effects model extensions have been proposed for the design and analysis of these trials. In this article, we explore these extensions under a unified perspective. We provide a general model representation and regard various model extensions as alternative ways to characterize the secular trend, intervention effect, as well as sources of heterogeneity. We review the key model ingredients and clarify their implications for the design and analysis. The article serves as an entry point to the evolving statistical literatures on stepped wedge designs.

A tutorial on sample size calculation for multiple-period cluster randomized parallel, cross-over and stepped-wedge trials using the Shiny CRT Calculator

It has long been recognized that sample size calculations for cluster randomized trials require consideration of the correlation between multiple observations within the same cluster. When measurements are taken at anything other than a single point in time, these correlations depend not only on the cluster but also on the time separation between measurements and additionally, on whether different participants (cross-sectional designs) or the same participants (cohort designs) are repeatedly measured. This is particularly relevant in trials with multiple periods of measurement, such as the cluster cross-over and stepped-wedge designs, but also to some degree in parallel designs. Several papers describing sample size methodology for these designs have been published, but this methodology might not be accessible to all researchers. In this article we provide a tutorial on sample size calculation for cluster randomized designs with particular emphasis on designs with multiple periods of measurement and provide a web-based tool, the Shiny CRT Calculator, to allow researchers to easily conduct these sample size calculations. We consider both cross-sectional and cohort designs and allow for a variety of assumed within-cluster correlation structures. We consider cluster heterogeneity in treatment effects (for designs where treatment is crossed with cluster), as well as individually randomized group-treatment trials with differential clustering between arms, for example designs where clustering arises from interventions being delivered in groups. The calculator will compute power or precision, as a function of cluster size or number of clusters, for a wide variety of designs and correlation structures. We illustrate the methodology and the flexibility of the Shiny CRT Calculator using a range of examples.

Sample size and power calculations for open cohort longitudinal cluster randomized trials

When calculating sample size or power for stepped wedge or other types of longitudinal cluster randomized trials, it is critical that the planned sampling structure be accurately specified. One common assumption is that participants will provide measurements in each trial period, that is, a closed cohort, and another is that each participant provides only one measurement during the course of the trial. However some studies have an "open cohort" sampling structure, where participants may provide measurements in variable numbers of periods. To date, sample size calculations for longitudinal cluster randomized trials have not accommodated open cohorts. Feldman and McKinlay (1994) provided some guidance, stating that the participant-level autocorrelation could be varied to account for the degree of overlap in different periods of the study, but did not indicate precisely how to do so. We present sample size and power formulas that allow for open cohorts and discuss the impact of the degree of "openness" on sample size and power. We consider designs where the number of participants in each cluster will be maintained throughout the trial, but individual participants may provide differing numbers of measurements. Our results are a unification of closed cohort and repeated cross-sectional sample results of Hooper et al (2016), and indicate precisely how participant autocorrelation of Feldman and McKinlay should be varied to account for an open cohort sampling structure. We discuss different types of open cohort sampling schemes and how open cohort sampling structure impacts on power in the presence of decaying within-cluster correlations and autoregressive participant-level errors.

Accounting for a decaying correlation structure in cluster randomized trials with continuous recruitment

A requirement for calculating sample sizes for cluster randomized trials (CRTs) conducted over multiple periods of time is the specification of a form for the correlation between outcomes of subjects within the same cluster, encoded via the within-cluster correlation structure. Previously proposed within-cluster correlation structures have made strong assumptions; for example, the usual assumption is that correlations between the outcomes of all pairs of subjects are identical ("uniform correlation"). More recently, structures that allow for a decay in correlation between pairs of outcomes measured in different periods have been suggested. However, these structures are overly simple in settings with continuous recruitment and measurement. We propose a more realistic "continuous-time correlation decay" structure whereby correlations between subjects' outcomes decay as the time between these subjects' measurement times increases. We investigate the use of this structure on trial planning in the context of a primary care diabetes trial, where there is evidence of decaying correlation between pairs of patients' outcomes over time. In particular, for a range of different trial designs, we derive the variance of the treatment effect estimator under continuous-time correlation decay and compare this to the variance obtained under uniform correlation. For stepped wedge and cluster randomized crossover designs, incorrectly assuming uniform correlation will underestimate the required sample size under most trial configurations likely to occur in practice. Planning of CRTs requires consideration of the most appropriate within-cluster correlation structure to obtain a suitable sample size.

Inference for the treatment effect in multiple-period cluster randomised trials when random effect correlation structure is misspecified

Multiple-period cluster randomised trials, such as stepped wedge or cluster cross-over trials, are being conducted with increasing frequency. In the design and analysis of these trials, it is necessary to specify the form of the within-cluster correlation structure, and a common assumption is that the correlation between the outcomes of any pair of subjects within a cluster is identical. More complex models that allow for correlations within a cluster to decay over time have recently been suggested. However, most software packages cannot fit these models. As a result, practitioners may choose a simpler model. We analytically examine the impact of incorrectly omitting a decay in correlation on the variance of the treatment effect estimator and show that misspecification of the within-cluster correlation structure can lead to incorrect conclusions regarding estimated treatment effects for stepped wedge and cluster crossover trials.

Community mobilization to modify harmful gender norms and reduce HIV risk: results from a community cluster randomized trial in South Africa

INTRODUCTION

Community mobilization (CM) is increasingly recognized as critical to generating changes in social norms and behaviours needed to achieve reductions in HIV. We conducted a CM intervention to modify negative gender norms, particularly among men, in order to reduce associated HIV risk.

METHODS

Twenty two villages in the Agincourt Health and Socio-Demographic Surveillance Site in rural Mpumalanga, South Africa were randomized to either a theory-based, gender transformative, CM intervention or no intervention. Two cross-sectional, population-based surveys were conducted in 2012 (pre-intervention, n = 600 women; n = 581 men) and 2014 (post-intervention, n = 600 women; n = 575 men) among adults ages 18 to 35 years. We used an intent-to-treat (ITT) approach using survey regression cluster-adjusted standard errors to determine the intervention effect by trial arm on gender norms, measured using the Gender Equitable Mens Scale (GEMS), and secondary behavioural outcomes.

