Ricochet is the best place on the internet to discuss the issues of the day, either through commenting on posts or writing your own for our active and dynamic community in a fully moderated environment. In addition, the Ricochet Audio Network offers over 50 original podcasts with new episodes released every day.
Drug Trials
There is a lot of noise now about azithromycin and hydroxychloroquine as therapy for Wuhan flu. People are demanding drug trials, people are claiming 100% cure rates, people are calling for free access to the drugs.
There’s a problem here. You can’t do a proper drug trial during an epidemic. The results of treating a critical illness will vary with an astonishing number of co-factors, many of which cannot be foreseen as you begin. For example, treating menopause with estrogens will increase strokes and heart attacks if the drug is given orally, but not when it is given transdermally. That took about 40 years to figure out.
So you have to do a proper drug trial.
A proper drug trial takes an early observation (here, HC and Azo may be effective for Wuhan flu), creates a null hypothesis (here, “treatment with HC and Azo is no better than saline”), creates a set of endpoint characteristics which are studied (days on a respirator, death rate) and controlled (sex, age, co-morbidities).
A proper drug trial then calculates the difference in the endpoint and number of cases needing to be studied to make the results statistically significant, meaning there is less than a 5% chance that said results arose by chance. Here, the difference might be that a 50% difference in death rates between the two groups, after treating 100 patients in each group, would disprove the null hypothesis.
A proper drug trial then goes through the institutional research review board to get approval for experimentation on human subjects. This, in my small experience, is often the longest part of the process.
A proper drug trial then recruits patients, either directly or at the behest of their treating Docs.
A proper drug trial then blinds the investigators as to what they are using, by providing identical-looking supplies, here perhaps bags of IV solution, half providing the drug and half providing the placebo.
A proper drug trial then consults each patient and explains that this is a randomized, controlled trial, that you must give your informed consent to participate, that there is a 50% chance you will receive placebo.
A proper drug trial then enrolls consenting patients and distributes supplies to the participating institution’s pharmacies. The pharmacies use some random method, chosen by the trial designers, to allocate therapy versus placebo.
A proper drug trial then lets events run their course until some pre-determined time when they do an interim analysis. If the study results by then conclusively disprove the null hypothesis, the study ends (that is to say, if there IS a difference between treatment with HC/Azo and saline, the study ends). If the study results do not disprove the null hypothesis, the study continues. If the results show undue harm from the study drug, say the initiation of lethal cardiac arrhythmias, the study is stopped. This is what happened to my favorite bugaboo drug trial, the Women’s Health Initiative, in 2001.
A proper drug trial thus is neither easy to design and conduct, nor quick to happen.
In an epidemic, a crisis situation, you can’t do a proper drug trial. You might follow your gut instincts and treat people with a drug if you believe it will be helpful, then see how they do. Maybe you will get good results, maybe you won’t. You’ll never know if your intervention was responsible because of the failure to randomize and control for variables.
So my hat is off to the brave ICU, general med, ER and pulmonary Docs who are treating this epidemic by the seats of their pants. They can’t wait for results from a proper drug trial. Given that fact, the politicos who know more about medicine than we do and the commentators who demand perfection have to be ignored.
Perfection is the enemy of good. We can’t do a proper drug trial here. We have to make do.
Published in Healthcare
I have several times wanted to yell just this in Fauci’s face during press conferences.
What I find reassuring about this drug being prescribed for this virus is that the drug has been around for a long time as an antimalarial drug. It is not a brand-new drug that someone just invented in a lab somewhere.
The patient should sign some sort of legal waiver.
From what I have read about it, if it doesn’t work, the patient can still pursue other existing treatments afterward.
So unless the patient has a precluding heart condition, there’s no harm done for most patients who take it. At least that the impression I have from what I’ve read.
The FDA should make it clear that the use of this drug against this virus has not gone through the customary drug trials and they cannot guarantee any results (which they can’t do anyway for any drug at any time–this is not different from their blessing on any other drug). Then they should simply get out of the way in this.
