Skip to main content

Table 4 Overview of main results for binary outcomes (n=120)

From: Evaluating agreement between bodies of evidence from randomized controlled trials and cohort studies in medical research: a meta-epidemiological study

Analysis Number of BoE-pairs Ratio of ratios; 95% CI Heterogeneity (I2 (%);τ2) 95% prediction interval
Main 120 1.04 (0.97 to 1.11) 69; 0.061 0.63 to 1.71
Stratified by type of binary effect measure
 Risk ratios 85 1.02 (0.94 to 1.11) 73; 0.072 0.60 to 1.75
 Odds ratios 30 1.11 (0.98 to 1.25) 48; 0.039 0.72 to 1.70
 Hazard ratios 5 1.01 (0.78 to 1.30) 31; 0.026 0.52 to 1.95
Stratified by degree of overall PI/ECO-similarity
 More or less identical 2 1.17 (0.90 to 1.51) 0; 0.00 -
 Similar but not identical 81 1.06 (0.99 to 1.14) 54; 0.034 0.73 to 1.54
 Broadly similar 37 0.99 (0.85 to 1.16) 82; 0.149 0.45 to 2.21
Stratified by type of interventiona
 Drugs 40 1.04 (0.89 to 1.21) 76; 0.139 0.48 to 2.24
 Invasive procedures 39 1.00 (0.91 to 1.10) 25; 0.011 0.79 to 1.26
 Nutrition 28 1.07 (0.98 to 1.16) 71; 0.023 0.77 to 1.48
Stratified by outcome-categorya
 All-cause mortality 28 0.94 (0.82 to 1.09) 80; 0.075 0.53 to 1.69
 Cardiovascular disease outcomes 26 1.12 (1.02 to 1.23) 43; 0.022 0.81 to 1.55
 Drug safety outcomes 20 1.06 (0.89 to 1.26) 67; 0.068 0.60 to 1.90
  1. BoE Body of evidence, CI Confidence interval, PI/ECO Population, intervention/ exposure, comparator, outcome
  2. aOnly results of the largest subgroups are shown; detailed results are reported in Additional file 1 (Figs. S2a-S7)