Developments in cluster randomized trials and <i>Statistics in Medicine</i>
Statistics in MedicinePublished 29 November 2006Open access
Michael J. Campbell, Allan Donner, Neil Klar
Citations248
SJR quartileQ1
SJR score1.27
SNIP1.33
Generate an AI Snapshot to get a quick, structured summary of this paper.
Study Snapshot
ObjectiveStudy objective
MethodsResearch methodology
PopulationPopulation studied
Sample sizeSample sizes
OutcomesStudy outcomes here
ResultsStudy results comes here
LimitationsResearch study limitations comes here
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
TL;DR
The design and analysis of cluster randomized trials has been a recurrent theme in Statistics in Medicine since the early volumes and recent developments, particularly those that featured in the journal are reviewed.
Abstract
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.
Keywords
Mathematics
BiometrikaLongitudinal data analysis using generalized linear models
18,123 Citations1986Kung‐Yee Liang, Scott L. Zeger
American Journal of EpidemiologyA Modified Poisson Regression Approach to Prospective Studies with Binary Data
9,099 Citations2004Guangyong Zou
Results from a limited simulation study indicate that this approach is very reliable even with total sample sizes as small as 100, and the method is illustrated with two data sets.
Journal of the American Statistical AssociationApproximate Inference in Generalized Linear Mixed Models
4,212 Citations1993N. E. Breslow, David Clayton
BMJAnalysis of serial measurements in medical research.
3,156 Citations1990J. N. S. Matthews, D G Altman +2 more
Use of summary measures to analyse serial measurements, though not new, is potentially a useful and simple tool in medical research.
The American StatisticianUsing Heteroscedasticity Consistent Standard Errors in the Linear Regression Model
1,414 Citations2000J. Scott Long, Laurie H. Ervin
BMJComplex interventions: how “out of control” can a randomised controlled trial be?
1,255 Citations2004Penelope Hawe, Alan Shiell +1 more
A radical departure from the way large scale interventions are typically conceptualised is proposed, which could liberate interventions to be responsive to local context and potentially more effective while still allowing meaningful evaluation in controlled designs.
Health Technology AssessmentMethods for evaluating area-wide and organisation-based interventions in health and health care: a systematic review.
1,175 Citations1999Ukoumunne, Martin Gulliford +3 more
Medicine & Science in Sports & ExerciseDesign and Analysis of Group-Randomized Trials
1,135 Citations1999Seema S. Sonnad
Design and Analysis of Group-Randomized Trials
1,124 Citations2007David M. Murray
This paper focuses on the design and analysis of nested cross-Sectional designs and the applications of these designs in the context of nested Cohort designs.
Journal of the American Statistical AssociationPattern-Mixture Models for Multivariate Incomplete Data
961 Citations1993Roderick J. A. Little
Statistics in MedicineTables of the number of patients required in clinical trials using the logrank test
725 Citations1982Laurence S. Freedman
The logrank test is commonly used in the analysis of clinical trials in chronic diseases such as cancer and the numbers required are considerably smaller than those in existing tables when the event-free proportions are small, but otherwise comparable.
American Journal of EpidemiologyRANDOMIZATION BY CLUSTER
696 Citations1981Allan Donner, Nicholas Birkett +1 more
American Journal of Public HealthDesign and Analysis of Group-Randomized Trials: A Review of Recent Methodological Developments
669 Citations2004David M. Murray, Sherri P. Varnell +1 more
Developments in estimates of intraclass correlation, power analysis, matched designs, designs involving one group per condition, and designs in which individuals are randomized to receive treatments in groups are reviewed.
New England Journal of MedicineDoes Free Care Improve Adults' Health?
668 Citations1983Robert H. Brook, John E. Ware +8 more
International Journal of EpidemiologySample size for cluster randomized trials: effect of coefficient of variation of cluster size and analysis method
537 Citations2006Sandra Eldridge, Deborah Ashby +1 more
This work shows how a simple formula can be used to judge the possible effect of unequal cluster sizes for various types of analyses and both continuous and binary outcomes and provides a good estimate of sample size requirements for trials analysed using cluster-level analyses weighting by cluster size.
BiometricsA Covariance Estimator for GEE with Improved Small‐Sample Properties
537 Citations2001Lloyd Mancl, Timothy A. DeRouen
This paper proposes a bias-corrected covariance estimator for generalized estimating equations (GEE) that gives tests with sizes close to the nominal level even when the number of subjects was 10 and cluster sizes were unequal, whereas the robust and jackknife covariances estimators gave Tests with sizes that could be 2-3 times the nominallevel.
