My second thought upon reading the headline is that the reference sample (people who don’t take the drug) isn’t comparable to the data they collected. People who feel good (survive) generally don’t get as much medical care and thus don’t show up in statistics. So thus I assumed that this study compared its data that would more likely include all people in the study with data that would leave out more survivors.

But this study made its own baseline data, so it has some merit. I did not look into sample size at all.