Punchline - "Although not statistically significant, patients receiving remdesivir had a numerically faster time to clinical improvement than those receiving placebo among patients with symptom duration of 10 days or less (hazard ratio 1·52 [0·95–2·43])."
That said, this morning Gilead reported that in their own pivotal trial that the drug met it's primary endpoint.[1] What we don't know is how big of an effect it was.
In general, yes. In an emergency when there are no other obvious better alternatives at hand, looking at results with relatively poor p-values might still be a pretty reasonable thing to do.
But it doesn't mean that drug doesn't work -- it means that the sample was too small to say that it works. If the sample were 100,000, then that would mean that it doesn't work. This was around 225, which isn't tiny, but given that we are well-endowed with people dying of COVID who would rather not, it makes sense to do more testing. Which they call for.
Yes, you should not look at any result in an emergency you wouldn't otherwise - in fact you should probably only increase scrutiny given how much incentive there is to merely get a headline with your drug mentioned as a potential covid19 cure.
The OP is basically arguing for a "something must be done and this is something" reasoning.
Yes there is. Not having effective options against a new virus is the null hypothesis. So by all means do whatever research has promising leads but don't lower burden of proof for accepting something as clinically effective.
I'm picturing the Feds and States shelling out desperately needed money to purchase an almost useless drug because the people on the Titanic are desperate for anything rather than the drowning.
If it does indeed turn out to be useless (which is far from obvious at the moment), it'll be dropped, and not much money will have been spent.
It's unfortunate, but a lot of what we're trying right now will turn out to be useless. One or two recent analyses suggest that the (far more expensive) lockdowns might turn out to have been almost useless. That doesn't mean that trying them was the wrong thing to do.
It is like the passengers discussing if they must carry a box with 10 pounds of oranges in each life boat, find some paddles, a plastic cover, or paint the Eyes of Horus in them before launching them.
The first three look useful, but I'm not sure if they improve the chance of survival too much. You need some time to find them and enough room in the boat, so it is a trade off if they are useful at all.
No, that go against the current standards of scientific ethics.
“Not statistically significant” does not mean the effect does not exist. It means that the study provides weak evidence, but weak evidence ≠ no evidence. It tags things for further study.
This is one of those cases where people get too hung up on P values. You have to look at the bigger picture; look at the entire scientific process.
I mean, it does provide no evidence for the specific hypothesis.
It may provide evidence of other hypotheses.
So you're right, but you're also wrong, without making yourself more clear.
Poor precision in the use of language around clinical data has been a major problem in all of this, in the press especially, where for the most part they clearly don't give a damn and just have an agenda to spin in most cases. Because of this, scientists publishing ought to try to use language that ensures their work cannot be misinterpreted. The notes in the results section here about non-statistically significant aspects to the data fail that test, since an average reader would likely walk away with a false mental model of what the data is showing.
edit: my comment is wrong, i'm tired and haven't had enough coffee. the criticisms below are valid. i do have a point, but won't jam it in here.
> I mean, it does provide no evidence for the specific hypothesis.
Again, this is incorrect. It provides weak evidence, not no evidence. You talk about precise language here but you are conflating “not statistically significant” with “no evidence”. You can’t just handwave the difference between these two things.
Having a cutoff of statistical significance is a side effect of wanting to divide the world into a black-and-white binary categorization.
This is a heuristic which helps things stay within our limited reasoning capacity, cuts down on the noise level from all the activity in the scientific community to focus on what's important, and helps with inherently binary organizational decisions (e.g. should this experiment be published in a journal or not? Should we give this medicine to patients or not?)
A better approximation of optimal Bayesian reasoning would be continuous updating of your prior with no cutoff. Even a study with a very weak p-value in the right direction should in theory slightly tweak your priors in favor of the hypothesis. (But you also have to tweak slightly against the hypothesis for studies with very weak p-values in the wrong direction, and compensate for selection effects, e.g. if experiments with underwhelming or wrong-direction p-values are underpublished.)
Even weak p-values below the binary decision threshold contain information. It's just that effectively denoising and acting on that information is more difficult and fraught with pitfalls than binary decisions based on strict thresholds. So science mostly standardizes on the safer approach, although there are some people who try to wring precious drops of information out of experiments that fall below the usual statistical cutoffs (often by so-called meta-analysis that combines results of multiple studies).
