A Harvard study or a Huh study?

A Harvard study (actually more Boston Univ authors than Harvard) preprint is receiving a lot of attention in the media, widely reported in the news, even though it has not been peer reviewed (https://dash.harvard.edu/bitstream/handle/1/42669767/Satellite_Images_Baidu_COVID19_manuscript_DASH.pdf?sequence=3&isAllowed=y).  In fact, it even formed the basis of a question posed to the Chinese Foreign Minister spokeswoman on Jun 9, 2020.  Government Q&A around the world is rarely of scientific nature or value, although what the spokeswoman replied was quite prescient: I don’t know what this kind of parking lot data can tell you anything definite.  

To be clear: while the undertone of the Western media is that Wuhan had an early outbreak but, again, China hid it from the world, the Harvard study said nothing of this sort.  The authors explained a lot of shortcomings with their data and analyses and pointed out it was only a correlation without controls.  Still, just on the scientific merit alone, it has a lot of problems.  It follows a long list and trend of questionable studies, preprints, and papers on the subject of COVID-19 that normally would not even have preformed or published.  Examples: a Jan 2020 Chinese paper saying snakes might be a host of COVID-19, a withdrawn Indian preprint suggesting COVID-19 was man-made, and a German paper about asymptomatic transmission (https://www.nejm.org/doi/full/10.1056/NEJMc2001468).  The German paper came on the heels of Chinese news reports of asymptomatic transmission and might be the first scientific publication, although it contained an embarrassing error.  It reported an asymptomatic Chinese traveler infected her German coworker(s).  It turned out she had already had flu-like symptom in her German hotel, so she was not asymptomatic.  She didn’t know she had COVID-19 though and only self-reported a few days later.  Regardless, these early German cases were well contained and didn’t lead to outbreaks in late Feb. 

But what exactly is the Harvard study?  Well, they basically looked at satellite data of the traffic and parking lots around several Wuhan hospitals and Baidu search phrases and suggested there was an increased activity as early as August 2019.  Linking that to COVID-19 is truly a huge leap that the authors didn’t even say it with any confidence.  But besides the hole in logic, there are myriad other problems.  The foremost is controls: how about other cities around the world, and in 2018 or early?  How often do these things happen?  This is so lacking that how can anyone draw any conclusion?  Just because Wuhan identified COVID-19 first?  If other places had the same pattern, does it mean they had COVID-19 but failed to identify it?  The Baidu search for the word “diarrhea” is likewise highly questionable: how do you know only Wuhan searched it but not other places?  That the search of “diarrhea” increased in August, hence the implication that COVID-19 started in August 2019, was eagerly devoured by the Western media but simply way over-interpreted and likely meaningless.   Wuhan is among the hottest cities in China in the summer, and poisoning due to spoiled food and fruit consumption is common, even in the days of frigs.  There could be a lag between hot weather and the search for “diarrhea”, perfectly explaining why the latter increased in August and later.

The most fatal problem with the Harvard study, however, as if it needs more, is that their data are so obtuse.  If there are 200 more cars in the parking lot, does it mean 200 COVID-19 cases?  Or 1, 10?  If it is 1, how are you sure it is not 0?  If it is 10 or 200, knowing what we know about COVID-19 now, why wasn’t the whole Wuhan infected by Dec 2019?  

To sum, from their data, they can’t conclude COVID-19 started in August, which they did not explicitly anyway.  But if anyone conclude COVID-19 started in August, he can’t explain what happened 4-5 months later.   

Of course all this is not say that there wasn’t a COVID-19 patient lurking somewhere, maybe in August or even earlier.  But there are just too many possibilities.  A close version of the virus might have circulated silently somewhere, Wuhan or not, China or not, for years.  Then it just mutated to cause COVID-19 in one person recently, and this person transmitted to others, eventually leading to the outbreak in Wuhan.  Or, the virus jumped from an animal to humans months or years ago, but those people are in the mountains or countryside and only recently coming out and leading to Wuhan outbreak.   Or the “first” human carriers were immune and asymptomatic (Class 4 in the 6/9/2020 blog) and their community lived with COVID-19 for months or years, then a susceptible outsider made a contact with them and got infected, which then through who-know-how-long-the-chain-of-transmission, caused the Wuhan outbreak.   No actual data for these scenarios yet, but they are consistent with the origins or theories about other infectious diseases.   

          

The Harvard study is another example of the COVID-19 bandwagon research with low quality and weak or unsupported conclusions.  Worthy of scientific critique but nothing more.    

Note: Wow, the beating is coming fast and furious before one can say the words.  While my critique is about the general problems with the Harvard study, this one strikes directly at its data (https://news.sina.com.cn/w/2020-06-10/doc-iircuyvi7796205.shtml).  In essence, none of their data mean anything any more.  First, the increase in parked cars can be explained by when during the day the pictures were taken.  Second, there were constructions at the hospital(s) resulting in changed parking space.  Third, COVID-19 patients are unlikely going to some of the hospitals.  Fourth, if one checks baidu.com search in 2017 and 2018, he will find the same pattern/increase in those years.  By the logic of the Harvard study, COVID-19 started in 2017 or 2018?  If so, the whole Earth is infected already!  

There is nothing special about this study.  Similar work about other diseases has been done.  The only reason it gets any media attention is because it fits the smear China mentality.  It didn't pass the smell test at first sight.  Now it won't pass any test.  This is what you get when you have dubious logic, shady data (pun intended), and no controls.  At this point, there is only one way this Harvard study can go, i.e., the way of the above Indian preprint: retraction.  But don't worry, WCEV will find another Harvard study later.