Table 2.
Characteristics of pilot cluster randomised trials included in this review
| Characteristic | Number of trials (%) |
| Publication year (earlier of the print and electronic publication date) | |
| 2010* | 1 (6) |
| 2011 | 3 (17) |
| 2012 | 3 (17) |
| 2013 | 7 (39) |
| 2014 | 4 (22) |
| Country | |
| UK | 10 (56) |
| Canada | 3 (17) |
| USA | 2 (11) |
| Germany | 1 (6) |
| New Zealand | 1 (6) |
| Australia | 1 (6) |
| Method of cluster randomisation† | |
| Simple | 1 (8) |
| Stratified with blocks | 9 (69) |
| Blocked only | 2 (15) |
| Bias coin method | 1 (8) |
| Number of clusters randomised‡ | |
| Median (IQR) | 8 (4 to 16) |
| Range | 2 to 50 |
| Average cluster size§ | |
| Median (IQR) | 32 (14 to 82) |
| Range | 7 to 588 |
*One paper has an extracted publication year outside of the 2011–2014 range. This is because the print publication date for this paper was 2011 but the online publication date was 2010, so the paper satisfies the inclusion criteria which states that the publication date, print or electronic, must be between 2011 and 2014, but we extract the earlier of the print and electronic dates.
†13 of the 18 trials reported their method of randomisation. Percentages are given as a percentage of these 13 trials.
‡Not reported for one trial.
§Defined as number of individuals randomised divided by number of clusters randomised, based on 12 trials that reported information on both.