Get involved

Claim

Opening the claim…

For each of the 15 largest comparisons a publication, a regulator's review or a sponsor's report on its placebo-controlled period was located (15 located). For any serious adverse event, 12 of them give a count of the same outcome over the same window in both arms (1 computed from a published percentage) and 3 do not; with the usable counts in place of the registry's the risk ratio is 0.98 (95% CI 0.95 to 1.00) against 0.98 (0.96 to 1.01) with the registry's. For death, 13 give usable counts and 2 do not: with the 10 whose registry record has a deaths table swapped it is 0.94 (0.90 to 0.99), and with the 3 whose registry record has no deaths table added as well, 0.97 (0.93 to 1.01), against 0.96 (0.91 to 1.01). · sciencejournal.ai