Estimating intervention effectiveness in trials of malaria interventions with contamination

Abstract Background In cluster randomized trials (CRTs) or stepped wedge cluster randomized trials (SWCRTs) of malaria interventions, mosquito movement leads to contamination between trial arms unless buffer zones separate the clusters. Contamination can be accounted for in the analysis, yielding an estimate of the contamination range, the distance over which contamination measurably biases the effectiveness. Methods A previously described analysis for CRTs is extended to SWCRTs and estimates of effectiveness are provided as a function of intervention coverage. The methods are applied to two SWCRTs of malaria interventions, the SolarMal trial on the impact of mass trapping of mosquitoes with odor-baited traps and the AvecNet trial on the effect of adding pyriproxyfen to long-lasting insecticidal nets. Results For the SolarMal trial, the contamination range was estimated to be 146 m ($$95\%$$ 95 % credible interval $$[0.052,\,0.923]$$ [ 0.052 , 0.923 ] km), together with a $$31.9\%$$ 31.9 % ($$95\%$$ 95 % credible interval $$[15.3,\,45.8]\%$$ [ 15.3 , 45.8 ] % ) reduction of Plasmodium infection, compared to the $$30.0\%$$ 30.0 % reduction estimated without accounting for contamination. The estimated effectiveness had an approximately linear relationship with coverage. For the AvecNet trial, estimated contamination effects were minimal, with insufficient data from the cluster boundary regions to estimate the effectiveness as a function of coverage. Conclusions The contamination range in these trials of malaria interventions is much less than the distances Anopheles mosquitoes can fly. An appropriate analysis makes buffer zones unnecessary, enabling the design of more cost-efficient trials. Estimation of the contamination range requires information from the cluster boundary regions and trials should be designed to collect this.

Download Full-text

Quantifying the impact of fixed effects modeling of clusters in multiple imputation for cluster randomized trials

Biometrical Journal ◽

10.1002/bimj.201000140 ◽

2011 ◽

Vol 53 (1) ◽

pp. 57-74 ◽

Cited By ~ 43

Author(s):

Rebecca R. Andridge

Keyword(s):

Multiple Imputation ◽

Fixed Effects ◽

Randomized Trials ◽

Cluster Randomized Trials ◽

Cluster Randomized ◽

The Impact

Download Full-text

Methodological issues in the design and analysis of cluster randomized trials

10.1093/med/9780198816805.003.0032 ◽

2021 ◽

pp. 113-128

Author(s):

Kathy J. Baisley ◽

Richard J. Hayes ◽

Lawrence H. Moulton

Keyword(s):

Gold Standard ◽

Randomized Trials ◽

Controlled Trials ◽

Cluster Randomized Trials ◽

Methodological Issues ◽

Eligibility Criteria ◽

Experimental Conditions ◽

Cluster Randomized ◽

And Control ◽

The Impact

Randomized controlled trials are the accepted gold standard for evaluating the effects of interventions to improve health. In the majority of such trials, individuals are randomly allocated to the experimental conditions under study, for example, to treatment and control arms. However, in some situations it is more appropriate to randomly allocate groups of individuals to the treatment arms. These groups are referred to as clusters, and trials of this kind are known as cluster randomized trials (CRTs). Examples of clusters include schools, villages, workplaces, or health facilities, but there are many other possible choices. In some CRTs, all individuals within the selected clusters are automatically included. In others, there may be additional eligibility criteria. Similarly, the impact of the intervention may be measured in all individuals in the cluster, or in a random subsample. This chapter aims to discuss methodological issues that arise in the design and analysis of CRTs

Download Full-text

Inference for the treatment effect in longitudinal cluster randomized trials when treatment effect heterogeneity is ignored

Statistical Methods in Medical Research ◽

10.1177/09622802211041754 ◽

2021 ◽

pp. 096228022110417

Author(s):

Rhys Bowden ◽

Andrew B Forbes ◽

Jessica Kasza

Keyword(s):

