---
title: "Via @[748510345:2048:Nash Yielding] "Some *very* promising results…"
date: 2020-04-30
source: facebook
type: Archer T. Ships shared a link.
---

# Via @[748510345:2048:Nash Yielding] "Some *very* promising results…

*April 30, 2020 · Facebook*

::::: {}
::: {}
[https://mainichi.jp/english/articles/20200425/p2a/00m/0na/007000c](https://mainichi.jp/english/articles/20200425/p2a/00m/0na/007000c){target="_blank"}
:::

::: {}
Via @\[748510345:2048:Nash Yielding\]\
\
\"Some \*very\* promising results from a potential COVID-19 treatment being tested in several places.\
\
Additional commentary (seen online):\
\
\"For those on ventilators, deaths during the study dropped from 21.3% to 7.3% (2/3 reduction). The mortality of all patients dropped from 8.5% to 1.4% (5/6 reduction). 84% fewer deaths. Wow! Side effects of treatment were minimal, also. This urgently needs to be (a) replicated in further research, and (b) rolled out \*\*immediately\*\* in all hospitals worldwide. Ivermectin is extremely safe and well tolerated and complications are likely to be rare and minor. It looks like we found the silver bullet we so desperately needed, folks. 🙏 If ivermectin is administered early, in a cocktail along with remdesivir or HCQ, we can likely save even those last 16% too.\"\
\
\"Ah, looks like they retroactively look at patient outcomes and do statistical regression to prove their result. Been burned too often by exactly such studies in just this pandemic. Hope someone does a randomized control group, double-blind test soon.\"\
\
\"RCTs are not going to be viable in this climate and in fact are very likely innately subject to detection and selection bias because of the situation. People need to stop pushing this meme. We have very good statistical analyses expertise compared to the past and acting like all other study types are useless is not helping.\" \<\-- Interesting point worth discussing further.\
\
\"But then when we look at the methodology, they didn't adjust for location. \[From the study: \'The cohort was derived from 169 hospitals across 3 continents (North America, Europe and Asia) and included 704 patients treated with Ivermectin. A similar sized propensity score matched cohort was developed from among 68,230 other hospitalized patients who were not treated with Ivermectin and matched for the variables described in the methodology.\'\] So for all we know, the whole subset \'n\' with the great outcomes comes from one region in one part of the world. Are we so sure that the base mortality rate is the same in every medical center in every continent? Or did this experiment simply detect the most extreme variation within the sample?\"
:::
:::::