Table 3. Number of major discrepancies per funding source.
Published RCTs that were registered before or during trial completion and have major discrepancies with their trial registries and the effect of these discrepancies on the statistical significance of published outcomes, by funding source.
| Number of published RCTs | Total | Private | Public | Industry | Mixed | Undisclosed | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 109 | 8 | 20 | 45 | 33 | 3 | |||||||
| Number of published RCTs with major discrepancies between registry and publication | 53 | 49% | 2 | 25% | 12 | 60% | 17 | 38% | 19 | 58% | 3 | 100% |
| Number of major discrepancies between registry and publication | 118 | 3 | 32 | 36 | 40 | 7 | ||||||
| Registered primary outcomes demoted in publication | 30 | 25% | 0 | 0% | 12 | 38% | 6 | 29% | 12 | 30% | 0 | 0% |
| Registered primary outcomes omitted from publication | 47 | 40% | 1 | 33% | 12 | 38% | 14 | 39% | 14 | 35% | 6 | 86% |
| Unregistered primary outcomes added to publication | 30 | 25% | 1 | 33% | 6 | 19% | 13 | 36% | 9 | 23% | 1 | 14% |
| Registered secondary outcomes promoted in publication | 3 | 3% | 0 | 0% | 1 | 3% | 2 | 6% | 0 | 0% | 0 | 0% |
| Timing of assessment of primary outcomes differs | 8 | 7% | 1 | 33% | 1 | 3% | 1 | 3% | 5 | 13% | 0 | 0% |
| Number of major discrepancies between registry and publication | 118 | 3 | 32 | 36 | 40 | 7 | ||||||
| Did not report p-values | 87 | 74% | 1 | 33% | 21 | 66% | 27 | 75% | 32 | 80% | 6 | 86% |
| Reported p-values | 31 | 26% | 2 | 67% | 11 | 34% | 9 | 25% | 8 | 20% | 1 | 14% |
| Major discrepancy favors statistical significance | 25 | 81% | 2 | 100% | 11 | 100% | 5 | 56% | 7 | 88% | 0 | 0% |
| Major discrepancy does not favor statistical significance | 6 | 19% | 0 | 0% | 0 | 0% | 4 | 44% | 1 | 13% | 1 | 100% |
| Number of RCTs containing major discrepancies favoring statistical significance | 19 | 2 | 6 | 5 | 6 | 0 | ||||||