American Journal of Public HealthPitfalls of and Controversies in Cluster Randomization Trials
448 Citations2004Allan Donner, Neil Klar
This work discusses several topics related to cluster randomized trials, with emphasis on the choices that must be made in the planning stages of a trial and on some potential pitfalls to be avoided.
Statistics in MedicineThe role of internal pilot studies in increasing the efficiency of clinical trials
443 Citations1990Janet Wittes, Erica Brittain
This paper proposes designs, which are term internal pilot studies, that designate a portion of the main trial as a pilot phase, and recommends considering a similar approach for a number of types of endpoints.
Clinical TrialsDesign and analysis of clinical trials with clustering effects due to treatment
301 Citations2005Chris Roberts, Stephen A. Roberts
It is argued that the design and analysis of this type of trial should take account of possible heterogeneity in cluster size and intracluster correlation, and the implications of differential clustering effects for sample size and power are addressed.
Clinical TrialsLessons for cluster randomized trials in the twenty-first century: a systematic review of trials in primary care
297 Citations2004Sandra Eldridge, Deborah Ashby +3 more
Methodologic quality of cluster randomized trials in primary health care is variable and reporting needs improvement, and the use of cluster randomization should be indicated in the title or abstract so these kinds of trials are easier to identify.
Statistics in MedicineApproaches to sample size estimation in the design of clinical trials—a review
254 Citations1984Allan Donner
Statistics in MedicineON DESIGN CONSIDERATIONS AND RANDOMIZATION-BASED INFERENCE FOR COMMUNITY INTERVENTION TRIALS
244 Citations1996Mitchell H. Gail, Steven D. Mark +3 more
Design considerations and the role of randomization-based inference in randomized community intervention trials are discussed, and it is stressed that longitudinal follow-up of cohorts within communities often yields useful information on the effects of intervention on individuals, whereas cross-sectional surveys can usefully assess the impact of intervention.
BMC Medical Research MethodologyCluster randomised trials in the medical literature: two bibliometric surveys
194 Citations2004Martin Bland
Statistics in MedicineSmall‐sample adjustments in using the sandwich variance estimator in generalized estimating equations
173 Citations2002Wei Pan, Melanie M. Wall
This work proposes using an approximate t- or F-test that takes account of the variability of the sandwich estimator and has a clear implication on constructing Wald-type confidence intervals or regions.
Statistics in MedicineCurrent and future challenges in the design and analysis of cluster randomization trials
172 Citations2001Neil Klar, Allan Donner
Statistics in MedicineA COMPARISON OF STATISTICAL METHODS FOR CLUSTERED DATA ANALYSIS WITH GAUSSIAN ERROR
156 Citations1996Ziding Feng, Dale McLerran +1 more
This work investigates by simulation the properties of four different estimation procedures under a linear model for correlated data with Gaussian error: maximum likelihood based on the normal mixed linear model; generalized estimating equations; a four-stage method, and a bootstrap method that resamples clusters rather than individuals.
BiometrikaA geometric approach to assess bias due to omitted covariates in generalized linear models
156 Citations1993John Neuhaus, Nicholas P. Jewell
Statistics in MedicineCohort versus cross‐sectional design in large field trials: Precision, sample size, and a unifying model
147 Citations1994Henry A. Feldman, Sonja M. McKinlay
A comprehensive design equation is provided, relating sample size to precision for cohort and cross-sectional designs, and it is shown that the follow-up cost and selection bias attending a cohort design may outweigh any theoretical advantage in precision.
Controlled Clinical TrialsAspects of statistical design for the community intervention trial for smoking cessation (COMMIT)
146 Citations1992Mitchell H. Gail, David P. Byar +2 more
The COMMIT design includes 11 pair-matched communities, which should provide good power to detect a 10% or greater difference in quit rates between the intervention and control communities in cohort of heavy smokers and in cohorts of light or moderate smokers.
Journal of Clinical EpidemiologyIntraclass correlation coefficient and outcome prevalence are associated in clustered binary data
141 Citations2005Martin Gulliford, Geoffrey G. Adams +4 more
The prevalence of an outcome may be used to make an informed assumption about the magnitude of the intraclass correlation coefficient in a range of outcomes in community and health services settings.
