Sunday, June 28, 2020

Evidence based science and COVID-19

There is so much of confusing rather conflicting data regarding spread and lethality of COVID19 infection that people are divided about it based on what political ideology they believe in. There are "believers" and then there are "non-believers". Most likely truth lies somewhere in the middle. And given the crumbling economies everywhere in the world, it has become absolutely necessary to find the "truth". To me following questions are important to be asked from government agencies so as the real threat of this virus can be assessed: 

1. In the US alone, lots of people die naturally of "flu" every year, gravity of which can be can be seen in the fact that as many as 61,000 people died from influenza (flu) just in one year (2017-2018). 
So it will be interesting to know as to how many people died of regular "flu" during this pandemic times? Do we have overlap of these two separate "diseases" (flu vs. COVID19) in the number of the patients? 

2. Cardiovascular diseases are the number one killer in the US. 647,000 Americans die from heart disease each year—that's 1 in every 4 deaths. 
Would the agencies release the data from the same period (pandemic period) as to how many people died of cardiovascular diseases during these months? Are these numbers independent from COVID19 patients who died during this time or are they same? More important question is as to how many people died of cardiovascular diseases during this period who did not have COVID19? 

Same question can be asked regarding other major killers such as real number of deaths from stroke (number 2 killer), COPD (number 3 killer), cancer, etc. and some other infectious diseases for lower income countries. 

Get these numbers and measure the overlap between these diseases and COVID19 by subtracting the number of deaths from these diseases from COVID19 related deaths, that will tell the real impact of COVID19. In absence of these numbers, vested interests will continue to manipulate the people's mind to suit their political agenda and millions of people will continue to suffer from non-COVID19 related issues such as unemployment etc.

An example of "believers" vs. "non-believers":

Thursday, June 4, 2020

Controversial Lancet study on Hydroxychloroquine retracted



https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7255293/

In the above paper published on May 22, 2020, investigators concluded that they were unable to confirm a benefit of hydroxychloroquine or chloroquine, when used alone or with a macrolide, on in-hospital outcomes for COVID-19. Each of these drug regimens was associated with decreased in-hospital survival and an increased frequency of ventricular arrhythmias when used for treatment of COVID-19.


Soon after it was published, it was considered as a major finding- Hydroxychloroquine (HCQ) touted by the Mr. Trump as possible COVID-19 treatment looked to be not only ineffective, but downright deadly. This study used hospital records procured by a little-known data analytics company called Surgisphere to conclude that coronavirus patients taking chloroquine or hydroxychloroquine were more likely to show an irregular heart rhythm and were more likely to die in the hospital.


Soon after this paper published others in the field started raising concern regarding the mysterious database company that supplied the content for this whole study. The database belongs to Surgisphere Corporation whose founder and CEO, is Dr. Sapan Desai, who is a lead co-author of the study. The Lancet started its internal investigation in search for the raw data but Dr. Desai refused to disclose the data – for independent confirmatory review. In fact, he refuses to identify the participating hospitals, or even the countries.
https://www.sciencemag.org/news/2020/06/mysterious-company-s-coronavirus-papers-top-medical-journals-may-be-unraveling?fbclid=IwAR3qJmis6GzYBQGq2hho3aRrSxwYxCBDjRRRQxIjHYIHNoT0RK6ka6XCF-4


From my little experience of having analyzed the thought pattern and modus operandi of controversial researchers, I have noticed that most of them would like to go with something that majority of the people in the field already believe in or has been established. In this case, we know that when Trump touted HCQ as probable drug for COVID19, of course based on a flawed French study, naturally biomedical community criticized that vehemently and as it is a human nature (that applies to scientists as well) most of us started developing a bias against HCQ, some of us just because of our adherence to a political ideology which is different from those those that come from Trump camp which is broadly considered as anti-science.


