Skip to main content

Table 4 Percentages accounting for clustering in sample size calculations and analysis in previous reviews

From: A systematic review of cluster randomised trials in residential facilities for older people suggests how to improve quality

Authors Source of trials Years Clustering accounted for
In sample size In analysis
Donner et al. [5] 16 non-therapeutic Intervention trials 1979-1989 19% 50%
Simpson et al. [6] 21 trials from Amer J Pub Health Prev Med 1990-1993 19% 57%
Chuang et al. [7] 24 trials of computer based decision support 1975-1998 0% 58%
Isaakidis et al. [8] 51 trials in Sub-Saharan Africa 1973-2001 20% 37%
Puffer et al. [3] 36 trials in BMJ, Lancet and NEJM 1997-2002 56% 92%
Eldridge et al. [10] 152 trials in primary health care 1997-2000 9% 59%
Varnell et al. [39] 60 trials in Amer J PubHealth Prev Med 1998-2002 15% 54%
Bland [9] 18 articles in BMJ 1983-2003 n/a 72%
Eldridge et al. [4] 34 articles in primary health care 2004-2005 62% 88%
Murray et al. [38] 75 trials in oncology 2002-2006 24% 45%
Bowater et al. [11] 35 trials in tropical parasitic diseases 1998-2007 29% 43%
Handlos et al. [24] 35 trials in maternal and child health 1998-2008 71% 80%
Froud et al. [27] 23 trials in oral health 2005-2009 65% 78%
Ivers et al. [12] 300 trials of cluster design randomly selected 2000-2008 33%h 70%
This review 73 trials in residential facilities for older people 1993-2010 27% i 74%
  1. h Calculated by us for consistency with other reports, as the reported 61% excludes those trials which did not report sample size calculations.
  2. i This increases to 47% when those that did not report sample size calculations are excluded.