And this ties into the whole field of experiment design, and the design of statistical tests. You often conduct an experiment with the goal of making some kind of binary decision. So you figure out what an acceptable risk of type I and type II error would look like, and then you can figure out how much data you need to collect, and where to put the cutoff when analyzing data.
But the binary decision “let’s use drug X to treat condition Y”, and the binary decision “let’s investigate the efficacy of drug X on condition Y” each come with radically different tolerances for type I and type II errors.
Statistical significance is not a bright-line test. The p value reported is the probability that the observed effects occur if there are actually no effects. Notably, p is always greater than zero.
In general, when less data is available, larger p values are considered more interesting, while when more data is available, p must be smaller to be interesting. In particle physics, p < 0.001 means almost nothing. In the particular quoted line, p = 0.07ish which is above the human-defined, conventional threshold of 0.05 but in this case there are almost no other useful data and so the phenomenon is interesting.
The issue with statistical significance is that it might have a strong effect in some people and no effect in most people and statistical significance cant detect that.
For example, statistically chocolate doesnt cause acne. Yet I 100% get acne from chocolate.
There is enough variance in people that many things that work anecdotally cannot be proven statistically.
We have hydroxychloroquine with z-paks- which seems to be pretty much the same effect. But much much cheaper. If you treat early with it- it works, late treatment- and it's not effective.
Lots of legitimate questions about statistical significance, but it's strong enough that Fauci likes it[1], which tells me it can't all be bunk (given how cautious he's been about *quine):
> White House health advisor Dr. Anthony Fauci said Wednesday that data from a coronavirus drug trial testing Gilead Sciences’ antiviral drug remdesivir showed “quite good news” and sets a new standard of care for Covid-19 patients.
> Speaking to reporters from the White House, Fauci said he was told data from the trial showed a “clear cut positive effect in diminishing time to recover.”
That's from a more powered US trial, not the one in China. The argument is that Giliad basically got "unlucky" with this China one (insignificant number of patients/deaths to nullify hypothesis Remdesivir does work)
Depending on the specifics of "time to recover", this could be pretty good news.
My wife became symptomatic with COVID-19 about a month ago. For the first ~8 days, she was in pretty rough shape, although happily she didn't need hospitalization. But the pace of recovery is slower than I've seen with any flu. She still needs to significantly limit her activity or else her pulse gets high and she gets short of breath. (The effect is much greater than you'd expect from someone being bedridden for 8 days.)
Each day is a little better, so hopefully she'll make a full recovery. But if I had to spitball a linear extrapolation of when she'll be back to 100%, I'd say another few weeks at the soonest.
Obviously we'd rather have a drug that saves lives and/or prevents long-term disability. But if remdesivir could have hastened my wife's recovery by several weeks, our kids and my professional work would have been much better off.
Notwithstanding its effectiveness... I'm not sure if this could quickly be produced on a large scale.
For those versed in organic chemistry: looks pretty complicated. Maybe someone with industry experience can guess how difficult this is to productionize on a global scale?
About a month ago Gilead said they have 1.5 million doses ready to distribute or in the last stages of production but that's only enough for ~360k courses (a course is a enough to treat one person). They said they hoped to have produced a total 1 million courses by the end of the year [1], so yeah it's not enough.
I honestly do not understand what the interpretation means:
"In this study of adult patients admitted to hospital for severe COVID-19, remdesivir was not associated with statistically significant clinical benefits. However, the numerical reduction in time to clinical improvement in those treated earlier requires confirmation in larger studies."
I interpreted this to mean that the outcomes were not improved by Remdesivir but perhaps the time to get to that outcome might have been improved (but not confirmed) which is a hospital load / economic benefit.
I believe that this is the correct interpretation.
From the findings section: Although not statistically significant, patients receiving remdesivir had a numerically faster time to clinical improvement than those receiving placebo among patients with symptom duration of 10 days or less (hazard ratio 1·52 [0·95–2·43])
In other words, the patients that were going to get better, got better regardless of if they got the drug or not. However, those that got the drug got better faster than those who got the placebo (but as gfodor points out, not a statistically significant change, but a trend).
Yeah, but it isn't statistically significant. So this claim is a statement about the data itself, not about anything beyond that. It does provide a reasonable direction for future research since it's a 'lead', but from a pure interventionary action basis, it seems to contain zero information.