RESULTS

Among men, there was a significant 2.7 point increase (Beta Coefficient 95% CI: 0.62, 4.78, p = 0.01) in GEMS between those in intervention compared to control communities. We did not observe a significant difference in GEMS scores for women by trial arm. Among men and women in intervention communities, we did not observe significant differences in perpetration of intimate partner violence (IPV), condom use at last sex or hazardous drinking compared to control communities. The number of sex partners in the past 12 months (AOR 0.29, 95% CI 0.11 to 0.77) were significantly lower in women in intervention communities compared to control communities and IPV victimization was lower among women in intervention communities, but the reduction was not statistically significant (AOR 0.53, 95% CI 0.24 to 1.16).

CONCLUSION

Community mobilization can reduce negative gender norms among men and has the potential to create environments that are more supportive of preventing IPV and reducing HIV risk behaviour. Nevertheless, we did not observe that changes in attitudes towards gender norms resulted in desired changes in risk behaviours suggesting that more time may be necessary to change behaviour or that the intervention may need to address behaviours more directly.

CLINICAL TRIALS NUMBER

ClinicalTrials.gov NCT02129530.

A multi-level intervention in worksites to increase fruit and vegetable access and intake: Rationale, design and methods of the 'Good to Go' cluster randomized trial

BACKGROUND

Fruit and vegetable (F&V) consumption is an important contributor to chronic disease prevention. However, most Americans do not eat adequate amounts. The worksite is an advantageous setting to reach large, diverse segments of the population with interventions to increase F&V intake, but research gaps exist. No studies have evaluated the implementation of mobile F&V markets at worksites nor compared the effectiveness of such markets with or without nutrition education.

METHODS

This paper describes the protocol for Good to Go (GTG), a cluster randomized trial to evaluate F&V intake change in employees from worksites randomized into three experimental arms: discount, fresh F&V markets (Access Only arm); markets plus educational components including campaigns, cooking demonstrations, videos, newsletters, and a web site (Access Plus arm); and an attention placebo comparison intervention on physical activity and stress reduction (Comparison). Secondary aims include: 1) Process evaluation to determine costs, reach, fidelity, and dose as well as the relationship of these variables with changes in F&V intake; 2) Applying a mediating variable framework to examine relationships of psychosocial factors/determinants with changes in F&V consumption; and 3) Cost effectiveness analysis of the different intervention arms.

DISCUSSION

The GTG study will fill important research gaps in the field by implementing a rigorous cluster randomized trial to evaluate the efficacy of an innovative environmental intervention providing access and availability to F&V at the worksite and whether this access intervention is further enhanced by accompanying educational interventions. GTG will provide an important contribution to public health research and practice. Trial registration number NCT02729675, ClinicalTrials.gov.

Impact of non-uniform correlation structure on sample size and power in multiple-period cluster randomised trials

Stepped wedge and cluster randomised crossover trials are examples of cluster randomised designs conducted over multiple time periods that are being used with increasing frequency in health research. Recent systematic reviews of both of these designs indicate that the within-cluster correlation is typically taken account of in the analysis of data using a random intercept mixed model, implying a constant correlation between any two individuals in the same cluster no matter how far apart in time they are measured: within-period and between-period intra-cluster correlations are assumed to be identical. Recently proposed extensions allow the within- and between-period intra-cluster correlations to differ, although these methods require that all between-period intra-cluster correlations are identical, which may not be appropriate in all situations. Motivated by a proposed intensive care cluster randomised trial, we propose an alternative correlation structure for repeated cross-sectional multiple-period cluster randomised trials in which the between-period intra-cluster correlation is allowed to decay depending on the distance between measurements. We present results for the variance of treatment effect estimators for varying amounts of decay, investigating the consequences of the variation in decay on sample size planning for stepped wedge, cluster crossover and multiple-period parallel-arm cluster randomised trials. We also investigate the impact of assuming constant between-period intra-cluster correlations instead of decaying between-period intra-cluster correlations. Our results indicate that in certain design configurations, including the one corresponding to the proposed trial, a correlation decay can have an important impact on variances of treatment effect estimators, and hence on sample size and power. An R Shiny app allows readers to interactively explore the impact of correlation decay.

Understanding the cluster randomised crossover design: a graphical illustraton of the components of variation and a sample size tutorial

BACKGROUND

In a cluster randomised crossover (CRXO) design, a sequence of interventions is assigned to a group, or 'cluster' of individuals. Each cluster receives each intervention in a separate period of time, forming 'cluster-periods'. Sample size calculations for CRXO trials need to account for both the cluster randomisation and crossover aspects of the design. Formulae are available for the two-period, two-intervention, cross-sectional CRXO design, however implementation of these formulae is known to be suboptimal. The aims of this tutorial are to illustrate the intuition behind the design; and provide guidance on performing sample size calculations.

METHODS

Graphical illustrations are used to describe the effect of the cluster randomisation and crossover aspects of the design on the correlation between individual responses in a CRXO trial. Sample size calculations for binary and continuous outcomes are illustrated using parameters estimated from the Australia and New Zealand Intensive Care Society - Adult Patient Database (ANZICS-APD) for patient mortality and length(s) of stay (LOS).

RESULTS

The similarity between individual responses in a CRXO trial can be understood in terms of three components of variation: variation in cluster mean response; variation in the cluster-period mean response; and variation between individual responses within a cluster-period; or equivalently in terms of the correlation between individual responses in the same cluster-period (within-cluster within-period correlation, WPC), and between individual responses in the same cluster, but in different periods (within-cluster between-period correlation, BPC). The BPC lies between zero and the WPC. When the WPC and BPC are equal the precision gained by crossover aspect of the CRXO design equals the precision lost by cluster randomisation. When the BPC is zero there is no advantage in a CRXO over a parallel-group cluster randomised trial. Sample size calculations illustrate that small changes in the specification of the WPC or BPC can increase the required number of clusters.

CONCLUSIONS

By illustrating how the parameters required for sample size calculations arise from the CRXO design and by providing guidance on both how to choose values for the parameters and perform the sample size calculations, the implementation of the sample size formulae for CRXO trials may improve.

Analyzing Organizational Growth in Repeated Cross-Sectional Designs Using Multilevel Structural Equation Modeling

Abstract. In repeated cross-sections of organizations, different individuals are sampled from the same set of organizations at each time point of measurement. As a result, common longitudinal data analysis methods (e.g., latent growth curve models) cannot be applied in the usual way. In this contribution, a multilevel structural equation modeling approach to analyze data from repeated cross-sections is presented. Results from a simulation study are reported which aimed at obtaining guidelines on appropriate sample sizes. We focused on a situation where linear growth occurs at the organizational level, and organizational growth is predicted by a single organizational level variable. The power to identify an effect of this organizational level variable was moderately to strongly positively related to number of measurement occasions, number of groups, group size, intraclass correlation, effect size, and growth curve reliability. The Type I error rate was close to the nominal alpha level under all conditions.