Great Post.
How do you explain to the love ones of the those who will die because they were denied the Hydorxycloroquinine cocktail because you just had to perform your damn study just to satisfy your psychotic need to be reassured by the results of a clinical trial? The cocktail has been proven to be 100% effective in trials elsewhere by some of the most prestigious researchers in the world. What more do you need? The delay waiting for these clinical trials is killing thousands of people. Who should be held to account for their deaths? And don’t tell me it was just something we had to do because of some damn regulations. Regulations can be changed. Once someone is dead that fact cannot be changed.
This should be the normal FDA role: first do no harm. Let others demonstrate that something actually does good. Safety, not efficacy.
Great comments, thanks.
I don’t think, in the trenches, that there is any need to get informed consent for ad hoc use. Both drugs are approved by the FDA, albeit for other diseases, and there is no bar to using a drug off label. I do it every day (except now, being furloughed as my specialty is “non-essential”).
Which is one reason it infuriates me to see poohbahs telling Docs not to use these drugs. Get out of the effing way and let the brave Docs and Nurses in the trenches save some lives. If the stuff doesn’t work, nothing has been lost.
To be fair, I don’t believe the good doctor was advocating the outlawing of individual off-label trials (something that is fairly common, I believe). He was describing for us laymen what “proper” drug trials involve and demonstrating that as an impossibility in the current crisis.
To be fair, I don’t believe the good doctor was advocating the outlawing of individual off-label trials (something that is fairly common, I believe)
Neither was I. I was only talking about the current Hydroxycloroquinine trials.
I know of at least 2 ricochet members who are obsessed with p-values and seem less concerned with reducing the body count … all in the name of ‘science’
This, so much this. All of the drugs being tested are approved for a different indication, with studies on the side effects. Let doctors use their medical training and experience.
By the way, most medical facilities in my area are looking for any kind of medical professional to help with testing and general care. If you want to lend a hand, there are probably options.
Yeah, though we could do a comparison of treatments with statistically matched groups. There are three treatments that I hear are being tossed around – HCQ / Z-pack, the HIV combination drug Kaletra, and another anti-viral from Japan. Randomize people to one of the different likely treatment groups, and compare them statistically. I think people on Ricochet have mentioned this approach before.
Precisely.
There’s a danger that, in discussions about this, people will assume that “we are not able to do a proper drug trial” is synonymous with saying “we can learn nothing of value from the ad hoc testing we can do.” I run into this surprisingly often. Imperfect needn’t mean irrelevant and, as you observe, sometimes it’s all we have.
Are you kidding me?
You have been saying the opposite thing to me and @unsk
you are a hypocrite
This is not true. I don’t have a psychotic need. I don’t want a full-blown, ordinary clinical study. I just want a decent study with a randomized trial, including a control group, that could establish the effectiveness or ineffectiveness of this treatment with some statistical significance.
No, I’ve been saying pretty much exactly this. We just haven’t communicated very well. It’s okay. I’ve been eager to hear the results of the ongoing, imperfect, ad hoc, limited, desperate testing that’s going on. I think it has value.
Doc R, I was with you until the last 3 paragraphs.
I don’t want a full-blown “proper drug trial.” I’d like to see a decent clinical trial, with randomization, to see if it establishes effectiveness. This shouldn’t be very hard. It seems to me that doctors would want to establish effectiveness, too, at least in a preliminary way, instead of flying by the seat of their pants.
This is particularly important as different docs and teams seem to be trying slightly different treatments — different dosages, combination of hydrochloroquine with other drugs, and so on.
imperfect and desperate… those are not loaded comments
Well, it was a rollicking, spirited, rip-snorting discussion, I’ll grant you that. But let’s not let it spill over into this thread. Anyone interested can go savor the drama where it happened, here. ;)
There’s one thing I keep seeing that is being grossly overblown.
That’s the “heart condition” claim.