Treatment Effect ◽

Randomized Trials ◽

Error Rates ◽

Type I ◽

Cluster Randomized Trials ◽

Stepped Wedge ◽

Treatment Effect Heterogeneity ◽

Cluster Randomized ◽

Effect Heterogeneity ◽

The Impact

In cluster-randomized trials, sometimes the effect of the intervention being studied differs between clusters, commonly referred to as treatment effect heterogeneity. In the analysis of stepped wedge and cluster-randomized crossover trials, it is possible to include terms in outcome regression models to allow for such treatment effect heterogeneity yet this is not frequently considered. Outside of some simulation studies of specific cases where the outcome is binary, the impact of failing to include terms for treatment effect heterogeneity on the variance of the treatment effect estimator is unknown. We analytically examine the impact of failing to include terms for treatment effect heterogeneity on the variance of the treatment effect estimator, when outcomes are continuous. Using analysis of variance and feasible generalized least squares we provide expressions for this variance. For both the cluster-randomized crossover design and the stepped wedge design, our analytic derivations indicate that failing to include treatment effect heterogeneity results in the estimates for variance of the treatment effect that are too small, leading to inflation of type I error rates. We therefore recommend assessing the sensitivity of sample size calculations and conclusions drawn from the analysis of cluster randomized trials to the inclusion of treatment effect heterogeneity.

Download Full-text

Power for Detecting Treatment by Moderator Effects in Two- and Three-Level Cluster Randomized Trials

Journal of Educational and Behavioral Statistics ◽

10.3102/1076998616655442 ◽

2016 ◽

Vol 41 (6) ◽

pp. 605-627 ◽

Cited By ~ 14

Author(s):

Jessaca Spybrook ◽

Benjamin Kelcey ◽

Nianbo Dong

Keyword(s):

Statistical Power ◽

Randomized Trials ◽

A Priori ◽

Individual Characteristics ◽

Cluster Randomized Trials ◽

Moderator Effects ◽

Effect Of Treatment ◽

Cluster Randomized ◽

Level Cluster ◽

The Impact

Recently, there has been an increase in the number of cluster randomized trials (CRTs) to evaluate the impact of educational programs and interventions. These studies are often powered for the main effect of treatment to address the “what works” question. However, program effects may vary by individual characteristics or by context, making it important to also consider power to detect moderator effects. This article presents a framework for calculating statistical power for moderator effects at all levels for two- and three-level CRTs. Annotated R code is included to make the calculations accessible to researchers and increase the regularity in which a priori power analyses for moderator effects in CRTs are conducted.

Download Full-text

Statistical approaches to computing sample size in cluster randomized trials: a simulation study

10.21203/rs.2.18978/v2 ◽

2020 ◽

Author(s):

Ashutosh Ranjan ◽

Guangzi Song ◽

Christopher S Coffey ◽

Leslie A McClure

Keyword(s):

