Table 2 Comparison of qualities between small and large trials in meta-analyses in critical care medicine.
| Â | No. of patients in component trials | P value | |
|---|---|---|---|
| Â | Large trials (>100 per arm, n = 73) | Small trials (<100 per arm, n = 244) | Â |
Sequence generating | Â | Â | 0.031 |
   Adequate | 29 (59%) | 50 (41%) |  |
   Inadequate/unclear | 20 (41%) | 72 (59%) |  |
Allocation concealment | Â | Â | <0.001 |
   Adequate | 56 (82%) | 114 (51%) |  |
   Inadequate/unclear | 12 (18%) | 111 (49%) |  |
Blinding | Â | Â | <0.001 |
   Adequate | 49 (67%) | 99 (41%) |  |
   Inadequate/unclear | 24 (33%) | 145 (59%) |  |
Incomplete follow-up data addressed | Â | Â | <0.001 |
   Adequate | 36 (86%) | 56 (48%) |  |
   Inadequate/unclear | 6 (14%) | 61 (52%) |  |
Intention to treat | Â | Â | <0.001 |
   Yes | 38 (83%) | 79 (52%) |  |
   No/unclear | 8 (17%) | 72 (48%) |  |
Sample size calculation | Â | Â | <0.001 |
   Yes | 29 (88%) | 32 (44%) |  |
   No/unclear | 4 (12%) | 40 (56%) |  |
Year of publication | Â | Â | <0.001 |
   Before 2002 | 55 (75%) | 127 (52%) |  |
   After 2002 | 18 (25%) | 117 (48%) |  |