When you start digging up studies, the handful of truly adverse reactions to HCQ that involve prior cardiac issues tend to be 1) Very large doses, given for very unusual reasons, or 2) Long term treatment with HCQ, on top of a plethora of other drugs given to people with very bad health.
When you dig out the “surprise” issues with normal, short-term doses of HCQ and cardiac incidents, they’re rare enough that people write serious papers on single cases. It’s a “wow- this actually happened” sort of situation.
In normal use, on the short term, it’s just not that common, and can be detected early when it does happen.
I can’t stand passive aggressive behavior especially from men.
Passive aggressive men have zero balls.
It sucks to be the placebo rat in a drug trial.
There’s been a lot of talk about the “French Trials” conducted by Didier Raoult and his colleagues. A paper detailing (well, sort of) the treatment of 80 COVID-19 patients was published in late March. Link here.
There is an accepted risk stratification used in France; the paper describes it:
The patients:
Parsing this: about half the patients had pneumonia. Most were not terribly ill, or (maybe) not yet terribly ill.
About a month ago, JAMA reported on data from China. Lots of patients, which is good, but… China.
Case-fatality rate
2.3% (1023 of 44 672 confirmed cases)
14.8% in patients aged ≥80 years (208 of 1408)
8.0% in patients aged 70-79 years (312 of 3918)
49.0% in critical cases (1023 of 2087)
But if that data holds true in general, the expected mortality in the French study would be… hard to tell but not very high. Which is what the study found.
Potential confounding factors:
The samples for the coronavirus test were taken by nasopharyngeal swab; this is a potential confounding factor for the purported endpoint of “not contagious” since there are reports of the lower respiratory tract being a more common site from which to recover virus, and also there is reported fecal shedding even after clinical recovery.
Best practices for using CT in COVID-19 cases require the disinfection of the CT unit after a COVID-19 patient is scanned before a non-infected patient is examined; this takes well over an hour. This is another way in which accurate rapid testing would be extremely desireable: If you test positive, and CT is indicated use a dedicated scanner.
I’d bill this study as a waste of time and resources. It did find lowered viral load, but…
the goal is to save lives and reduce the transmission.
the goal is not to be published in the new england medical journal.
Doc Robert actually oversimplifies. There are ways to do statistically valid tests that do NOT use a Control Group who are given a placebo. At least this is what my sister, who has worked as a statistician for numerous drug companies, tells me.
In other words, we don’t have to deny access to a potentially life-saving drug to get results that even mathematicians could accept.
It would be nice if the French study added useful information to move us towards either goal but it doesn’t. Maybe if they publish another couple of hundred cases.
I think running antibody tests on Dr. Zelenko’s choroquine/azithromycin treated patients would be more useful.
It does seem that as long as someone is keeping decent records on people who weren’t part of the study, data on such persons ought to be usable. It won’t be quite as good, but it will not be worthless either.
As a physician who is now outside of the fray, I am quite sure that the majority of ER physicians who are working in the hot spots of covid19 activity are taking hydroxycloroquine and zinc as prophylaxis. These are people who are well versed in the distinctions between controlled studies and observational studies. The vast majority of clinical breakthroughs begin with observational trials followed by more rigorous blinded studies if results look promising. The observational evidence for the benefit of the combination of hydroxycloroquine, azythromycin (or other antibiotic) with zinc early in the disease process is rapidly becoming overwhelming.
Is the antibiotic just a chemical that luckily happens to correlate with less virus trouble?
Or is it a prophylaxis against bacterial infections setting in as a side-effect of the virus problems?
That is true, but so far it hasn’t been done, or done well.
The two things I’ve read in various places are:
1) It’s not really acting as an antibiotic, technically: they’re taking advantage of a side effect. Azithromycin helps inhibit mucus buildup deep in the lungs, in this case.
2) The other thing is that it supposedly helps zinc transport, which (also supposedly) inhibits virus replication. This is the more-iffy claim, but the doctors who have been using it claim they get better results with zinc supplements added to the HCQ and the Az, and that’s the most popular theory going for how that works.