Planning a cluster randomized trial with unequal cluster sizes: practical issues involving continuous outcomes

BackgroundCluster randomization design is increasingly used for the evaluation of health-care, screeening or educational interventions. At the planning stage, sample size calculations usually consider an average cluster size without taking into account any potential imbalance in cluster size. However, there may exist high discrepancies in cluster sizes.MethodsWe performed simulations to study the impact of an imbalance in cluster size on power. We determined by simulations to which extent four methods proposed to adapt the sample size calculations to a pre-specified imbalance in cluster size could lead to adequately powered trials.ResultsWe showed that an imbalance in cluster size can be of high influence on the power in the case of severe imbalance, particularly if the number of clusters is low and/or the intraclass correlation coefficient is high. In the case of a severe imbalance, our simulations confirmed that the minimum variance weights correction of the variation inflaction factor (VIF) used in the sample size calculations has the best properties.ConclusionPublication of cluster sizes is important to assess the real power of the trial which was conducted and to help designing future trials. We derived an adaptation of the VIF from the minimum variance weights correction to be used in case the imbalance can be a priori formulated such as "a proportion (γ) of clusters actually recruit a proportion (τ) of subjects to be included (γ ≤ τ)".

[1]  O. Ukoumunne.  A comparison of confidence interval methods for the intraclass correlation coefficient in cluster randomized trials , 2002, Statistics in medicine.

[2]  J. Ioannidis,et al.  The Gini coefficient as a measure for understanding accrual inequalities in multicenter clinical studies. , 2004, Journal of clinical epidemiology.

[3]  D M Murray,et al.  Intraclass Correlation for Measures from a Worksite Health Promotion Study: Estimates, Correlates, and Applications , 1999, American journal of health promotion : AJHP.

[4]  J. Sterne,et al.  Methods for evaluating area-wide and organisation-based interventions in health and health care: a systematic review. , 1999, Health technology assessment.

[5]  Deborah Ashby,et al.  Lessons for cluster randomized trials in the twenty-first century: a systematic review of trials in primary care , 2004, Clinical trials.

[6]  S M Kerry,et al.  Unequal cluster sizes for trials in English and Welsh general practice: implications for sample size calculations. , 2001, Statistics in medicine.

[7]  Neil Klar,et al.  Sample size re‐estimation in cluster randomization trials , 2002, Statistics in medicine.

[8]  D. Altman,et al.  CONSORT statement: extension to cluster randomised trials , 2004, BMJ : British Medical Journal.

[9]  Geert Molenberghs,et al.  Linear Mixed Models in Practice , 1997 .

[10]  Liam Smeeth,et al.  Intraclass correlation coefficients for cluster randomized trials in primary care: data from the MRC Trial of the Assessment and Management of Older People in the Community. , 2002, Controlled clinical trials.

[11]  S. Chinn,et al.  Intraclass correlation coefficient and outcome prevalence are associated in clustered binary data. , 2005, Journal of clinical epidemiology.

[12]  R. Littell.  SAS System for Mixed Models , 1996 .

[13]  A C Wagenaar,et al.  Intraclass correlations from a community-based alcohol prevention study: the effect of repeat observations on the same communities. , 2000, Journal of studies on alcohol.

[14]  F B Hu,et al.  Intraclass correlation estimates in a school-based smoking prevention study. Outcome and mediating variables, by sex and ethnicity. , 1996, American journal of epidemiology.

[15]  David M. Murray,et al.  Design and Analysis of Group- Randomized Trials , 1998 .

[16]  D. Jacobs,et al.  PARAMETERS TO AID IN THE DESIGN AND ANALYSIS OF COMMUNITY TRIALS: INTRACLASS CORRELATIONS FROM THE MINNESOTA HEART HEALTH PROGRAM , 1994, Epidemiology.

[17]  R. Reading,et al.  Cluster randomised trials in maternal and child health: implications for power and sample size , 2000, Archives of Disease in Childhood.

[18]  P. Ravaud,et al.  A priori postulated and real power in cluster randomized trials: mind the gap , 2005, BMC medical research methodology.

[19]  Richard J Lilford,et al.  Ethical issues in the design and conduct of cluster randomised controlled trials , 1999, BMJ.

[20]  Allan Donner,et al.  Design and Analysis of Cluster Randomization Trials in Health Research , 2001 .

[21]  Amita K. Manatunga,et al.  Sample Size Estimation in Cluster Randomized Studies with Varying Cluster Size , 2001 .

[22]  L. Ryan,et al.  Analysis of dichotomous outcome data for community intervention studies , 2000, Statistical methods in medical research.

[23]  B. Short,et al.  Intraclass correlation among measures related to tobacco use by adolescents: estimates, correlates, and applications in intervention studies. , 1997, Addictive behaviors.

[24]  D. Torgerson,et al.  Evidence for risk of bias in cluster randomised trials: review of recent trials published in three general medical journals , 2003, BMJ : British Medical Journal.

[25]  David M. Murray,et al.  Intraclass Correlation for Measures from a Middle School Nutrition Intervention Study: Estimates, Correlates, and Applications , 2001, Health education & behavior : the official publication of the Society for Public Health Education.

[26]  Evangelos Evangelou,et al.  Intraclass correlation coefficients for cluster randomized trials in primary care: the cholesterol education and research trial (CEART). , 2005, Contemporary clinical trials.