Application of a nonrandomized stepped wedge design to evaluate an evidence-based quality improvement intervention: a proof of concept using simulated data on patient-centered medical homes

Background

Stepped wedge designs have gained recognition as a method for rigorously assessing implementation of evidence-based quality improvement interventions (QIIs) across multiple healthcare sites. In theory, this design uses random assignment of sites to successive QII implementation start dates based on a timeline determined by evaluators. However, in practice, QII timing is often controlled more by site readiness. We propose an alternate version of the stepped wedge design that does not assume the randomized timing of implementation while retaining the method’s analytic advantages and applying to a broader set of evaluations. To test the feasibility of a nonrandomized stepped wedge design, we developed simulated data on patient care experiences and on QII implementation that had the structures and features of the expected data from a planned QII. We then applied the design in anticipation of performing an actual QII evaluation.

Methods

We used simulated data on 108,000 patients to model nonrandomized stepped wedge results from QII implementation across nine primary care sites over 12 quarters. The outcome we simulated was change in a single self-administered question on access to care used by Veterans Health Administration (VA), based in the United States, as part of its quarterly patient ratings of quality of care. Our main predictors were QII exposure and time. Based on study hypotheses, we assigned values of 4 to 11 % for improvement in access when sites were first exposed to implementation and 1 to 3 % improvement in each ensuing time period thereafter when sites continued with implementation. We included site-level (practice size) and respondent-level (gender, race/ethnicity) characteristics that might account for nonrandomized timing in site implementation of the QII. We analyzed the resulting data as a repeated cross-sectional model using HLM 7 with a three-level hierarchical data structure and an ordinal outcome. Levels in the data structure included patient ratings, timing of adoption of the QII, and primary care site.

Results

We were able to demonstrate a statistically significant improvement in adoption of the QII, as postulated in our simulation. The linear time trend while sites were in the control state was not significant, also as expected in the real life scenario of the example QII.

Conclusions

We concluded that the nonrandomized stepped wedge design was feasible within the parameters of our planned QII with its data structure and content. Our statistical approach may be applicable to similar evaluations.

The Effects of an Injury Prevention Program on Landing Biomechanics Over Time

BACKGROUND

Knowledge is limited regarding how long improvements in biomechanics remain after completion of a lower extremity injury prevention program.

PURPOSE

To evaluate the effects of an injury prevention program on movement technique and peak vertical ground-reaction forces (VGRF) over time compared with a standard warm-up (SWU) program.

STUDY DESIGN

Controlled laboratory study.

METHODS

A total of 1104 incoming freshmen (age range, 17-22 years) at a military academy in the United States volunteered to participate. Participants were cluster-randomized by military company to either the Dynamic Integrated Movement Enhancement (DIME) injury prevention program or SWU. A random subsample of participants completed a standardized jump-landing task at each time point: immediately before the intervention (PRE), immediately after (POST), and 2 (POST2M), 4 (POST4M), 6 (POST6M), and 8 months (POST8M) after the intervention. VGRF data collected during the jump-landing task were normalized to body weight (%BW). The Landing Error Scoring System (LESS) was used to evaluate movement technique during the jump landing. The change scores (Δ) for each variable (LESS, VGRF) between the group's average value at PRE and each time point were calculated. Separate univariate analyses of variance were performed to evaluate group differences.

RESULTS

The results showed a greater decrease in mean (±SD) VGRF in the DIME group compared with the SWU group at all retention time points: POST2M (SWU [Δ%BW], -0.13 ± 0.82; DIME, -0.62 ± 0.91; P = .001), POST4M (SWU, -0.15 ± 0.98; DIME,-0.46 ± 0.64; P = .04), POST6M (SWU, -0.04 ± 0.96; DIME, -0.53 ± 0.83; P = .004), and POST8M (SWU, 0.38 ± 0.95; DIME, -0.11 ± 0.98; P = .003), but there was not a significant improvement in the DIME group between PRE and POST8M (Δ%BW, -0.11 ± 0.98). No group differences in Δ LESS were observed.

CONCLUSION

The study findings demonstrated that an injury prevention program performed as a warm-up can reduce vertical ground-reaction forces compared with a standard warm-up but a maintenance program is likely necessary in order for continued benefit.

CLINICAL RELEVANCE

Injury prevention programs may need to be performed constantly, or at least every sport season, in order for participants to maintain the protective effects against injury.

The need to balance merits and limitations from different disciplines when considering the stepped wedge cluster randomized trial design

Background

Various papers have addressed pros and cons of the stepped wedge cluster randomized trial design (SWD). However, some issues have not or only limitedly been addressed. Our aim was to provide a comprehensive overview of all merits and limitations of the SWD to assist researchers, reviewers and medical ethics committees when deciding on the appropriateness of the SWD for a particular study.

Methods

We performed an initial search to identify articles with a methodological focus on the SWD, and categorized and discussed all reported advantages and disadvantages of the SWD. Additional aspects were identified during multidisciplinary meetings in which ethicists, biostatisticians, clinical epidemiologists and health economists participated. All aspects of the SWD were compared to the parallel group cluster randomized design. We categorized the merits and limitations of the SWD to distinct phases in the design and conduct of such studies, highlighting that their impact may vary depending on the context of the study or that benefits may be offset by drawbacks across study phases. Furthermore, a real-life illustration is provided.

Results

New aspects are identified within all disciplines. Examples of newly identified aspects of an SWD are: the possibility to measure a treatment effect in each cluster to examine the (in)consistency in effects across clusters, the detrimental effect of lower than expected inclusion rates, deviation from the ordinary informed consent process and the question whether studies using the SWD are likely to have sufficient social value. Discussions are provided on e.g. clinical equipoise, social value, health economical decision making, number of study arms, and interim analyses.

Conclusions

Deciding on the use of the SWD involves aspects and considerations from different disciplines not all of which have been discussed before. Pros and cons of this design should be balanced in comparison to other feasible design options as to choose the optimal design for a particular intervention study.