Statistics in MedicineBreaking the matches in a paired <i>t</i>‐test for community interventions when the number of pairs is small
132 Citations1995Paula Diehr, Donald C. Martin +2 more
Computer simulation is used to examine the performance of an approach in which the authors matched communities but performed an unmatched analysis, and a variant of this procedure is discussed, in which an unmatchedAnalysis is performed only if the matching 'did not work'.
Statistics in MedicineSample size formulae for intervention studies with the cluster as unit of randomization
131 Citations1988F. Hsieh
This paper presents sample size formulae for both continuous and dichotomous endpoints obtained from intervention studies that use the cluster as the unit of randomization, derived from Student's t-test with use of cluster summary measures and a variance that consists of within and between cluster components.
Statistics in MedicineBayesian methods of analysis for cluster randomized trials with binary outcome data
126 Citations2001Rebecca Turner, Rumana Omar +1 more
The sensitivity of results to the prior specified is investigated and it is found that the estimate of intervention effect changes very little in this data set, while its interval estimate is more sensitive.
Statistics in MedicineUnequal cluster sizes for trials in English and Welsh general practice: implications for sample size calculations
125 Citations2001Sally Kerry, Martin Bland
The effect of variability in practice list size on sample size calculations is demonstrated using the General Medical Services Statistics for England and Wales, 1997, and minimum variance weights are shown to be superior to uniform, particularly when clusters are small, and to cluster size weights, particularlywhen clusters are large.
Statistics in MedicineAnalysis of data arising from a stratified design with the cluster as unit of randomization
125 Citations1987Allan Donner, Alan Donald
A weighted paired t-test based on the empirical logistic transform is proposed for designs that randomize large aggregate clusters in each of several strata for assessing the statistical significance of the intervention effect over all strata.
Statistics in MedicineAnalysis of data from group-randomized trials with repeat observations on the same groups
118 Citations1998David M. Murray, Peter J. Hannan +3 more
This study used Monte Carlo simulations to evaluate the performance of alternative models for the analysis of group-randomized trials having more than two time intervals for data collection, and indicates that the mixed-model ANOVA yields unbiased estimates of sampling variation and nominal type I error rates when the group-specific time trends are homogenous.
Statistics in MedicineThe merits of matching in community intervention trials: a cautionary tale
115 Citations1997Neil Klar, Allan Donner
It is shown in this paper that there are several analytic limitations with pair-matched designs, including the restriction of prediction models to cluster-level baseline risk factors, the inability to test for homogeneity of odds ratios, and difficulties in estimating the intracluster correlation coefficient.
Statistical Methods in Medical ResearchCluster randomized trials in general (family) practice research
115 Citations2000Michael J. Campbell
Issues discussed are the estimation and range of values of the intra-cluster correlation coefficient found in general practice, and the associated sample size problems.
Statistics in MedicineBayesian methods for cluster randomized trials with continuous responses
110 Citations2001David J. Spiegelhalter
Bayesian methods for cluster randomized trials extend the random-effects formulation by allowing both the use of external evidence on parameters and straightforward relaxation of the standard normality and constant variance assumptions, and recommends a standard 'non-informative' prior distribution for the within-cluster sampling variance and an independent prior on the intraclass correlation coefficient.
Statistics in MedicineThe effect of matching on the power of randomized community intervention studies
108 Citations1993Donald C. Martin, Paula Diehr +2 more
It is shown here that if the number of communities is small, the matched design will probably have less power than the unmatched design, which outweighs the benefits of matching on any but the strongest correlates of changes in behaviour.
Statistics in MedicineAn integrated population‐averaged approach to the design, analysis and sample size determination of cluster‐unit trials
107 Citations2003John S. Preisser, Mary Lynn Young +2 more
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), and new sample size formulae are applied to a large non-randomized study to reduce underage drinking.
Statistics in MedicineAnalysis of a cluster randomized trial with binary outcome data using a multi-level model
105 Citations2000Rumana Omar, Simon G. Thompson
The use of multi-level logistic regression models was explored for the analysis of data from a cluster randomized trial investigating whether a training programme for general practitioners' reception staff could improve women's attendance at breast screening.
Statistics in MedicineThe design and analysis of paired cluster randomized trials: an application of meta-analysis techniques
99 Citations1997Simon G. Thompson, Stephen Pyke +1 more
The design and analysis of trials in which pairs of clusters are randomized in the context of one recent trial, the British Family Heart Study, are discussed.
