Good writeup, but made the mistake of assuming rel
Post# of 155485
(Total Views: 509)
Posted On: 03/08/2021 11:16:53 AM
Re: nicehedges #81870
Good writeup, but made the mistake of assuming relative risk reduction was absolute in figuring number needed to treat to save a life.
Steroids overall reduced critical (mech. ventilated) mortality by 12.1% absolute (41.4 to 29.3%). Serious (oxygen only) had less mortality improvement of just 2.9% (26.2 to 23.3%), and for those not on any oxygen, steroids made death MORE likely (14.0% to 17.8% on dex).
https://www.nejm.org/doi/full/10.1056/NEJMoa2021436
Similar to our results, but I think ours are better and on top of benefit from steroids, just that we have smaller numbers.
Looking forward to seeing all the actual numbers presented so we don't have to calculate them.
But a relative risk reduction of death of 24% is not the same as saving 1 in 4 subjects. If the mortality was 50% on placebo, and we reduced it 24%, that would be a mortality of 38% with leronlimab. That is a 12% absolute reduction in mortality, so only 1 in 8 patients would have life saved.
For smaller numbers, the benefit is even less. Say 30% on placebo, and 24% relative risk reduction would be 22.8%, or a 7.2% absolute reduction, which means need to treat 14 patients to save one life.
Why not release those *before* the conference, so people can look at them and ask questions???
Steroids overall reduced critical (mech. ventilated) mortality by 12.1% absolute (41.4 to 29.3%). Serious (oxygen only) had less mortality improvement of just 2.9% (26.2 to 23.3%), and for those not on any oxygen, steroids made death MORE likely (14.0% to 17.8% on dex).
https://www.nejm.org/doi/full/10.1056/NEJMoa2021436
Similar to our results, but I think ours are better and on top of benefit from steroids, just that we have smaller numbers.
Looking forward to seeing all the actual numbers presented so we don't have to calculate them.
But a relative risk reduction of death of 24% is not the same as saving 1 in 4 subjects. If the mortality was 50% on placebo, and we reduced it 24%, that would be a mortality of 38% with leronlimab. That is a 12% absolute reduction in mortality, so only 1 in 8 patients would have life saved.
For smaller numbers, the benefit is even less. Say 30% on placebo, and 24% relative risk reduction would be 22.8%, or a 7.2% absolute reduction, which means need to treat 14 patients to save one life.
Why not release those *before* the conference, so people can look at them and ask questions???
(6)
(0)