Methods for sample size determination in cluster randomized trials

BACKGROUND

The use of cluster randomized trials (CRTs) is increasing, along with the variety in their design and analysis. The simplest approach for their sample size calculation is to calculate the sample size assuming individual randomization and inflate this by a design effect to account for randomization by cluster. The assumptions of a simple design effect may not always be met; alternative or more complicated approaches are required.

METHODS

We summarise a wide range of sample size methods available for cluster randomized trials. For those familiar with sample size calculations for individually randomized trials but with less experience in the clustered case, this manuscript provides formulae for a wide range of scenarios with associated explanation and recommendations. For those with more experience, comprehensive summaries are provided that allow quick identification of methods for a given design, outcome and analysis method.

RESULTS

We present first those methods applicable to the simplest two-arm, parallel group, completely randomized design followed by methods that incorporate deviations from this design such as: variability in cluster sizes; attrition; non-compliance; or the inclusion of baseline covariates or repeated measures. The paper concludes with methods for alternative designs.

CONCLUSIONS

There is a large amount of methodology available for sample size calculations in CRTs. This paper gives the most comprehensive description of published methodology for sample size calculation and provides an important resource for those designing these trials.

The CORE study protocol: a stepped wedge cluster randomised controlled trial to test a co-design technique to optimise psychosocial recovery outcomes for people affected by mental illness in the community mental health setting

INTRODUCTION

User engagement in mental health service design is heralded as integral to health systems quality and performance, but does engagement improve health outcomes? This article describes the CORE study protocol, a novel stepped wedge cluster randomised controlled trial (SWCRCT) to improve psychosocial recovery outcomes for people with severe mental illness.

METHODS

An SWCRCT with a nested process evaluation will be conducted over nearly 4 years in Victoria, Australia. 11 teams from four mental health service providers will be randomly allocated to one of three dates 9 months apart to start the intervention. The intervention, a modified version of Mental Health Experience Co-Design (MH ECO), will be delivered to 30 service users, 30 carers and 10 staff in each cluster. Outcome data will be collected at baseline (6 months) and at completion of each intervention wave. The primary outcome is improvement in recovery score using the 24-item Revised Recovery Assessment Scale for service users. Secondary outcomes are improvements to user and carer mental health and well-being using the shortened 8-item version of the WHOQOL Quality of Life scale (EUROHIS), changes to staff attitudes using the 19-item Staff Attitudes to Recovery Scale and recovery orientation of services using the 36-item Recovery Self Assessment Scale (provider version). Intervention and usual care periods will be compared using a linear mixed effects model for continuous outcomes and a generalised linear mixed effects model for binary outcomes. Participants will be analysed in the group that the cluster was assigned to at each time point.

ETHICS AND DISSEMINATION

The University of Melbourne, Human Research Ethics Committee (1340299.3) and the Federal and State Departments of Health Committees (Project 20/2014) granted ethics approval. Baseline data results will be reported in 2015 and outcomes data in 2017.

TRIAL REGISTRATION NUMBER

Australian and New Zealand Clinical Trials Registry ACTRN12614000457640.

Twenty Years of Neighborhood Effect Research: An Assessment

This paper reviews the magnitude and empirical findings of social epidemiological neighborhood effects research. An electronic keyword literature search identified 1369 empirical and methodological neighborhood effects papers published in 112 relevant journals between 1990 and 2014. Analyses of temporal trends were conducted by focus, journal type (e.g., epidemiology, public health, or social science), and specific epidemiologic journal. Select papers were then critically reviewed. Results show an ever-increasing number of papers published, notably since the year 2000, with the majority published in public health journals. The variety of health outcomes analyzed is extensive, ranging from infectious disease to obesity to criminal behavior. Papers relying on data from experimental designs are thought to yield the most credible results, but such studies are few and findings are inconsistent. Papers relying on data from observational designs and multilevel models typically show small statistically significant effects, but most fail to appreciate fundamental identification problems. Ultimately, of the 1170 empirically focused neighborhood effects papers published in the last 24 years, only a handful have clearly advanced our understanding of the phenomena. The independent impact of neighborhood contexts on health remains unclear. It is time to expand the social epidemiological imagination.

Sample size requirements to detect an intervention by time interaction in longitudinal cluster randomized clinical trials with random slopes

In longitudinal cluster randomized clinical trials (cluster-RCT), subjects are nested within a higher level unit such as clinics and are evaluated for outcome repeatedly over the study period. This study design results in a three level hierarchical data structure. When the primary goal is to test the hypothesis that an intervention has an effect on the rate of change in the outcome over time and the between-subject variation in slopes is substantial, the subject-specific slopes are often modeled as random coefficients in a mixed-effects linear model. In this paper, we propose approaches for determining the samples size for each level of a 3-level hierarchical trial design based on ordinary least squares (OLS) estimates for detecting a difference in mean slopes between two intervention groups when the slopes are modeled as random. Notably, the sample size is not a function of the variances of either the second or the third level random intercepts and depends on the number of second and third level data units only through their product. Simulation results indicate that the OLS-based power and sample sizes are virtually identical to the empirical maximum likelihood based estimates even with varying cluster sizes. Sample sizes for random versus fixed slope models are also compared. The effects of the variance of the random slope on the sample size determinations are shown to be enormous. Therefore, when between-subject variations in outcome trends are anticipated to be significant, sample size determinations based on a fixed slope model can result in a seriously underpowered study.

Optimal Experimental Design With Nesting of Persons in Organizations

This paper introduces optimal design of randomized experiments where individuals are nested within organizations, such as schools, health centers, or companies. The focus is on nested designs with two levels (organization, individual) and two treatment conditions (treated, control), with treatment assignment to organizations, or to individuals within organizations. For each type of assignment, a multilevel model is first presented for the analysis of a quantitative dependent variable or outcome. Simple equations are then given for the optimal sample size per level (number of organizations, number of individuals) as a function of the sampling cost and outcome variance at each level, with realistic examples. Next, it is explained how the equations can be applied if the dependent variable is dichotomous, or if there are covariates in the model, or if the effects of two treatment factors are studied in a factorial nested design, or if the dependent variable is repeatedly measured. Designs with three levels of nesting and the optimal number of repeated measures are briefly discussed, and the paper ends with a short discussion of robust design.