Yeah, you're right about the significance... but (just using the data they provided in the summary), it does look like a trend, which might be better fleshed out with more numbers. It does seem to at least demonstrate the safety of the drug, but I'm not sure if that was the point of this study.
This was the trial with patients with severe disease, so there is still a decent amount to be learned about potential for effects earlier in treatment. But there are other ongoing studies for earlier timepoints.
I am interpreting, however, "not statistically significant" as: "looking without calculating it appears it is faster, once it's calculated it turns out that appearance could be just a noise, not a real signal."
From the limited data in the summary, it looks like the mean time to outcome (getting better) was decreased with drug treatment, but the 95% CI is pretty wide and includes "no effect" in the CI. Honestly, there isn't much data here to work with, so it's hard to know exactly what they saw. But, I believe that the last patient would have come off study a little over a week ago (from the summary and clinical trial criteria), so this is all moving very fast.
This sounds exactly like what tamiflu does, if you have the flu, get it prescribed early, your duration of having the flu goes from a week+ to 4-5 days. With that scenario being: taking the drug(tamiflu) very early. It is my suspicion that the earlier Remdesivir is taken the better the outcome. These patients are all somewhat late in the disease progression. So not surprised to see these results although the time to improve is a nice bonus.
It means the study group was big enough to rule out benefit for those who began treatment in severe stage but not big enough to establish benefits for those who began treatment in a mild stage, which in turn indicates an upper bound to such possible benefit.
I’ve just tried to write down my very clear initial understanding of this but I misunderstood. It seems we can only guess at what the authors mean from that rather unclear paragraph.
It means it didn't help people who already had severe symptoms, but might help people if it is given to them earlier.
This is actually quite well known when treating the flu with anti-virals. My doctor has said in previous years, if I wake up with full body ache and feel like it's the flu, come in immediately and he'll prescribe anti-virals. He said if I wait, there's no point, though.
Makes sense the anti-virals might only be enough to stop viral replication if given early on before it's all over the place.
And while neither effect was found to be stat sig, the “treat earlier” group was close to stat sig (eyeballing it, it probably would have been stat sig at 90% confidence) so that’s good enough reason to try again with a larger sample size. If the effect stayed roughly similar, a larger N would lead to a smaller confidence interval that would not include a non-effect if an odds ratio of 1, and therefore be stat sig evidence of benefit.
Edit: Also it looks like they didn’t find an effect overall, but if they slice the data to just those who were treated early, they’re closer to having something stat sig. Hopefully that was a pre-planned cut of the data, but either way the interpretation is still probably “Somewhat promising but needs more data.”
And the difficulty is that RDV is only available IV, so depending on what "early" means, you would have to come in to the hospital at the first sign of symptoms.
What the NAISD study and Fauci were talking about is that it accelerates time to recover (seemingly among those who were going to recover anyways), which is a big deal as that increases the hospital throughput.
No effect when given in late stage severe cases. Unsure effect when given earlier in disease progression but a reduction in “days to improve” in the early administration cohort merits additional research.
This happens a lot with covid drug research. Many times treatments given too late have no effect on improving outcomes. That is, if you give it too late, it cannot save you. However, there does appear to be some drugs that improve outcomes when given early in disease progression. Think of it like tamiflu. If you take it 2-3 days post onset, it works. If you take it past that timeframe, it does not work.
Nobody seems to be talking much about favipiravir, which seems a bit like not-invented-here syndrome. If it stood to make a lot of money for some US big pharma company I suspect we'd be hearing more about it.
It seems to me like people "talking much about" unproven drugs is the problem, and that favipiravir (and its patients) is lucky to have been spared much of that nonsense.
People want to talk about this stuff because it feels like a magic bullet. Viral infections don't have magic bullets[1], they just don't. We have to beat this with elbow great: huge amounts of testing and tracing once the outbreak is at a manageable baseline, and high uptake of mitigation strategies like social distancing until it gets there. And yes, that costs a lot of money.
[1] Rather: they do, but they're called vaccines and take time.
Not if the effect size is large. That would be proved in no time. The problem is so far all effect sizes are small, so it will take trials with large power (and longer term) to discern any positive benefit.
They do. Consider HIV/AIDS. Nowadays, someone who gets HIV-positive at the age of 20 can expect to live to 70. It's quite astounding what modern combination therapy can do.
And the current treatments for hepatitis C have mild side effects and usually eradicate the infection even if the infection has been going on for 20 years.
