Skip to main content

Table 3 Presence and Pre-specification of Adjusted Analyses in Published RCTs a

From: Ideal vs. real: a systematic review on handling covariates in randomized controlled trials

  Adjusted Analyses Present (N = 298)   Pre-Specified Adjustment (N = 249) b  
No Yes p-value No Yes p-value
N % N %   N % N %  
Type of Study      0.6856      0.3791
 Superiority 40 16 215 84   19 9 196 91  
 Non-Inferiority 5 22 18 78   1 6 17 94  
 Other 4 20 16 80   3 19 13 81  
Number of Arms      0.4884      0.3666
 2 37 16 199 84   19 10 180 90  
 3 5 21 19 79   3 16 16 84  
 4 4 15 22 85   1 5 21 95  
 5 or more 3 25 9 75   0 0 9 100  
Multicenter Study      0.0865      0.0458
 No 7 29 17 71   4 24 13 76  
 Yes 42 15 232 85   19 8 213 92  
Number of Centers, median (IQR) 8 (3,32) 25 (7,77) 0.1562 10 (2,63) 28 (8,77) 0.3187
Number of Subjects, median (IQR) 449 (124,1030) 580 (250,1489) 0.6507 700 (179,1408) 568 (267,1502) 0.6488
Length of Study, Years, median (IQR) 3 (2,4) 3 (2,5) 0.0996 3 (2,6) 3 (2,5) 0.6060
Months of Subject Follow-up, median (IQR) 6 (3,14) 12 (5,24) 0.1011 12 (3,24) 12 (6,23) 0.6069
Study Initiation Year      0.7281      0.0029
 Unable to Determine 3 60 2 40   0 0 2 100  
 Before 2000 1 6 16 94   5 31 11 69  
 2000–2004 12 16 63 84   7 11 56 89  
 2005–2009 22 18 101 82   9 9 92 91  
 2010 and later 11 14 67 86   2 3 65 97  
Primary Outcome      0.0526      0.2586
 Continuous 12 13 77 87   6 8 71 92  
 Binary 27 24 86 76   11 13 75 87  
 Time-to-Event 8 10 76 90   4 5 72 95  
 Other 2 17 10 83   2 20 8 80  
Covariates Involved in Allocation      0.0100      0.1980
 No 16 28 41 72   6 15 35 85  
 Yes 33 14 208 86   17 8 191 92  
Number of Covariates      0.0312      0.2949
 N/A 16 28 41 72   6 15 35 85  
 1 19 20 76 80   6 8 70 92  
 2 8 9 78 91   5 6 73 94  
 3 6 15 34 85   2 6 32 94  
 4 0 0 11 100   3 27 8 73  
 5 or more 0 0 9 100   1 11 8 89  
Baseline Test for Significant Differences      0.3033      0.7216
 Unable to Determine 2 40 3 60   0 0 3 100  
 No 30 18 137 82   12 9 125 91  
 Yes 17 13 109 87   11 10 98 90  
  1. aP-values correspond to logistic regression results
  2. bThis subset of 249 studies includes those studies reporting either adjusted analyses only (N = 87) or both unadjusted and adjusted (N = 162)
\