Hacker News Reader

Comment by tlb | Hacker News Reader

Not a randomized trial. In fact:

> Several comorbidities were significantly less frequent in the HCQ group, including cardiovascular diseases, arterial hypertension, chronic renal disease, neurological and cognitive disorders, solid cancer, obesity as well as the proportion of active smokers.

IOW, doctors gave HCQ to healthier people, so you can't conclude much from the death rate being lower. There may be some value in this paper if you really dig into the details, but the overall death rate isn't a meaningful indication of whether HCQ helps.

View on HN

>>> Several comorbidities were significantly less frequent in the HCQ group, including cardiovascular diseases, arterial hypertension, chronic renal disease, neurological and cognitive disorders, solid cancer, obesity as well as the proportion of active smokers.

>IOW, doctors gave HCQ to healthier people,

Literally the sentence after the one you quoted contradicts clarifies, though:

>> On the other hand, at admission, patients in the HCQ group appeared to be sicker as reflected by the higher frequency of radiological pneumonia, acute respiratory distress syndrome, ICU transfer within the 24h after admission and invasive ventilation support as well as the higher frequency of elevated LDH and CRP levels.

I wouldn't call that a contradiction or clarification. It's yet another confound and there's no reason to believe it cancels out the other ones.

In this case, people arriving sicker at the hospital with a given disease is not necessarily an indication of the likelihood of death with no treatment. For instance, robust young people may wait until they are extremely sick before heading to the hospital, while old or immunocompromised people may go at the first sign of illness. In that case, you would expect death rate to be negatively correlated with how sick people are when they arrive at the hospital.

And the VA observational study showing no HCQ benefit had the opposite problem, with those given HCQ much worse on many markers-of-health.

Non-randomized studies have wasted a lot of everyone's attention on this & related issues.

It's difficult (nigh impossible) to do a proper controlled and randomized trial during a pandemic. We wont see any definitive science done this year.

Thankfully, the UK went ahead and carried out a large-scale randomized controlled trial in the middle of the pandemic anyway despite the difficulty - otherwise we'd all be completely in the dark. Unfortunately, we seem to be pretty much the only country which did, which means that the evidence available is limited to the treatment and dosage combinations that happened to be included in that trial.

At least five peer-reviewed RCTs were published, more are on their way..Each of them negative. Science was done.

Not disagreeing with you, but you should probably include the next sentence:

> patients in the HCQ group appeared to be sicker as reflected by the higher frequency of radiological pneumonia, acute respiratory distress syndrome, ICU transfer within the 24h after admission and invasive ventilation support as well as the higher frequency of elevated LDH and CRP levels.

But yes, it does not look good for this study that the HCQ group tended to be younger and healthier.

The RECOVERY study on hydroxychloroquine was a randomised clinical trial from the UK with a large number of participants and concluded hydroxychloroquine had no clinical benefit:

> "A total of 1542 patients were randomised to hydroxychloroquine and compared with 3132 patients randomised to usual care alone. There was no significant difference in the primary endpoint of 28-day mortality (25.7% hydroxychloroquine vs. 23.5% usual care; hazard ratio 1.11 [95% confidence interval 0.98-1.26]; p=0.10). There was also no evidence of beneficial effects on hospital stay duration or other outcomes."

The press release from June: No clinical benefit from use of hydroxychloroquine in hospitalised patients with COVID-19

https://www.recoverytrial.net/news/statement-from-the-chief-...

> doctors gave HCQ to healthier people

It would be interesting to see someone build an IPTW model to estimate the average treatment effect of HCQ in all these non-randomized studies. My guess is that the effect would be statistically indistinguishable from zero.

Plus once they adjusted for the multiple variables the effect is not huge. This will probably be more valuable in a much larger meta-analysis.

Reply on news.ycombinator.com