It sounds like the patients in the study were a bit far into the illness already. Maybe lots of damage is already done and they are just having to additionally cope with the side-effects of the drug and not getting much of the benefit due to it having been started too late.
I haven't read the full study, so I don't know whether those have been addressed, but judging from the summary, it may not have been a consideration. I've heard of a few other studies recently where this was completely overlooked; they pick patients who are the most severe and give them the drugs.
Ok, excuse me for just complaining, but this study setup seems... bad.
That is, I believe the prior indication was fairly low to null that antiviral drugs would significantly help the treatment of a severe respiratory infection -- in the first place. You don't give someone Tamiflu when they have a sufficiently severe flu pneumonia and fever to be hospitalized, do you?
Antivirals are usually indicated for treatment when an infection is still in its early, mild, pre-hospitalization stages. I understand that it's easier to recruit study subjects with severe COVID-19 than mild at this stage, but still, doesn't this study amount to looking for one's keys under the lamp-post because it's brighter there?
From a nurse in Wuhan who is a family friend: It seemed clear to her and her coworkers that Remdesivir was working when given early, but had little or no effect after the disease had progressed.
My take: sounds like great news for a product, since everyone would want this on hand to catch infections early. Unfortunately it is an IV treatment and very difficult to manufacture, and is therefore unlikely to be able to fill that role. I bought a few shares anyway.
Sounds like it (I've also heard of zinc being added in with HCQ and the z-pak as part of that treatment protocol, sometimes called the "Raoult protocol" and referred to hereafter as such for brevity).
Both the Raoult protocol and remdesivir appear to help if used earlier in the disease progression; neither seem to do anything if the disease has progressed past a certain point.
If early use is the key to efficacy (and if both demonstrate the same degree of efficacy, as it seems currently), the Raoult protocol is the only one that makes sense to scale up. The ingredients for the Raoult protocol are much more easily mass-produced and administered, whereas remdesivir is proprietary, very expensive, and IV-only (as other posters have noted).
My barely informed interpretation and resulting take here is that:
1) This really only shows any efficacy when given shortly after the onset of symptoms
2) Remdesivir is an expensive and IV only drug that is difficult to produce in high quantities
3) The combination of the above makes this not particularly useful for treating the general population - we do not want everyone that shows mild symptoms to be rushing to the doctor for a treatment that has to be provided by a professional, and by the time we know if the symptoms will be severe and life threatening it is too late for remdesivir to be useful
4) That would seem to make it most useful for people that have known risk factors and co-morbidities to be candidates for remdesivir use. I do not know if the results are similarly promising for people that have those criteria, however.
If these results hold up it's good news, but I don't know if it's great news - it seems like the factors involved here make it impractical for the large scale treatment of cv19
90 comments
[ 3.0 ms ] story [ 144 ms ] threadThat said, this morning Gilead reported that in their own pivotal trial that the drug met it's primary endpoint.[1] What we don't know is how big of an effect it was.
[1]https://endpts.com/gilead-pivotal-covid-19-study-of-remdesiv...
The OP is basically arguing for a "something must be done and this is something" reasoning.
It's unfortunate, but a lot of what we're trying right now will turn out to be useless. One or two recent analyses suggest that the (far more expensive) lockdowns might turn out to have been almost useless. That doesn't mean that trying them was the wrong thing to do.
We're living in unusually uncertain times.
The first three look useful, but I'm not sure if they improve the chance of survival too much. You need some time to find them and enough room in the boat, so it is a trade off if they are useful at all.
If it works! If it does not, in fact, work, no sample of any size will show that it works.
“Not statistically significant” does not mean the effect does not exist. It means that the study provides weak evidence, but weak evidence ≠ no evidence. It tags things for further study.
This is one of those cases where people get too hung up on P values. You have to look at the bigger picture; look at the entire scientific process.
It may provide evidence of other hypotheses.
So you're right, but you're also wrong, without making yourself more clear.
Poor precision in the use of language around clinical data has been a major problem in all of this, in the press especially, where for the most part they clearly don't give a damn and just have an agenda to spin in most cases. Because of this, scientists publishing ought to try to use language that ensures their work cannot be misinterpreted. The notes in the results section here about non-statistically significant aspects to the data fail that test, since an average reader would likely walk away with a false mental model of what the data is showing.
edit: my comment is wrong, i'm tired and haven't had enough coffee. the criticisms below are valid. i do have a point, but won't jam it in here.
