Hi Kumara,
Perhaps an illustration can help.
Incidentally, your questions reminded me on my work on regret theory (I helped developed something called “acceptable regret” theory at which you will not regret being wrong) , which forces you to explicitly consider trade-off between the consequences of failing to provide potential benefits vs unnecessary harming patients. You may ask yourself question like this:
How many more times is worse not treating the patient whom I possibly can benefit comparing with treating unnecessary someone whom I can possibly harm (=RV)?
Formally, this can be expressed as decision threshold (T)=RXharms/(RRR-RRR*RXharms)
Where RRR= relative risk reduction of a given treatment.
The equation tells you that you should treat ONLY if morbidity/mortality without Rx is greater than T (note that morbidity/mortality ranges from 0 to 1; thus if it is 0, you should always treat, if it is 1, you should never treat)
[The equation for including RV and multiple harms is a bit more complicated but let’s stick to one harms only and assume RV=1]
Harms: Lancet article lists the % of De-novo ventricular arrhythmia in control at 0.3% and 4.3% to 8.1% in patients using HCQ in the various combination with antibiotics . Let’s assume it is about 5%.
Benefits (effectiveness): The Lancet article also states that mortality was 9% in the control arm , and was 16.8% to 23.8% . So, RRR=0.
(This would also mean that you would treat if T>9%).
If you replace these numbers in the equation above, T>>>1, which means you should NEVER treat the patient.
However, you believe that the Lancet article is at high risk of bias (although as Mohammed reminded us high risk for bias is not the same as risk of high bias). Nevertheless, highest risk of bias that I am aware in terms of distorting true effect of treatment is about 50-70% (but is typically in 10-20% range). Let’s then assume that harms=2%. How high RRR should be to satisfy you that you should treat your patient at T=9%?
Solving the equation above, you get that you need RRR~ 22% in order to justify giving HCQ.
Do you really believe that HCQ is so efficacious? If yes, you really need data to support this belief. If not, then you are really harming your patients. [Most people counter this argument with remark: “I have never seen the patient with harms due to XXX”. The problem is that in our individual practice we can never have enough patients to get personal experience such as the collective experience reported in the Lancet report, which reports 96,000+ patients. It is famously said 40 years of the experience with one drug for one disease cannot compete with one, well-designed large RCT]
Of course, this just an illustration but provides you with an explicit framework how one can replace or augment his/her intuition.
Hopefully this may convince you to discontinue using HCQJ
Best
ben
From: Evidence based health (EBH) [mailto:[log in to unmask]] On Behalf Of Kumara Mendis
Sent: Tuesday, May 26, 2020 10:09 AM
To: [log in to unmask]
Subject: Re: Re Critical Appraisal of Lancet paper
Dear all
Just heard BBC news that Brazil is using HCQ for all patients despite the warning from the WHO.
One possible reason is that the President of Brazil is a good friend of President Trump!
This is where politics gets into healthcare!
a) Patients expectations, what they want and expect
c) Best evidence that guides the decisions
A typical scenario from a LMI country physician who has only a few choices...
So....patients wants a drug to cure or shorten the COVID-19 disease
As a clinicians who has treated maybe 10-50 or more patients with HCQ, may be 1-2 had only had heart problems and many have been discharged home. In addition CLQ has been known and used for decades to treat malaria....
The evidence so far is on the fence...(This is why I was asking for a critical appraisal of the Lancet paper....can anyone do this as critical appraisal is one aspect of what you do when you find evidence?)
So as a clinician I keep on treating patients (without doing clinical trials as I do not have time or funds...)
I cannot find any articles on comparison between HCQ and Remdesivir
Dear Vivek,
Yes, I can understand why those are all powerful arguments. But still insufficient evidence.
A priori reasoning that it works come from in vitro studies. But a priori reasoning it may do harm also exists - QT interval.
Yes, safety profile when used in rheumatological conditions and malaria is good but when used in people who have myocarditis from SARS-2??? Don’t have enough evidence yet
Hello respected everyone,
Today, I was watching an Indian press brief, where the concerned person briefed that the biological plausibility, evidence from in-vitro studies, and experience over the time of using hydroxychloroquine in malaria along with relative safety profile from local studies have prompted to use it as prophylaxis. A similar thought is running in my country where such indications go viral in social media and people start using it on their own. On the other hand, hardly any data come out about the effectiveness and safety from local patients to support the guideline recommendations. At times, it becomes very difficult to discuss evidence as common statements are "something is better than nothing during a pandemic." It further influences me to think "What would I do if I was that COVID-19 positive patient in the hospital ICU, would I opt to take it?"
https://en.wikipedia.org/wiki/Medicus_curat,_natura_sanat still have a place. The beauty of science lies in its many explanatory stories, and the wonderful findings that someinterventions cure disease, prolong life and diminish human suffering. Surely, with our unedifying past examples, we know by now that it is unethical to guess rather than test?
Hi Kumara,
Let’s assume the Lancet study is flawed, does this really justify giving Rx with unknown benefits and well known harms? That is, regardless if the study described in the Lancet is credible or not, what EBM has taught us is that the probability of harming patients when giving them not well tested treatments is much higher than benefited them (+ lead to waste of resources). On my last count, there are 30+ different treatments that have been claimed to have effect against SARS-CoV2 (from vitamin C to HCQ to plasma of patients who recovered from COVID19). Should we use them all?
This urge “to do something”, to “treat ourselves rather than patients”, or even to provide false hope is , unfortunately, characteristic of all physicians regardless where we practice. But, we ought to resist it. Or, better, we should use the opportunities like these to formally test the effects of these Rxs. For example, perhaps you have enough patients treated with HCQ that you can analyze and we all can learn from you. This is not criticism of your practice- tendency for overRx dominates the current practice of medicine all around the world- but I am sad to see that after more than quarter century EBM is still In its infancy.
Thanks for openly engaging about your practice - only by sharing our experience we can learn from each other
Sent from my iPad - excuse typos and brevity
I guess when you have no treatment at all fir Covid and you have a lots of experience with using Chloroquine for Malaria you want to do something fir your patients
Patients will also be happy to know that they are getting some specific medications
This is exactly the situation in most of the LMI countries
Isn’t it not how most of the treatments for diseases have come about?
But now with this Lancet study things take a different turn
We have done evidence of the efficacy of the treatment
My question is how good is the study to inform us and can it be generalized ?