s it is a novel disease, inevitably there have been numerous suggestions for treatments for Covid-19, ranging from herbal tonics to the anti-malarial hydroxychloroquine, as advocated by the former US president. The question is: what treatments work best? It is not enough just to compare what happened to people who did or did not have the treatment, which may, for example, have been given to healthier patients. The only reliable method is to allocate volunteers at random to either receive the novel treatment or a control, and, if possible, neither they nor the medical team know which. Randomised trials reduce statistical biases and, if they are large enough, researchers can robustly say whether the intervention helps. The Randomised Evaluation of Covid-19 Therapy (Recovery) trials began in the UK more than a year ago and have been extraordinarily successful. The world’s largest Covid-19 trial organisation, with about 40,000 hospital patients so far taking part, Recovery takes advantage of the NHS to simultaneously run overlapping trials, so that each patient may be in many studies. Recovery trials have been hugely influential. Low-dose dexamethasone is a cheap and widely used steroid that reduced deaths in Covid patients receiving invasive mechanical ventilation by 36% (uncertainty interval 19%-49%) and those receiving oxygen by 18% (6%-28%). This finding alone is estimated to have saved 22,000 lives in the UK and more than a million worldwide. Another study showed that for every 25 patients treated with tocilizumab, one additional life would be saved. The trials have also found what doesn’t work: hydroxychloroquine showed no clear clinical benefit, and neither did convalescent plasma, another of Trump’s enthusiasms; they have not tested injecting bleach. Recovery is one of the UK’s greatest contributions to world health, all based on the deceptively simple principle of deciding treatment by the flip of a (digital) coin. It has demonstrated what can be achieved through efficient organisation, a national health service, dedicated medical teams and vast numbers of volunteers. David Spiegelhalter is chair of the Winton Centre for Risk and Evidence Communication at Cambridge. Anthony Masters is statistical ambassador for the Royal Statistical Society