Will the Uncertainties of COVID Science Resurrect Blogs?

Health news was popular before the pandemic. Now, almost all news is health news.

It’s not only a rapt audience contributing to the deluge of COVID19 news. Two other factors: 1) the availability of preprint servers, digital archives where a scientific paper can be published without formal peer-review and 2) the attention economy.

Attention is currency. Since the business model of both scientific journals, internet-based medical news sites and mainstream media is attention (citations, views), both groups are eager to publish all that is COVID.

The slew of COVID papers are outpacing the normal vetting process. It happened with hydroxychloroquine. In normal times, a study as weak as the one which launched this unproven drug would have never passed muster.

Media has the same problem as medical journals. They struggle to keep up with the public’s need for COVID information. Vetting was far from perfect before COVID, but there used to be more time to assess a study, seek expert opinion and add a bit of caution.

For instance, pre-COVID, journalists would often get an embargoed copy of a study days before it was published. The luxury of time does not exist during the pandemic.

Enter blogs

Blogs were hip 10-15 years ago, but have given way to Instagram, Twitter, and podcasts. Content consumption via the printed word seems quaint.

But these days I find myself increasingly drawn to the raw and candid nature of un-edited blogs. Here is why: A study gets covered in mainstream media. It fits an anointed narrative. Once one media site covers it, others feel like they have to cover it as well. Then Twitter and Facebook amplify it.

Boom, a message becomes accepted, often codified. If slow science pre-COVID was imperfect, the fast science of COVID can be deeply problematic. I’ve already mentioned hydroxychloroquine.

Another example is a study showing that cities that started social distancing earlier and stayed with it longer had better economic outcomes during the 1918 flu pandemic.

That narrative aligned well with the notion that stricter/longer lockdowns were the right choice. Keeping the virus from spreading is good for health and good for the economy.

Here’s the problem: other researchers could not reproduce it. When they tried, they found noise instead of signal.

Where did I read that? No, not in the NYT or WashPo or NPR; they don’t have the time or the incentive to correct a story. I read it on a blog from a distinguished professor of statistics at Columbia.

Professor Andrew Gelman writes;

The larger issue is that there seem to be more outlets for positive claims than negative claims. “We found X” can get you publication in a top journal and major media coverage—in this case, even before publication. “We didn’t find X” . . . that’s a tougher sell.

Another example, one which I may opine on in coming days:

A prominent group of researchers in cardiology published an observational study of patients with COVID who received clot-blocking drugs called anticoagulants. The top-line result was that patients who got the anticoagulants did better.

One of the authors of the study is the editor-in-chief of the journal that published the paper. He’s influential. Mainstream media covered the story. The WashPo used an upbeat headline and weaved a positive narrative.

This, too, fits with another common theme: keep socially distancing because if you get the virus later rather than sooner doctors will have found better treatments. And this may be true, but the study on anticoagulants was far too flawed to make any such conclusion.

Where did I get that? Well, since this is cardiology, I could critically appraise it myself. But say I wasn’t a cardiologist.

I, or you, or anyone, could have read the blog of intensive-care doctor, Josh Farkas, who writes on topics related to ICU medicine. Here is his assessment of the study’s fatal problem, including something technical called immortal time bias.

Numerous sources of bias exist.  Perhaps most notable is immortal time bias – patients who live longer may survive long enough to be diagnosed with DVT/PE and be treated with anticoagulation (whereas other patients may die rapidly, before being able to be prescribed anticoagulation).

You might wonder how one is supposed to know if information from a blog is reliable. This is where thinking comes in; it’s under-rated. You start with the content. Make a judgement.

You can also use meta-data: for instance, a blog with adverts for life-hacks or proprietary supplements might downgrade your confidence. Keep in mind, though, that many medical journals make you watch an ad before looking at a study.

One of my favorite American heroes, Mr Rogers, famously said when you are scared look for the helpers. I wonder… in this scary pandemic will more people start looking for help from the doctors, professors and thinkers taking time to write online. One of my recent favorites is Marginal Revolution.

Medical studies often get discussed on Twitter on the day of publication. Indeed, it is a great place for watching and conversing in realtime, but, as Doctor Bryan Vartabedian writes, in a blog, the ephemeral nature of Twitter makes it a lousy place to park ideas.

Finally, and this is big: science is not supposed to be held up as law; it’s supposed to be corrected. Being wrong in science is normal. Humility is essential. More and more, I am interested in how we communicate that core tenet to the public.

The tension, of course, is that the public can’t handle uncertainty.

I wonder if that is true. Perhaps it is the opposite: is it the faux certainty that bolsters distrust and division?

JMM