Paperonium

Liver transplants · J Hepatol · 2020

“Liver-transplant patients with COVID die less often.” 20 deaths, against 21 expected, can’t show that.

The paper:3.2% to 5.8% fewerclearly fewer deathsIts 20 deaths: 42% fewerto 48% moreno clear difference4.5% fewer0.6×0.8×1×1.2×1.5×as many deaths as expectedfewer deathsmore
The paper: 3.2% to5.8% fewerclearly fewer deathsIts 20 deaths:42% fewerto 48% moreno clear difference4.5% fewer0.5×0.7×1×1.5×as many deathsas expectedfewer deathsmore

Deaths as a multiple of the expected number, with each range

The paper says

111 Spanish liver-transplant patients got COVID; 20 died, against about 21 expected at the death rate of Spaniards of their age and sex with COVID. The paper says that is 4.5% fewer deaths, with the true figure in a range of 3.2% to 5.8% fewer.

Its own data show

Working out the range from its 20 deaths, as its methods say it did, gives 42% fewer to 48% more, which includes no difference. The printed range is what 20,741 deaths would give: the paper’s estimate, in its extra tables, of how many of Spain’s COVID patients would die at the transplant patients’ rate.

A range is the span of values the data allow, and the fewer the cases, the wider it gets. So 20 deaths can’t back the paper’s claim, made five times, that these patients die less often, or its suggestion that their anti-rejection drugs, which damp down the immune system, may protect them from severe COVID.

“The summary” means the paper’s abstract: the short summary printed at its top, which is the part most readers see.

Read the paper →Full technical write-up →

Epidemiological pattern, incidence, and outcomes of COVID-19 in liver transplant patients · J Hepatol (2020) · doi:10.1016/j.jhep.2020.07.040