Researchers from Oxford University had 'cherry-picked' data from animal studies to give a positive spin to the new vaccine in order to win extra funding. As a result, 2,800 infants in South Africa were given the MVA85A TB vaccine when there was no evidence that it was safe.
The researchers had dismissed any results from four animal studies that hadn't been positive about the new vaccine, which was designed as a booster for the standard BCG jab. The researchers were granted a further £8m ($10.9m) funding and, in 2009, tested the vaccine on small children in South Africa.
But a 2015 review of the original trials discovered that the data had been 'cherry-picked'; worse, it seemed the MVA85A vaccine impaired the effectiveness of the BCG. Some of the animals given the two vaccines were dying quicker, but this was not reported in the trial results.
(Source: BMJ, 2018; 360: j5845)