Skip to main content

Table 1 Characteristics of eligible trials (N = 101)

From: Evidence of unexplained discrepancies between planned and conducted statistical analyses: a review of randomised trials

CharacteristicN (%)
Journal (n, %)
Annals of Internal Medicine3 (3%)
The BMJ3 (3%)
JAMA19 (19%)
Lancet28 (28%)
NEJM42 (42%)
PLOS Medicine6 (6%)
Funding (n, %)
 Pharmaceutical21 (21%)
 Other for-profit medical company8 (8%)
 Government37 (37%)
 Charity5 (5%)
 Multiple including pharmaceutical/other for-profit medical4 (4%)
 Multiple excluding pharmaceutical/other for-profit medical22 (22%)
 Others4 (4%)
Type of intervention (n, %)
 Pharmacologic52 (51%)
 Surgical13 (13%)
 Psychosocial/behavioural/educational9 (9%)
 Others24 (24%)
 Multiple types3 (3%)
Cluster trial (n, %)14 (14%)
Factorial trial (n, %)3 (3%)
Crossover trial (n, %)3 (3%)
Non-inferiority trial (n, %)20 (20%)
No. of treatment arms (n, %)
 Two85 (84%)
 Three or more16 (16%)
Sample size
 Median, IQR758 (306, 2129)
 Min, max36, 415357