Promoting energy-balance behaviors among ethnically diverse adolescents: overview and baseline findings of The Central Texas CATCH Middle School Project

The Central Texas Coordinated Approach To Child Health (CATCH) Middle School Project is a 3.5-year school-based project aimed at promoting physical activity (PA), healthy eating, and obesity prevention among public middle school students in Texas. This article describes the CATCH intervention model and presents baseline findings from spring 2009. CATCH comprises six core components: CATCH Team, CATCH PE, CATCH Classroom, CATCH Eat Smart Cafeteria, CATCH Family, and CATCH Social Marketing. A group randomized serial cross-sectional design is being employed to test the effect of three program support conditions (n = 10 schools each) on energy-balance behaviors: Basic (training and curriculum only), Basic Plus (training and curriculum plus CATCH facilitator support), and Basic Plus Social Marketing (all inputs plus social marketing component). The study sample is composed of a cross-sectional sample of eighth-grade students (primary outcome evaluation sample) and sixth- and seventh-grade students (PE process evaluation sample) who are selected and measured each year. At baseline, 37.9% of eight-grade students (n = 2,841; 13.9 years) were overweight/obese and 19.2% were obese. Eighth-grade students reported, on average, consuming sugar-sweetened beverages more than two times on the previous day and fruits and vegetables roughly three times on the previous day; only two of five school districts surpassed the recommended 50% cut-point for class time spent in moderate-and-vigorous PA as measured in classes of sixth- and seventh-grade students. Additional behavioral findings are reported. Body mass index and behaviors were comparable across conditions. Baseline findings underscore the need to promote student energy-balance behaviors.

Application of a hierarchical model incorporating intrafamily correlation and cluster effects

BACKGROUND

Research interventions at the family level often include individual- and group-level data that can present an analytic challenge. The study that motivated this article was an intervention study conducted with elementary school children and their parents. Randomization occurred at the school level, with families nested within schools. Repeated measurements collected from children and parents at different time points presented modeling challenges, including how to specify the covariance structure correctly among all measurements.

OBJECTIVES

The aim of this study was to introduce a mixed model with random effects to model the correlations among family members, repeated measures, and the grouping effect.

METHODS

A hierarchical random-effect model was used that included both fixed and random effects; time and intervention-by-time variables were included as fixed effects, the school-specific variable was included as random effect, and the intrafamily correlation was modeled through a spatial autoregression covariance matrix. Comparisons were made between the performance of the proposed modeling method and that of other parsimony models using Akaike's Information Criterion (AIC).

RESULTS

The proposed modeling method produced a 3% and 9% reduction in AIC values, respectively, compared with the two other models. The likelihood ratio test further confirmed that the full model was better than the other two models (p < .0001 for both models).

DISCUSSION

The data suggest that using the proposed mixed model technique will produce a significantly better model fit for intrafamily correlation with a nested study design.

Sample size considerations for GEE analyses of three-level cluster randomized trials

Cluster randomized trials in health care may involve three instead of two levels, for instance, in trials where different interventions to improve quality of care are compared. In such trials, the intervention is implemented in health care units ("clusters") and aims at changing the behavior of health care professionals working in this unit ("subjects"), while the effects are measured at the patient level ("evaluations"). Within the generalized estimating equations approach, we derive a sample size formula that accounts for two levels of clustering: that of subjects within clusters and that of evaluations within subjects. The formula reveals that sample size is inflated, relative to a design with completely independent evaluations, by a multiplicative term that can be expressed as a product of two variance inflation factors, one that quantifies the impact of within-subject correlation of evaluations on the variance of subject-level means and the other that quantifies the impact of the correlation between subject-level means on the variance of the cluster means. Power levels as predicted by the sample size formula agreed well with the simulated power for more than 10 clusters in total, when data were analyzed using bias-corrected estimating equations for the correlation parameters in combination with the model-based covariance estimator or the sandwich estimator with a finite sample correction.

An overview of the "Positive Action for Today's Health" (PATH) trial for increasing walking in low income, ethnic minority communities

BACKGROUND

Ethnic minorities and lower-income adults have among the highest rates of obesity and lowest levels of regular physical activity (PA). The Positive Action for Today's Health (PATH) trial compares three communities that are randomly assigned to different levels of an environmental intervention to improve safety and access for walking in low income communities.

DESIGN AND SETTING

Three communities matched on census tract information (crime, PA, ethnic minorities, and income) were randomized to receive either: an intervention that combines a police-patrolled-walking program with social marketing strategies to promote PA, a police-patrolled-walking only intervention, or no-walking intervention (general health education only). Measures include PA (7-day accelerometer estimates), body composition, blood pressure, psychosocial measures, and perceptions of safety and access for PA at baseline, 6, 12, 18, and 24 months.

INTERVENTION

The police-patrolled walking plus social marketing intervention targets increasing safety (training community leaders as walking captains, hiring off-duty police officers to patrol the walking trail, and containing stray dogs), increasing access for PA (marking a walking route), and utilizes a social marketing campaign that targets psychosocial and environmental mediators for increasing PA. MAIN HYPOTHESES/OUTCOMES: It is hypothesized that the police-patrolled walking plus social marketing intervention will result in greater increases in moderate-to-vigorous PA as compared to the police-patrolled-walking only or the general health intervention after 12 months and that this effect will be maintained at 18 and 24 months.

CONCLUSIONS

Implications of this community-based trial are discussed.

Identifying, recruiting, and assessing social networks at high risk for HIV/AIDS: Methodology, practice, and a case study in St Petersburg, Russia

Population segments at highest risk for HIV are often hidden, marginalized, and hard to reach by conventional prevention programmes. This pattern is especially true in Central and Eastern Europe, where major HIV epidemics have recently appeared, where population members do not perceive themselves as belonging to a community, and where there is little precedence for strong community-based organization service programmes. In these circumstances, naturally existing intact social networks still can be targeted by prevention programmes. HIV prevention interventions undertaken with at-risk social networks can establish new group norms, reduce the risk behaviour of network members, and can reach 'hidden' members of a population known personally to leaders of the social networks. This article illustrates a methodology and a practical description for: (1) accessing high-risk social networks in a community population; (2) identifying and enumerating the membership of the social networks; (3) identifying the social leadership of the networks; and (4) establishing the HIV risk behaviour levels of the recruited networks. To illustrate how social network methods can be applied in the field, the article provides case study reports of HIV prevention fieldwork practice targeting high-risk networks of young men who have sex with men and young heterosexual adults in St Petersburg, Russia. Although there is an extensive conceptual literature on the influence of social networks on risk behaviour, this article describes specific and practical techniques that can be in the development of approaches for social network-based interventions.

