ACLUSTER - Software for the Design and Analysis of Cluster Randomized Trials

ACLUSTER is designed to facilitate the analysis of data from cluster randomized trials (trials in which the unit of randomization is a cluster-family, clinic, hospital, etc.-as apposed to an individual study subject.) The program covers methods that are not available in standard statistical packages for randomized clinical trials. ACLUSTER takes you through the following steps: computation of the intraclass correlation coefficient; estimation of the sample size for three types of cluster randomized study design, i.e. completely randomized, paired-matched, and stratified; use of closed-formulae techniques for the analysis of binary outcomes for the three study designs; use of simple techniques for the analysis of continuous outcomes for all the three study designs; use of simple techniques for the analysis of count and time-to-event outcomes; and random allocation of clusters to the arms of the trial. The software has been prepared by Mr. Alain Pinol and Dr. Gilda Piaggio of the UNDP /UNFPA /WHO /World Bank Special Programme of Research, Development and Training in Human Reproduction of the World Health Organization.

Version 2.1 Update
Version 2.1 of ACLUSTER is available as a free update for anyone who has purchased a previous version of the software. Version 2.1 of ACLUSTER resolves several issues, these are as follows;

• Correction of a problem in the probability computation routine that caused a "freeze" of the program
• Development of a set of routines for the reading of data from Excel files saved as CSV, TXT or PRN file type. The first row must contain a variable name for each column (or variable). Definition of a dictionary for these files is not necessary.
• Fixing of minor problems in several functions/panels of the program
• Correction of erroneous messages and of typographical mistakes.

DOWNLOAD VERSION 2.1

Technical Support
If you require any technical support for Acluster please email acluster@who.int