UCL Discovery
UCL home » Library Services » Electronic resources » UCL Discovery

Cluster randomised trials with different numbers of measurements at baseline and endline: Sample size and optimal allocation

Copas, AJ; Hooper, R; (2019) Cluster randomised trials with different numbers of measurements at baseline and endline: Sample size and optimal allocation. Clinical Trials 10.1177/1740774519873888. (In press). Green open access

[img]
Preview
Text
Full article as accepted.pdf - Accepted version

Download (427kB) | Preview

Abstract

Background/Aims: Published methods for sample size calculation for cluster randomised trials with baseline data are inflexible and primarily assume an equal amount of data collected at baseline and endline, that is, before and after the intervention has been implemented in some clusters. We extend these methods to any amount of baseline and endline data. We explain how to explore sample size for a trial if some baseline data from the trial clusters have already been collected as part of a separate study. Where such data aren’t available, we show how to choose the proportion of data collection devoted to the baseline within the trial, when a particular cluster size or range of cluster sizes is proposed. Methods: We provide a design effect given the cluster size and correlation parameters, assuming different participants are assessed at baseline and endline in the same clusters. We show how to produce plots to identify the impact of varying the amount of baseline data accounting for the inevitable uncertainty in the cluster autocorrelation. We illustrate the methodology using an example trial. Results: Baseline data provide more power, or allow a greater reduction in trial size, with greater values of the cluster size, intracluster correlation and cluster autocorrelation. Conclusion: Investigators should think carefully before collecting baseline data in a cluster randomised trial if this is at the expense of endline data. In some scenarios, this will increase the sample size required to achieve given power and precision.

Type: Article
Title: Cluster randomised trials with different numbers of measurements at baseline and endline: Sample size and optimal allocation
Location: England
Open access status: An open access version is available from UCL Discovery
DOI: 10.1177/1740774519873888
Publisher version: https://doi.org/10.1177/1740774519873888
Language: English
Additional information: This version is the author accepted manuscript. For information on re-use, please refer to the publisher’s terms and conditions.
Keywords: Cluster randomised trial, baseline data, sample size, power, efficiency, study design
UCL classification: UCL
UCL > Provost and Vice Provost Offices
UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences
UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences > Faculty of Population Health Sciences
UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences > Faculty of Population Health Sciences > Institute for Global Health
UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences > Faculty of Population Health Sciences > Institute for Global Health > Infection and Population Health
URI: https://discovery.ucl.ac.uk/id/eprint/10084167
Downloads since deposit
20Downloads
Download activity - last month
Download activity - last 12 months
Downloads by country - last 12 months

Archive Staff Only

View Item View Item