So what’s up, I’m not expert but I do have a university background that includes research processes, pharmacology, and statistics.
So the purpose of this study investors.gilead.com/node/36871/pdf was to examine whether a 5 versus 10 day course of remdesivir was preferable. It wasn’t whether it was effective, it wasn’t whether it was safe, simply comparing the 5 and 10 day course of treatment.
So for any drug you expect what’s called a dose-response curve, which includes things like ed50, ld50, etc., and the more of a substance a person is given, the more of a response up until it’s an effective dose, and if you keep going it’s lethal.
Anyhoo, the one major finding that can be taken from this paper as released isn’t that remdesivir helped anyone, it isn’t that remdesivir was safe or effective, it was that the longer a patient was on remdesivir the worse they did.
Quoth the paper,
“These data are encouraging as they indicate that patients who received a shorter, 5-day course of remdesivir experienced similar clinical improvement as patients who received a 10-day treatment course,” said Aruna Subramanian, MD, Clinical Professor of Medicine, Chief, Immunocompromised Host Infectious Diseases, Stanford University School of Medicine, and one of the lead investigators of the study. “While additional data are still needed, these results help to bring a clearer understanding of how treatment with remdesivir may be optimized, if proven safe and effective.”
Basically in almost any circumstance you would not expect less treatment to yield a larger and better effect, other than if there was some sort of floor or ceiling effect. Instead the most probable explanation is that remdesivir doesn’t work.
They’re trying to pre-emptively spin that one niggling fact.
edit: otherwise this study was not controlled, blinded, or compared in any way, and wasn’t actually really a research study as most people would understand it. There are blinded and controlled studies, but those results would be unknown because they’re blinded.
edit 2: To make it clear, similar outcome = substantially worse for the 10 day group.
edit: they lay out their figures poorly, but deaths were actually 8% in the 5 day, and 11% in the 10 day. Way higher. All the figures are laid out in a really confusing way
Clinical recovery 129 (65) 106 (54) 0.17 Discharge 120 (60) 103 (52) 0.44 Death 16 (8) 21 (11) 0.70 Safety Any adverse event (AE) 141 (71) 145 (74) 0.86 Grade ≥3 study drug-related AE 8 (4) 10 (5) 0.65 Study drug-related serious adverse event (SAE) 3 (2) 4 (2) 0.73 AE leading to discontinuation 9 (5) 20 (10) 0.07
The last numbers seem to be p values, which if so means nothing was significant in this study, as p < .05 is the minimal standard for anything. So 5 day remdesivir condition shows a better recovery rate and a better death rate than the 10 day condition. The difference is non significant, therefore “similar”.
This is a POOR study result that is masquerading as a home run.
Disclaimer: This is a guest post and it doesn’t necessarily represent the views of IWB.