However, in this Lancet study, I consider these authors more guilty because all of them are very senior investigators (Professor level), no resident/trainee/postdoctoral fellow involved so they can not blame junior members of the team.
To me, looks like investigators of this controversial Lancet study sensed that air against HCQ in biomedical community and exploited it by extrapolating probably a smaller observation/trend regarding HCQ's clinical benefit or lack of it under disguise of now a dubious "clinical database company" run by Sapan Desai so that they can publish it in a reputed journal.


I would still give benefit of doubt to authors of this paper (except Dr Sapan Desai) as they simply may have relied on the "database" provided by Dr Desai who is the chief of the database company - Surgisphere. However, they still are accountable for being careless and not doing due diligence in collaborating with Dr. Desai who has a bad reputation of having been sued multiple times for medical malpractice.

I personally do not think that Lancet editorial board can be blamed for this disaster as they rely on peer review. However, reviewers can partially be blamed for not taking red signal when they came across a "database" that has no reputation in the field. This is where our biases come into play, when we see a study submitted from Harvard (Mandeep Mehra) most of us take it at face value and become less critical than same data coming from a smaller center/third world country.


Another thing that I want to note that this paper is a perfect example of how reputation of modern medicine/biomedical research has a greater risk of being tarnished by too much reliance on "big data". I am quite old fashioned in this regard and have personally suffered in my career because of my hesitations regarding the use of "data analytics", "cloud computing", "machine learning", "artificial intelligence" etc which is dominating the field these days. I can not say for clinical sciences, but in the basic biomedical sciences it is getting harder and harder to publish studies that investigate the role of one gene/one protein/one pathway in a disease context. It has become fashion to use big data sets derived from whole genome sequencing which is totally dependent on computer based data analytics and things can be changed/tailored very easily just by tweaking one parameter to favor one's hypothesis. Even in the case of Lancet study, I do not think same thing was possible if they had to go and collect the clinical data old fashioned way from one hospital to another.


I might sound retrogressive in this regard but I think its reminder for all of us in the field to slow down a bit when it comes to making real progress in biomedical sciences, we need a balance between real life scientific observation in lab/clinic and use of cloud based " big databases".



Fast forward, while I was writing this blog, retraction notice came up for this paper: 

https://marlin-prod.literatumonline.com/…/S0140673620313246…

As most of us expected who have faith in science and goodwill of it, 3 authors of the controversial Lancet paper came forward and realized their mistake and asked for retraction of the their research paper on findings. So now we know that the sole culprit of this fraudulent study is only one person - Dr. Sapan Desai.


In brief, it was an observational study on allegedly a large number of COVID19 patient dataset claimed to be in possession of doctor named Dr Sapan Desai. Dr. Desai apparently does not have much reputation in the field but it is strange that he got hold of three eminent doctors/investigators namely Mandeep R Mehra, Frank Ruschitzka, and Amit N Patel from reputed institutes.


Scientific/medical studies always need collaboration, more often multiple teams from multiple institutes with different specialities collaborate and work together to conduct a biomedical study. Trust is everything in a team work and people work together in good faith. To me it appears that these 3 authors of this study were oblivious of the background of this guy Sapan Desai who is a mastermind of the fraudulent study. Read the statement of these 3 authors:

"We all entered this collaboration to contribute in good faith and at a time of great need during the COVID-19 pandemic."


Like any other field we scientists too have our share of evil. There are bad people who sometimes commit wrongdoing intentionally and sometimes it is committed due to their carelessness. Whatever may be the case, scientific fraternity is very self critical. We are taught to doubt our own data, ourselves, so whenever any wrongdoing is caught, community takes its cognizance and corrective measures are taken. So as soon as this study came up, biomedical fraternity started raising concerns and the journal started it's internal investigation.


So retraction of a research paper is not a new things for those of us who are insider, but this episode of wrongdoing by Sapan Desai is certainly very disappointing especially at a time when the whole world is watching us with hope. I am still very upset and would personally hope I am still very upset and would hope that some kind of criminal prosecution be launched against Dr. Sapan Desai.



Moral of the story 

Do your due diligence background check before falling in love with your collaborator/s.