China
Africa
Explore chapters and articles related to this topic
Randomised trials in primary care
Published in Felicity Goodyear-Smith, Bob Mash, How to Do Primary Care Research, 2018
Gillian Bartlett-Esquilant, Miriam Dickinson, Tibor Schuster
Cluster randomised trials are particularly useful for primary care research where interventions are more likely to be under evaluation, rather than medications or individualised treatments.17 For this trial design, complete social units or groups, i.e. the clusters, are randomised to receive the intervention.18 For example, the unit of randomisation might be by clinic, or by practitioner, rather than by the patient who is receiving either the intervention or the control. This is done to prevent contamination that might occur when an educational intervention is applied within a family medicine clinic. For example, as providers often discuss treatment options or practice decisions, it would be impossible to blind the other providers in the clinic to the intervention and not have each other's behaviours affected. There may be other reasons why individual randomisation is not possible for logistical, financial or ethical reasons. As a result, the cluster trial design is commonly used for the evaluation of service delivery or policy intervention at the level of the cluster.
Introduction
Published in Shein-Chung Chow, Jun Shao, Hansheng Wang, Yuliya Lokhnygina, Sample Size Calculations in Clinical Research: Third Edition, 2017
Shein-Chung Chow, Jun Shao, Hansheng Wang, Yuliya Lokhnygina
Chapter 11 summarizes sample size calculation for dose-ranging studies, including the determination of minimum effective dose (MED) and maximum tolerable dose (MTD). Chapter 12 considers sample size calculation for microarray studies controlling false discovery rate (FDR) and family-wise error rate (FWER). Bayesian sample size calculation is discussed in Chapter 13. Sample size calculation based on nonparametrics for comparing means with one or two samples is discussed in Chapter 14. Chapter 15 discusses statistical methods for data analysis under a cluster randomized trial. Sample size requirements for analysis at the cluster level and at individual subject level (within the cluster) are examined.
Introduction
Published in Richard J. Hayes, Lawrence H. Moulton, Cluster Randomised Trials, 2017
Richard J. Hayes, Lawrence H. Moulton
In some trials, however, groups of individuals may be randomly allocated to treatment arms. These groups are referred to as clusters, and such trials are known as cluster randomised trials. Examples of clusters include schools, communities, factories, hospitals or medical practices, but there are many other possible choices. Depending on the definition of a cluster, there are a number of alternative terms for such trials, the most common of which are group randomised trials and community randomised trials. The term community trial is sometimes used, but is ambiguous since it does not distinguish between individually randomised trials carried out in the community and trials in which communities are randomised to treatment arms. We shall use the abbreviation CRT for cluster randomised trials throughout this book.
Efficient design of geographically-defined clusters with spatial autocorrelation
Published in Journal of Applied Statistics, 2022
Consider the linear model Y is an μ is the population mean, ϵ is an independent and identically distributed 17] shows that the effective sample size is given by 1) reduces to: n is the average number of individuals in a cluster, J is the number of clusters, and 35]. In the same way, one can derive the ‘design effect’ for various other covariance matrices with compound symmetry such as designs with repeated measures and different specifications of temporal autocorrelation. Hemming et al. [20] provide such a list in the context of cluster randomised trial design.
A pilot and feasibility study of the effectiveness of care mapping on person-centred care in neurorehabilitation settings
Published in Neuropsychological Rehabilitation, 2020
Laura J. E. Brown, Daniel Blake, Katherine Berry, Russell Sheldrick
The aims of this study were to conduct pilot and feasibility testing of a cluster-randomized controlled trial to assess the effectiveness of DCM-NR at increasing levels of PCC in neurorehabilitation settings. Seventy-one percent of staff briefed about the study consented to participate, although 30% of these were lost to follow up. Whilst DCM-NR and the trial were generally acceptable to participants, levels of exposure were relatively low. There was no evidence of contamination between conditions, although a quarter of staff did report working on at least one other ward during the timeframe of the study. A higher proportion of PCC behaviour was observed in the experimental group compared to the control group at follow-up, although this was due to a decrease in PCC behaviour in the control group that was not observed in the experimental group. High levels of variability in interaction types were also observed within conditions. Neither group showed a change in attitudes towards people with brain injury, and participants in the control, but not the experimental, group showed an increase in PCC self-efficacy over time. Taken together, the findings suggest that this cluster-randomized trial design is acceptable, and also highlights modifications to the protocol and outcome measures that could increase the success and theoretical value of a future large-scale study.
Effectiveness of a Problem-Solving, Story-Bridge Mental Health Literacy Programme in Improving Ghanaian Community Leaders’ Attitudes towards People with Mental Illness: A Cluster Randomised Controlled Trial
Published in Issues in Mental Health Nursing, 2020
Yaw Amankwa Arthur, Gayelene H. Boardman, Amy J. Morgan, Terence V. McCann
A parallel group cluster randomised controlled trial design was conducted adhering to the CONSORT guidelines for cluster randomised trials (Campbell et al., 2012; Schulz et al., 2010). Simple computer-generated randomisation to cluster was done off-site in Australia by another researcher not part of the recruitment process. Randomisation was concealed and emailed to the researcher in Ghana who allocated clusters to the control and intervention groups. This process enabled the researcher to consider socio-demographic composition and geographical distribution of districts, to allocate the clusters to group and reduce the effect of contamination.