[BostonCard]

This man disagrees:

https://www.westernjournal.com/coronavir...aved-life/

Anecdotes ≠ data

That's not true; it's not an anecdote, it is artisanal data.

---

BC, the study NCT04261517 is the one reported on yesterday.  The other study of 150 patients, comparing HCQ versus Lopinavir/ritonavir, is in phase 2.

I'm surprised anyone with any credentials would jump on one small study (which seems flawed to me) and declare "nada" when there was another study of about the same size (and seemed flawed to me too) that seemed to have different results. 

30 treatment-naÏve patients, split randomly into a test group of 15 and a control group of 15.

I had to use Google Translate to pull some of the tabular results out    It indicates "Mean course of the disease" (I presume time since first symptom?) was 6.6 days for test group, 5.9 for control.  It indicates 9 of the test group had a fever while 13 of the control group had a fever.  That seems a bit odd to me.

The trail seems to only take one subsequent test for the virus, at 7 days, where 90% of all patients were without the virus.  That's not much room for discrimination.  That was the study's primary endpoint.

Only 1(in the test group) of the 30 became severe.  That's odd.   3% here versus 16% in the 1099-patient study.  Better outcome due to improved care?  How?

The test group median age was about 4 years older. (50.5 vs 46.7) which would bias things for worse outcome for the test group.
The test group had 5 of the 15 with hypertension, versus 3 in the control group.  The control group had one person with COPD. (Each group had one with diabetes.)
Overall 11 coexisting conditions (36%, 40% test group, 27% control group) compared to 24% in the 1099-patient study).

To me, the jury is still out.  My hopes are down though.

To be clear, I didn't jump on one study as dispositive (I didn't even read it until you pointed it out, because, other than being the first randomized trial, I would not consider it to be high quality).  Indeed, the conclusion of the study "Larger sample size study [sic] are needed to investigate the effects of HCQ in the treatment of COVID-19".  A thirty patient study with an endpoint of RT-PCR clearance of virus is not much evidence it doesn't work, other than the fact that you can probably rule out a massive effect (I'm talking an effect the size of the early penicillin "trials", which were not randomized or placebo controlled, but for which the effect was so obvious as to be unambiguous).  But I think we would all call a drug "successful" if it resulted in a reduction in the need for ventilators among patients admitted to the hospital from 30% to 20% or cut mortality among hospitalized patients from 10% to 5%, which is something you need a much bigger study to evaluate.

My point was just to jump on the "chloroquine cured me" case report published in the prestigious journal The Western Journal.  That is even less credible than an under-powered negative study.

BC

[BostonCard]

This man disagrees:

https://www.westernjournal.com/coronavir...aved-life/

Anecdotes ≠ data

That's not true; it's not an anecdote, it is artisanal data.

---

BC, the study NCT04261517 is the one reported on yesterday.  The other study of 150 patients, comparing HCQ versus Lopinavir/ritonavir, is in phase 2.

I'm surprised anyone with any credentials would jump on one small study (which seems flawed to me) and declare "nada" when there was another study of about the same size (and seemed flawed to me too) that seemed to have different results. 

30 treatment-naÏve patients, split randomly into a test group of 15 and a control group of 15.

I had to use Google Translate to pull some of the tabular results out    It indicates "Mean course of the disease" (I presume time since first symptom?) was 6.6 days for test group, 5.9 for control.  It indicates 9 of the test group had a fever while 13 of the control group had a fever.  That seems a bit odd to me.

The trail seems to only take one subsequent test for the virus, at 7 days, where 90% of all patients were without the virus.  That's not much room for discrimination.  That was the study's primary endpoint.

Only 1(in the test group) of the 30 became severe.  That's odd.   3% here versus 16% in the 1099-patient study.  Better outcome due to improved care?  How?

The test group median age was about 4 years older. (50.5 vs 46.7) which would bias things for worse outcome for the test group.
The test group had 5 of the 15 with hypertension, versus 3 in the control group.  The control group had one person with COPD. (Each group had one with diabetes.)
Overall 11 coexisting conditions (36%, 40% test group, 27% control group) compared to 24% in the 1099-patient study).

To me, the jury is still out.  My hopes are down though.

To be clear, I didn't jump on one study as dispositive (I didn't even read it until you pointed it out, because, other than being the first randomized trial, I would not consider it to be high quality).  Indeed, the conclusion of the study "Larger sample size study [sic] are needed to investigate the effects of HCQ in the treatment of COVID-19".  A thirty patient study with an endpoint of RT-PCR clearance of virus is not much evidence it doesn't work, other than the fact that you can probably rule out a massive effect (I'm talking an effect the size of the early penicillin "trials", which were not randomized or placebo controlled, but for which the effect was so obvious as to be unambiguous).  But I think we would all call a drug "successful" if it resulted in a reduction in the need for ventilators among patients admitted to the hospital from 30% to 20% or cut mortality among hospitalized patients from 10% to 5%, which is something you need a much bigger study to evaluate.

My point was just to jump on the "chloroquine cured me" case report published in the prestigious journal The Western Journal.  That is even less credible than an under-powered negative study.

BC