A comparison of the statistical power of different methods for the analysis of repeated cross-sectional cluster randomization trials with binary outcomes

Repeated cross-sectional cluster randomization trials are cluster randomization trials in which the response variable is measured on a sample of subjects from each cluster at baseline and on a different sample of subjects from each cluster at follow-up. One can estimate the effect of the intervention on the follow-up response alone, on the follow-up responses after adjusting for baseline responses, or on the change in the follow-up response from the baseline response. We used Monte Carlo simulations to determine the relative statistical power of different methods of analysis. We examined methods of analysis based on generalized estimating equations (GEE) and a random effects model to account for within-cluster homogeneity. We also examined cluster-level analyses that treated the cluster as the unit of analysis. We found that the use of random effects models to estimate the effect of the intervention on the change in the follow-up response from the baseline response had lower statistical power compared to the other competing methods across a wide range of scenarios. The other methods tended to have similar statistical power in many settings. However, in some scenarios, those analyses that adjusted for the baseline response tended to have marginally greater power than did methods that did not account for the baseline response.

Effects of the 5th and 7th grade enhanced versions of the keepin' it REAL substance use prevention curriculum

This study assessed the outcomes of adapting the culturally-grounded, middle school, substance-use prevention intervention, keepin ' it REAL (kiR), to target elementary school students and to address acculturation. At the beginning of 5th grade, 29 schools were randomly assigned to conditions obtained by crossing grade of implementation (5th, 7th, 5th + 7th, and control/comparison) by curriculum version [kiR-Plus vs. kiR-Acculturation Enhanced (AE)]. Students (n = 1984) completed 6 assessments through the end of 8th grade. The kiR curricula generally appear no more effective than the comparison schools' programming. Students receiving either version of the kiR intervention in only the 5th grade report greater increases in substance use than did control students. Receiving the kiR-AE version twice (both 5th and 7th grades) has benefits over receiving it once.

Social mobilization and social marketing to promote NaFeEDTA-fortified soya sauce in an iron-deficient population through a public–private partnership

Objective

The present pilot project aimed to assess the effectiveness of social mobilization and social marketing in improving knowledge, attitudes and practices (KAP) and Fe status in an Fe-deficient population.

Design

In an uncontrolled, before–after, community-based study, social mobilization and social marketing strategies were applied. The main outcomes included KAP and Hb level and were measured at baseline, 1 year later and 2 years later.

Setting

One urban county and two rural counties in Shijiazhuang Municipality, Hebei Province, China.

Subjects

Adult women older than 20 years of age and young children aged from 3 to 7 years were selected from three counties to attend the evaluation protocol.

Results

After 1 year, most knowledge and attitudes had changed positively towards the prevention and control of anaemia. The percentage of women who had adopted NaFeEDTA-fortified soya sauce increased from 8·9 % to 36·6 % (P ≤ 0·001). After 2 years, Hb levels had increased substantially, by 9·0 g/l (P ≤ 0·001) in adult women and 7·7 g/l (P ≤ 0·001) in young children.

Conclusion

Social mobilization and social marketing activities had a positive impact on the KAP of adult women, and resulted in marked improvements in Hb levels in both adult women and young children. This should be recommended as a national preventive strategy to prevent and control Fe deficiency and Fe-deficiency anaemia.

Comparison of the risk difference, risk ratio and odds ratio scales for quantifying the unadjusted intervention effect in cluster randomized trials

This paper evaluates methods for unadjusted analyses of binary outcomes in cluster randomized trials (CRTs). Under the generalized estimating equations (GEE) method the identity, log and logit link functions may be specified to make inferences on the risk difference, risk ratio and odds ratio scales, respectively. An alternative, 'cluster-level', method applies the t-test to summary statistics calculated for each cluster, using proportions, log proportions and log odds, to make inferences on the respective scales. Simulation was used to estimate the bias of the unadjusted intervention effect estimates and confidence interval coverage, generating data sets with different combinations of number of clusters, number of participants per cluster, intra-cluster correlation coefficient rho and intervention effect. When the identity link was specified, GEE had little bias and good coverage, performing slightly better than the log and logit link functions. The cluster-level method provided unbiased point estimates when proportions were used to summarize the clusters. When the log proportion and log odds were used, however, the method often had markedly large bias for two reasons: (i) bias in the modified summary statistic used for cluster-level estimation when a cluster has zero cases with the outcome of interest (arising when the number of participants sampled per cluster is small and the outcome prevalence is low) and (ii) asymptotically, the method estimates the ratio of geometric means of the cluster proportions or odds, respectively, between the trial arms rather than the ratio of arithmetic means.

Imputation strategies for missing continuous outcomes in cluster randomized trials

In cluster randomized trials, intact social units such as schools, worksites or medical practices - rather than individuals themselves - are randomly allocated to intervention and control conditions, while the outcomes of interest are then observed on individuals within each cluster. Such trials are becoming increasingly common in the fields of health promotion and health services research. Attrition is a common occurrence in randomized trials, and a standard approach for dealing with the resulting missing values is imputation. We consider imputation strategies for missing continuous outcomes, focusing on trials with a completely randomized design in which fixed cohorts from each cluster are enrolled prior to random assignment. We compare five different imputation strategies with respect to Type I and Type II error rates of the adjusted two-sample t -test for the intervention effect. Cluster mean imputation is compared with multiple imputation, using either within-cluster data or data pooled across clusters in each intervention group. In the case of pooling across clusters, we distinguish between standard multiple imputation procedures which do not account for intracluster correlation and a specialized procedure which does account for intracluster correlation but is not yet available in standard statistical software packages. A simulation study is used to evaluate the influence of cluster size, number of clusters, degree of intracluster correlation, and variability among cluster follow-up rates. We show that cluster mean imputation yields valid inferences and given its simplicity, may be an attractive option in some large community intervention trials which are subject to individual-level attrition only; however, it may yield less powerful inferences than alternative procedures which pool across clusters especially when the cluster sizes are small and cluster follow-up rates are highly variable. When pooling across clusters, the imputation procedure should generally take intracluster correlation into account to obtain valid inferences; however, as long as the intracluster correlation coefficient is small, we show that standard multiple imputation procedures may yield acceptable type I error rates; moreover, these procedures may yield more powerful inferences than a specialized procedure, especially when the number of available clusters is small. Within-cluster multiple imputation is shown to be the least powerful among the procedures considered.