Sample Size ◽

Cluster Size ◽

Randomized Trials ◽

Sample Size Calculation ◽

Minimum Variance ◽

Type I ◽

Cluster Randomized Trials ◽

Planning Stage ◽

Cluster Randomized ◽

The Impact

Abstract Background: Cluster randomized trials, which randomize groups of individuals to an intervention, are common in health services research when one wants to evaluate improvement in a subject's outcome by intervening at an organizational level. For many such trials, sample size calculation is performed under the assumption of equal cluster size. For a variety of reasons, many trials that set out to recruit clusters of the same size end up with unequal clusters. This leads to a misalignment between the method used for sample size calculation and the data analysis, which may affect trial power. Various weighted analysis methods for analyzing cluster means have been suggested to overcome the problem introduced by unbalanced clusters; however, the performance of such methods has not been evaluated extensively. Methods: We examine the use of the general linear model for analysis of clustered randomized trials that assume equal cluster sizes during the planning stage, but for which the realized cluster sizes are unequal. We demonstrate the performance of three approaches using different weights for analyzing the cluster means: (1) the standard analysis of cluster means, (2) weighting by cluster size, and (3) minimum variance weights. Several distributions are used to generate cluster sizes to assess a range of patterns of imbalance. The variability in cluster size is measured by the coefficient of variation (CV). We assess the impact of using each of the three methods of analysis with respect to type I error and power of the study and how each are impacted by the variability in cluster size via simulations. Results: Analyses that assumes equal clusters provide a reasonable approximation when cluster sizes vary minimally (CV < 0.30). For analyses weighted by cluster size type I errors were inflated, and that worsened as the variation in cluster size increases, despite reasonable power. However, minimum variance weighted analyses best maintain target power and level of significance under scenarios considered. Conclusion: Unweighted analyses work well as an approximate method when variation in cluster size is minimal. However, using minimum variance weights performs much better across the full range of variation in cluster size and is recommended.

Download Full-text

Analysis of contamination in cluster randomized trials of malaria interventions

Trials ◽

10.1186/s13063-021-05543-8 ◽

2021 ◽

Vol 22 (1) ◽

Author(s):

Lea Multerer ◽

Tracy R. Glass ◽

Fiona Vanobberghen ◽

Thomas Smith

Keyword(s):

Randomized Trials ◽

Random Effect ◽

Sigmoid Function ◽

Cluster Randomized Trials ◽

Intervention Effectiveness ◽

Insecticide Treated Nets ◽

Approximately Unbiased ◽

Buffer Zones ◽

Unbiased Estimates ◽

Cluster Randomized

Abstract Background In cluster randomized trials (CRTs) of interventions against malaria, mosquito movement between households ultimately leads to contamination between intervention and control arms, unless they are separated by wide buffer zones. Methods This paper proposes a method for adjusting estimates of intervention effectiveness for contamination and for estimating a contamination range between intervention arms, the distance over which contamination measurably biases the estimate of effectiveness. A sigmoid function is fitted to malaria prevalence or incidence data as a function of the distance of households to the intervention boundary, stratified by intervention status and including a random effect for the clustering. The method is evaluated in a simulation study, corresponding to a range of rural settings with varying intervention effectiveness and contamination range, and applied to a CRT of insecticide treated nets in Ghana. Results The simulations indicate that the method leads to approximately unbiased estimates of effectiveness. Precision decreases with increasing mosquito movement, but the contamination range is much smaller than the maximum distance traveled by mosquitoes. For the method to provide precise and approximately unbiased estimates, at least 50% of the households should be at distances greater than the estimated contamination range from the discordant intervention arm. Conclusions A sigmoid approach provides an appropriate analysis for a CRT in the presence of contamination. Outcome data from boundary zones should not be discarded but used to provide estimates of the contamination range. This gives an alternative to “fried egg” designs, which use large clusters (increasing costs) and exclude buffer zones to avoid bias.

Download Full-text

Covariate-constrained Randomization Routine for Achieving Baseline Balance in Cluster-randomized Trials

The Stata Journal Promoting communications on statistics and Stata ◽

10.1177/1536867x1701700214 ◽

2017 ◽

Vol 17 (2) ◽

pp. 503-510 ◽

Cited By ~ 2

Author(s):

Eva Lorenz ◽

Sabine Gabrysch

Keyword(s):

Randomized Trials ◽

Theoretical Background ◽

Cluster Randomized Trials ◽

Cluster Randomized ◽

Baseline Covariates

In cluster-randomized trials, groups or clusters of individuals, rather than individuals themselves, are randomly allocated to intervention or control. In this article, we describe a new command, ccrand, that implements a covariate-constrained randomization procedure for cluster-randomized trials. It can ensure balance of one or more baseline covariates between trial arms by restriction to allocations that meet specified balance criteria. We provide a brief overview of the theoretical background, describe ccrand and its options, and illustrate it using an example.

Download Full-text