Again, this is incorrect. It provides weak evidence, not no evidence. You talk about precise language here but you are conflating “not statistically significant” with “no evidence”. You can’t just handwave the difference between these two things.
It does provide some evidence: we are more likely to see this result in a world where the hypothesis is true than a world where it is false.
This is a heuristic which helps things stay within our limited reasoning capacity, cuts down on the noise level from all the activity in the scientific community to focus on what's important, and helps with inherently binary organizational decisions (e.g. should this experiment be published in a journal or not? Should we give this medicine to patients or not?)
A better approximation of optimal Bayesian reasoning would be continuous updating of your prior with no cutoff. Even a study with a very weak p-value in the right direction should in theory slightly tweak your priors in favor of the hypothesis. (But you also have to tweak slightly against the hypothesis for studies with very weak p-values in the wrong direction, and compensate for selection effects, e.g. if experiments with underwhelming or wrong-direction p-values are underpublished.)
Even weak p-values below the binary decision threshold contain information. It's just that effectively denoising and acting on that information is more difficult and fraught with pitfalls than binary decisions based on strict thresholds. So science mostly standardizes on the safer approach, although there are some people who try to wring precious drops of information out of experiments that fall below the usual statistical cutoffs (often by so-called meta-analysis that combines results of multiple studies).
But the binary decision “let’s use drug X to treat condition Y”, and the binary decision “let’s investigate the efficacy of drug X on condition Y” each come with radically different tolerances for type I and type II errors.
In general, when less data is available, larger p values are considered more interesting, while when more data is available, p must be smaller to be interesting. In particle physics, p < 0.001 means almost nothing. In the particular quoted line, p = 0.07ish which is above the human-defined, conventional threshold of 0.05 but in this case there are almost no other useful data and so the phenomenon is interesting.
For example, statistically chocolate doesnt cause acne. Yet I 100% get acne from chocolate.
There is enough variance in people that many things that work anecdotally cannot be proven statistically.
There's a reason these are drug TRIALS. They're TRYING things.
"It doesn't seem to work" would be an important finding, and one that has nothing to do with Trump.
> White House health advisor Dr. Anthony Fauci said Wednesday that data from a coronavirus drug trial testing Gilead Sciences’ antiviral drug remdesivir showed “quite good news” and sets a new standard of care for Covid-19 patients.
> Speaking to reporters from the White House, Fauci said he was told data from the trial showed a “clear cut positive effect in diminishing time to recover.”
[1] https://www.cnbc.com/2020/04/29/dr-anthony-fauci-says-data-f...
It also means this isn't any sort of "miracle" drug though.
I hope other viable treatments emerge from the other trials on drugs set to close by June.
My wife became symptomatic with COVID-19 about a month ago. For the first ~8 days, she was in pretty rough shape, although happily she didn't need hospitalization. But the pace of recovery is slower than I've seen with any flu. She still needs to significantly limit her activity or else her pulse gets high and she gets short of breath. (The effect is much greater than you'd expect from someone being bedridden for 8 days.)
Each day is a little better, so hopefully she'll make a full recovery. But if I had to spitball a linear extrapolation of when she'll be back to 100%, I'd say another few weeks at the soonest.
Obviously we'd rather have a drug that saves lives and/or prevents long-term disability. But if remdesivir could have hastened my wife's recovery by several weeks, our kids and my professional work would have been much better off.
For those versed in organic chemistry: looks pretty complicated. Maybe someone with industry experience can guess how difficult this is to productionize on a global scale?
https://en.m.wikipedia.org/wiki/File:Synthesis_of_Remdesivir...
https://www.acsh.org/news/2020/03/26/problem-remdesivir-maki...
[1] https://www.fiercepharma.com/manufacturing/gilead-to-donate-...
"In this study of adult patients admitted to hospital for severe COVID-19, remdesivir was not associated with statistically significant clinical benefits. However, the numerical reduction in time to clinical improvement in those treated earlier requires confirmation in larger studies."
From the findings section: Although not statistically significant, patients receiving remdesivir had a numerically faster time to clinical improvement than those receiving placebo among patients with symptom duration of 10 days or less (hazard ratio 1·52 [0·95–2·43])
In other words, the patients that were going to get better, got better regardless of if they got the drug or not. However, those that got the drug got better faster than those who got the placebo (but as gfodor points out, not a statistically significant change, but a trend).