Statistics in MedicineExtending the CONSORT statement to cluster randomized trials: for discussion
91 Citations2001Diana R Elbourne, Marion Campbell
The need for clear reporting of randomized controlled trials has been emphasized recently and evidence-based suggestions for a checklist and a patient flow diagram are suggested for discussion.
Statistics in MedicineSample size re‐estimation in cluster randomization trials
91 Citations2002Stephen Lake, Erin Kammann +2 more
This work proposes using a portion of the data to estimate the nuisance parameters and to re‐estimate sample size based on the estimates, which extends the standard internal pilot study methods to the setting of cluster randomization trials.
Design and Analysis of Cluster Randomization Trials
88 Citations2003David M. Zucker
The cluster randomization design poses special issues in the areas of design, logistics, sample size calculation, and statistical analysis in trials in which the intervention is applied at the level of the individual.
Clinical TrialsThe use of random effects models to allow for clustering in individually randomized trials
86 Citations2005Katherine J. Lee, Simon G. Thompson
Clustering is an important issue in many individually randomized trials and ignoring it can lead to underestimates of the uncertainty and too extreme P-values, which can have a large impact on the estimation and interpretation of the treatment effect.
Statistics in MedicineComparison of statistical methods for analysis of clustered binary observations
74 Citations2004Moonseong Heo, Andrew C. Leon
The results show that the performance of the full likelihood and the penalized quasi‐likelihood methods is superior for analysis of clustered binary observations, and is not necessarily inferior to that of the fixed‐effects logistic regression fit even when within‐cluster correlations are zero.
Statistics in MedicineMethods for modelling change in cluster randomization trials
63 Citations2004Neil Klar, Gerarda Darlington
This paper considers methods for modelling change when data are obtained from cluster randomization trials where the unit of allocation is a family, school or community, and focuses on mixed effects linear regression extensions of two-sample t-tests and analysis of covariance.
Statistics in MedicineAnalysis of cluster randomized trials with repeated cross-sectional binary measurements
63 Citations2001Obioha C. Ukoumunne, Simon G. Thompson
Comparisons of methods that can be used to evaluate the impact of an intervention on dichotomous outcomes conclude however that analyses of change from baseline can be misleading since they are affected too much by baseline results, and that analysis of covariance approaches are preferable.
Statistics in MedicineCorrelated binomial variates: Properties of estimator of intraclass correlation and its effect on sample size calculation
63 Citations1992Ziding Feng, James E. Grizzle
This work compares by simulation the moment method and the more standard ANOVA method of estimating the intraclass correlation and finds the former is less biased for a small to moderate number of clusters but the difference disappears when the appropriate degree of freedom is used for the ANOVA estimator.
Statistics in MedicineAllowing for imprecision of the intracluster correlation coefficient in the design of cluster randomized trials
60 Citations2004Rebecca Turner, A Toby Prevost +1 more
Methods which allow for the uncertainty in a previously observed ICC, using a variety of distributional assumptions, are developed and can be used by applied researchers to protect against loss of power, or to choose a design which reduces the impact of uncertainty in the ICC.
Statistics in MedicineSmall-sample performance of the robust score test and its modifications in generalized estimating equations
59 Citations2005Xu Guo, Wei Pan +3 more
This work investigates the small-sample performance of the robust score test for correlated data and proposes several modifications to improve the performance, including a modification based on a simple adjustment to the usual robust score statistic by a factor of J/(J - 1) (where J is the number of clusters).
Statistics in MedicineAssessing the gain in efficiency due to matching in a community intervention study
55 Citations1990Laurence S. Freedman, S. B. Green +1 more
This work has used baseline smoking quit rates in the communities as a surrogate for the outcome measure to evaluate the gain in efficiency from the matching, and estimates an efficiency gain of at least 50 per cent using the matched design.
Statistics in MedicinePower and money in cluster randomized trials: when is it worth measuring a covariate?
45 Citations2005Mirjam Moerbeek
It is shown that the strategy that uses covariates is more cost-efficient in detecting a treatment effect when the costs to measure these covariates are small and the correlation between the covariates and outcome is sufficiently large.
BMJExtending CONSORT to include cluster trials
44 Citations2004Michael J. Campbell
Using the extended CONSORT statement should help reduce bias and help readers to understand a cluster trial's conduct and to assess thevalidity of its conduct.