Design and analysis of group-randomized trials in cancer: a review of current practices

BACKGROUND

Previous reviews have identified problems in the design and analysis of group-randomized trials in a number of areas. Similar problems may exist in cancer research, but there have been no comprehensive reviews.

METHODS

We searched Medline and PubMed for group-randomized trials focused on cancer prevention and control that were published between 2002 and 2006. We located and reviewed 75 articles to determine whether articles included evidence of taking group randomization into account in establishing the size of the trial, such as reporting the expected intraclass correlation, the group component of variance, or the variance inflation factor. We also examined the analytical approaches to determine their appropriateness.

RESULTS

Only 18 (24%) of the 75 articles documented appropriate methods for sample size calculations. Only 34 (45%) limited their reports to analyses judged to be appropriate. Fully 26 (34%) failed to report any analyses that were judged to be appropriate. The most commonly used inappropriate analysis was an analysis at the individual level that ignored the groups altogether. Nine articles (12%) did not provide sufficient information.

CONCLUSIONS

Many investigators who use group-randomized trials in cancer research do not adequately attend to the special design and analytic challenges associated with these trials. Failure to do so can lead to reporting type I errors as real effects, mislead investigators and policy-makers, and slow progress toward control and prevention of cancer. A collaborative effort by investigators, statisticians, and others will be required to ensure that group-randomized trials are planned and analyzed using appropriate methods so that the scientific community can have confidence in the published results.

The Importance and Role of Intracluster Correlations in Planning Cluster Trials

There is increasing recognition of the critical role of intracluster correlations of health behavior outcomes in cluster intervention trials. This study examines the estimation, reporting, and use of intracluster correlations in planning cluster trials. We use an estimating equations approach to estimate the intracluster correlations corresponding to the multiple-time-point nested cross-sectional design. Sample size formulae incorporating 2 types of intracluster correlations are examined for the purpose of planning future trials. The traditional intracluster correlation is the correlation among individuals within the same community at a specific time point. A second type is the correlation among individuals within the same community at different time points. For a "time x condition" analysis of a pretest-posttest nested cross-sectional trial design, we show that statistical power considerations based upon a posttest-only design generally are not an adequate substitute for sample size calculations that incorporate both types of intracluster correlations. Estimation, reporting, and use of intracluster correlations are illustrated for several dichotomous measures related to underage drinking collected as part of a large nonrandomized trial to enforce underage drinking laws in the United States from 1998 to 2004.

Sizing a trial to alter the trajectory of health behaviours: methods, parameter estimates, and their application

Group-randomized trials often involve repeat observations on the same participants. When there are no more than two observations from each participant, standard mixed-model regression methods for a pretest-posttest design can be used. When there are more than two observations from each participant, random coefficients models may be useful. This paper describes the random coefficients analysis appropriate to data from an extended nested cohort design and presents the methods for power analysis and sample size calculations based on that analysis. We provide estimates for the parameters required for those calculations for a number of adolescent health behaviours. We also show how the estimates can be used to plan a future trial.

Developments in cluster randomized trials and Statistics in Medicine

The design and analysis of cluster randomized trials has been a recurrent theme in Statistics in Medicine since the early volumes. In celebration of 25 years of Statistics in Medicine, this paper reviews recent developments, particularly those that featured in the journal. Issues in design such as sample size calculations, matched paired designs, cohort versus cross-sectional designs, and practical design problems are covered. Developments in analysis include modification of robust methods to cope with small numbers of clusters, generalized estimation equations, population averaged and cluster specific models. Finally, issues on presenting data, some other clustering issues and the general problem of evaluating complex interventions are briefly mentioned.

Longitudinal and repeated cross-sectional cluster-randomization designs using mixed effects regression for binary outcomes: bias and coverage of frequentist and Bayesian methods

As medical applications for cluster randomization designs become more common, investigators look for guidance on optimal methods for estimating the effect of group-based interventions over time. This study examines two distinct cluster randomization designs: (1) the repeated cross-sectional design in which centres are followed over time but patients change, and (2) the longitudinal design in which individual patients are followed over time within treatment clusters. Simulations of each study design stipulated a multiplicative treatment effect (on the log odds scale), between 5 and 15 clusters in each of two treatment arms, and followed over two time periods. Estimation options included linear mixed effects models using restricted maximum likelihood (REML), generalized estimating equations (GEE), mixed effects logistic regression using both penalized quasi likelihood (PQL) and numerical integration, and Bayesian Monte Carlo analysis. For the repeated cross-sectional designs, most methods performed well in terms of bias and coverage when clusters were numerous (30) and variability across clusters of baseline risk and treatment effect was modest. With few clusters (two groups of five) and higher variability, only the Bayesian methods maintained coverage. In the longitudinal designs, the common methods of REML, GEE, or PQL performed poorly when compared to numerical integration, while Bayesian methods demonstrated less bias and better coverage for estimates of both log odds ratios and risk differences. The performance of common statistical tools for the analysis of cluster randomization designs depends heavily on the precise design, the number of clusters, and the variability of baseline outcomes and treatment effects across centres.

Impact evaluation of a Dutch community intervention to improve health-related behaviour in deprived neighbourhoods

This study investigates the impact of a 2-year community intervention on health-related behaviour among adults aged 18-65 years living in deprived neighbourhoods in Eindhoven, The Netherlands. The intervention is evaluated in a community intervention trial with a quasi-experimental design in a longitudinal cohort survey (n=1926 and attrition rate: 31%) using postal questionnaires. In the 2-year implementation phase, more than 40 intervention activities were planned and delivered by intersectoral neighbourhood coalitions. Outcome measures were fruit consumption, vegetable consumption, physical activity, smoking, alcohol consumption and intermediate outcomes of behaviour (i.e. attitudes, self-efficacy, awareness, knowledge and stages of change). The intervention demonstrated no evidence for an impact on vegetable consumption, physical activity, smoking and alcohol consumption and weak evidence for a small impact on (intermediate) outcomes of fruit consumption.

Proliferation of Cardiac Technology in Canada

Background— Critics remain skeptical about the long-term sustainability of Medicare in Canada because of the proliferation of health technology and escalating expenditures. The objective of this study was to examine the temporal trends in the utilization and costs of cardiovascular technologies for the evaluation and/or management of patients with ischemic heart disease in Canada.