This was the trial with patients with severe disease, so there is still a decent amount to be learned about potential for effects earlier in treatment. But there are other ongoing studies for earlier timepoints.
https://www.gilead.com/purpose/advancing-global-health/covid...
https://clinicaltrials.gov/ct2/show/NCT04257656
This is actually quite well known when treating the flu with anti-virals. My doctor has said in previous years, if I wake up with full body ache and feel like it's the flu, come in immediately and he'll prescribe anti-virals. He said if I wait, there's no point, though.
Makes sense the anti-virals might only be enough to stop viral replication if given early on before it's all over the place.
Edit: Also it looks like they didn’t find an effect overall, but if they slice the data to just those who were treated early, they’re closer to having something stat sig. Hopefully that was a pre-planned cut of the data, but either way the interpretation is still probably “Somewhat promising but needs more data.”
This happens a lot with covid drug research. Many times treatments given too late have no effect on improving outcomes. That is, if you give it too late, it cannot save you. However, there does appear to be some drugs that improve outcomes when given early in disease progression. Think of it like tamiflu. If you take it 2-3 days post onset, it works. If you take it past that timeframe, it does not work.
However they found exploratory results in the data that they think deserve further study.
That's something.
People want to talk about this stuff because it feels like a magic bullet. Viral infections don't have magic bullets[1], they just don't. We have to beat this with elbow great: huge amounts of testing and tracing once the outbreak is at a manageable baseline, and high uptake of mitigation strategies like social distancing until it gets there. And yes, that costs a lot of money.
[1] Rather: they do, but they're called vaccines and take time.
So, by definition, we will only have unproven drugs to use for the foreseeable future.
But in the legal sense, my impression is that FDA never approves anything in under 2 years. Happy to be proven wrong!
They do. Consider HIV/AIDS. Nowadays, someone who gets HIV-positive at the age of 20 can expect to live to 70. It's quite astounding what modern combination therapy can do.
Antiviral drugs won't reverse damage that's already done.
There is newer, positive data coming out today from the ongoing study of advanced covid: https://www.niaid.nih.gov/news-events/nih-clinical-trial-sho...
That is, I believe the prior indication was fairly low to null that antiviral drugs would significantly help the treatment of a severe respiratory infection -- in the first place. You don't give someone Tamiflu when they have a sufficiently severe flu pneumonia and fever to be hospitalized, do you?
Antivirals are usually indicated for treatment when an infection is still in its early, mild, pre-hospitalization stages. I understand that it's easier to recruit study subjects with severe COVID-19 than mild at this stage, but still, doesn't this study amount to looking for one's keys under the lamp-post because it's brighter there?
My take: sounds like great news for a product, since everyone would want this on hand to catch infections early. Unfortunately it is an IV treatment and very difficult to manufacture, and is therefore unlikely to be able to fill that role. I bought a few shares anyway.
Both the Raoult protocol and remdesivir appear to help if used earlier in the disease progression; neither seem to do anything if the disease has progressed past a certain point.
If early use is the key to efficacy (and if both demonstrate the same degree of efficacy, as it seems currently), the Raoult protocol is the only one that makes sense to scale up. The ingredients for the Raoult protocol are much more easily mass-produced and administered, whereas remdesivir is proprietary, very expensive, and IV-only (as other posters have noted).
Does remdesivir work on coronavirus in a similar fashion?
Tamiflu blocks influenza viruses from being released from its host cell (https://en.wikipedia.org/wiki/Neuraminidase_inhibitor)
Remdesivir interferes with viral replication itself (https://en.wikipedia.org/wiki/Remdesivir)
My barely informed interpretation and resulting take here is that:
1) This really only shows any efficacy when given shortly after the onset of symptoms
2) Remdesivir is an expensive and IV only drug that is difficult to produce in high quantities
3) The combination of the above makes this not particularly useful for treating the general population - we do not want everyone that shows mild symptoms to be rushing to the doctor for a treatment that has to be provided by a professional, and by the time we know if the symptoms will be severe and life threatening it is too late for remdesivir to be useful
4) That would seem to make it most useful for people that have known risk factors and co-morbidities to be candidates for remdesivir use. I do not know if the results are similarly promising for people that have those criteria, however.
If these results hold up it's good news, but I don't know if it's great news - it seems like the factors involved here make it impractical for the large scale treatment of cv19