Statistics in MedicineDesign and sample size estimation in clinical trials with clustered survival times as the primary endpoint
43 Citations2003Tailiang Xie, Joel Waksman
A sample size formula for clustered survival data is derived via Lee, Wei and Amato's marginal model and can be easily used to plan a clinical trial in which clustered survival times are of primary interest.
Statistics in MedicineRecruitment strategies in a cluster randomized trial – cost implications
40 Citations2002Terry N. Flynn, Elise Whitley +1 more
Examples of how problems in recruitment arise are taken from cluster randomized trials and more intuitive methods of describing clustering are summarized to provide a framework that enables triallists to consider explicitly the trade‐off between power and cost that is often inherent in such trials.
Statistics in MedicineExplaining community-level variance in group randomized trials
35 Citations1999Ziding Feng, Paula Diehr +4 more
It is found that adjusting for covariates will often substantially reduce the between-community variance component and suggests that blocking may not be necessary for the design of the group randomized trials where the change from baseline is of primary interest.
Statistics in MedicineThe merits of breaking the matches: a cautionary tale
32 Citations2006Allan Donner, Monica Taljaard +1 more
Focusing on the case of a continuous outcome variable, it is shown that the practice of performing an unmatched analysis on data arising from a matched-pair design can lead to bias in the estimated regression coefficient, and a corresponding test of significance which is overly liberal.
Statistics in MedicineBayesian methods for analysis of binary outcome data in cluster randomized trials on the absolute risk scale
26 Citations2003Simon G. Thompson, David E. Warn +1 more
This paper develops this method to allow estimation on the absolute risk scale, which facilitates clinical interpretation of both the treatment effect and the between‐cluster variance, and describes a range of models that 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.
Statistics in MedicineBaseline adjustments for binary data in repeated cross‐sectional cluster randomized trials
26 Citations2003Richard M. Nixon, Simon G. Thompson
It is found 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.
Statistics in MedicineComparing completely and stratified randomized designs in cluster randomized trials when the stratifying factor is cluster size: a simulation study
25 Citations2004J D Lewsey
The results show that when cluster size is strongly associated with a cluster level factor which is predictive of outcome, the stratified randomized design has superior power results to the completely randomized design and that the superiority is related to the number of clusters.
Statistics in MedicineAccounting for correlation and compliance in cluster randomized trials
20 Citations2001Tom Loeys, Stijn Vansteelandt +1 more
It is argued that cluster randomization carries the potential for post‐randomization exposures which involve differentially selective compliance between treatment arms, even for an all or nothing exposure at the individual level.
Statistics in MedicineEstimation of the design effect in community intervention studies
19 Citations1991Ruth M. Mickey, Gregory D. Goodwin +1 more
The approach in estimation of this cluster sampling variance is to determine the variance for simple random sampling and multiply it by a design effect which is calculated with use of information obtained from other sources.
Statistics in MedicinePerson-time analysis of paired community intervention trials when the number of communities is small
19 Citations1998Ron Brookmeyer, Ying Qing Chen
The objective of this paper is to evaluate person-time methods of analysis of paired community intervention trials when the number of community pairs is small and considers methods to account for individual level covariates.
Statistics in MedicineComparison of tests for categorical data from a stratified cluster randomized trial
15 Citations2002Timothy Dobbins, Judy M. Simpson
An extension of a method from survey sampling that uses the design effect to reduce the effective sample size to extend the standard Cochran–Mantel–Haenszel (CMH) statistic by using appropriate covariance matrices and with a bootstrap method is proposed.
Statistics in MedicineAn evaluation of methods for the stratified analysis of clustered binary data in community intervention trials
10 Citations2003James Song, Chul Ahn
A simulation study is conducted in a community intervention setting of stratified analysis of clustered binary data and recommends the use of chi(2) (RS) and chi( 2) (GLMM) since they have better overall performance in terms of empirical significance levels and empirical power levels.
Statistics in MedicineAnalysis of incidence rates in cluster-randomized trials of interventions against recurrent infections, with an application to trachoma
5 Citations2005Neâl Alexander, Paul M. Emerson
A method is developed for the Bayesian analysis of rate ratios of repeated events in matched cluster‐randomized trials, and applied to the Flies and Eyes trial against trachoma in The Gambia, which yields incidence and clearance rate ratios for age and the proportion of transmission due to flies was not possible to estimate accurately.