Methods and Results— This repeated cross-sectional population-based study of Ontario residents examined the temporal trends in the utilization and costs associated with echocardiography, stress (imaging and nonimaging) testing, coronary angiography, percutaneous coronary intervention (PCI), and bypass surgery between 1992 and 2001. Annual costs increased by nearly 2-fold over the 10-year study period and cumulatively accounted for more than $2.8 billion (Canadian) in expenditures. The proliferation in use of cardiac testing/interventions over time outstripped both demographic shifts and changes in the prevalence of coronary artery disease. Annual increases were widespread for all procedures ( P <0.001) and ranged from 2% per year for nonimaging stress tests to 12% per year for PCI, after adjustment for age and sex. Generally, utilization rates were higher among the elderly, males, and those of low socioeconomic status. With few exceptions, annual increases in the utilization rates of cardiac tests and procedures were disproportionately higher among the elderly and women, but they were similar across socioeconomic subgroups. Increases in utilization appeared to reflect referrals toward higher-risk populations.

Conclusions— Although definitive conclusions about the appropriateness of temporal patterns cannot be ascertained, the proliferation of cardiac testing challenges the sustainability of Medicare in Canada, especially given uncertainty as to whether the accompanying incremental rise in total expenditures translates into significant outcome benefits in the population.

Design of the Trial of Activity in Adolescent Girls (TAAG)

The primary aim of the Trial of Activity in Adolescent Girls (TAAG) is to test an intervention to reduce by half the age-related decline in moderate to vigorous physical activity (MVPA) in middle school girls. The intervention will be evaluated using a group-randomized trial involving 36 middle schools. The primary endpoint is the mean difference in intensity-weighted minutes (i.e., MET-minutes) of MVPA between intervention and comparison schools assessed using accelerometry. The TAAG study design calls for two cross-sectional samples, one drawn from 6th graders at the beginning of the study and the second drawn from 8th graders at the end of the study following the 2-year implementation of the intervention. An important strength of this design over a cohort design is the consistency with the goals of TAAG, which focus on environmental-level rather than individual-level interventions to produce change. The study design specifies a recruitment rate of 80% and a smaller sample of girls at baseline (n=48 per school) than at follow-up (n=96 per school). A two-stage model will be used to test the primary hypothesis. In the first stage, MET-weighted minutes of MVPA will be regressed on school, time (baseline or follow-up), their interaction, ethnicity and week of data collection. The second stage analysis will be conducted on the 72 adjusted means from the first stage. In the main-effects model, we will regress the follow-up school mean MET-weighted minutes of MVPA on study condition, adjusting for the baseline school mean. The TAAG study addresses an important health behavior, and also advances the field of group-randomized trials through the use of a study design and analysis plan tailored to serve the main study hypothesis.

The (mis)estimation of neighborhood effects: causal inference for a practicable social epidemiology

The resurgence of interest in the effect of neighborhood contexts on health outcomes, motivated by advances in social epidemiology, multilevel theories and sophisticated statistical models, too often fails to confront the enormous methodological problems associated with causal inference. This paper employs the counterfactual causal framework to illuminate fundamental obstacles in the identification, explanation, and usefulness of multilevel neighborhood effect studies. We show that identifying useful independent neighborhood effect parameters, as currently conceptualized with observational data, to be impossible. Along with the development of a dependency-based methodology and theories of social interaction, randomized community trials are advocated as a superior research strategy, one that may help social epidemiology answer the causal questions necessary for remediating disparities and otherwise improving the public's health.

Bayesian methods for analysis of binary outcome data in cluster randomized trials on the absolute risk scale

A Bayesian hierarchical modelling approach to the analysis of cluster randomized trials has advantages in terms of allowing for full parameter uncertainty, flexible modelling of covariates and variance structure, and use of prior information. Previously, such modelling of binary outcome data required use of a log-odds ratio scale for the treatment effect estimate and an approximation linking the intracluster correlation (ICC) to the between-cluster variance on a log-odds scale. In this paper we develop this method to allow estimation on the absolute risk scale, which facilitates clinical interpretation of both the treatment effect and the between-cluster variance. We describe a range of models and apply them to data from a trial of different interventions to promote secondary prevention of coronary heart disease in primary care. We demonstrate how these models can be used to incorporate prior data about typical ICCs, to derive a posterior distribution for the number needed to treat, and to consider both cluster and individual level covariates. Using these methods, we can benefit from the advantages of Bayesian modelling of binary outcome data at the same time as providing results on a clinically interpretable scale.

Baseline adjustments for binary data in repeated cross-sectional cluster randomized trials

Analysis of covariance models, which adjust for a baseline covariate, are often used to compare treatment groups in a controlled trial in which individuals are randomized. Such analysis adjusts for any baseline imbalance and usually increases the precision of the treatment effect estimate. We assess the value of such adjustments in the context of a cluster randomized trial with repeated cross-sectional design and a binary outcome. In such a design, a new sample of individuals is taken from the clusters at each measurement occasion, so that baseline adjustment has to be at the cluster level. Logistic regression models are used to analyse the data, with cluster level random effects to allow for different outcome probabilities in each cluster. We compare the estimated treatment effect and its precision in models that incorporate a covariate measuring the cluster level probabilities at baseline and those that do not. In two data sets, taken from a cluster randomized trial in the treatment of menorrhagia, the value of baseline adjustment is only evident when the number of subjects per cluster is large. We assess the generalizability of these findings by undertaking a simulation study, and find that increased precision of the treatment effect requires both large cluster sizes and substantial heterogeneity between clusters at baseline, but baseline imbalance arising by chance in a randomized study can always be effectively adjusted for.

An integrated population-averaged approach to the design, analysis and sample size determination of cluster-unit trials

While the mixed model approach to cluster randomization trials is relatively well developed, there has been less attention given to the design and analysis of population-averaged models for randomized and non-randomized cluster trials. We provide novel implementations of familiar methods to meet these needs. A design strategy that selects matching control communities based upon propensity scores, a statistical analysis plan for dichotomous outcomes based upon generalized estimating equations (GEE) with a design-based working correlation matrix, and new sample size formulae are applied to a large non-randomized study to reduce underage drinking. The statistical power calculations, based upon Wald tests for summary statistics, are special cases of a general power method for GEE.