April 28, 2020

"Lies, damned lies, and statistics"

Mark Twain said it well. We are entering the Pandemic phase of “who are you gonna believe?” Claims of cures are being questioned by our State Attorney General, conspiracy theories are flying across the internet, videos of doctors dressed in scrubs claiming it is safe to open up the country based on data they gathered have gone viral. Others claiming the deaths are not that high after all, despite evidence to the contrary. Reading these things can make you crazy. Trying to convince someone these things aren’t true can be a losing argument. Even people working with COVID patients can make the same mistake. We can easily be convinced by data, or what we perceive to be data. This editorial in the New England Journal of Medicine addresses many of the mistakes so many of us are making: we want to help, we want to get out of this uncomfortable place, we see a problem and want to fix it even though we don’t have the right tools for the job yet. At the same time we want to help, we don’t want to cause harm but can inadvertently do so.

In medicine, we talk about “anecdotal evidence” which is basically what we have experienced and observed on a case by case basis, often these are based on hunches, expectations, or other beliefs. When we see someone get better, we attribute it to something we did or observed, but maybe the patient would have gotten better anyway. That is what research is for: to contrast, compare, and analyze for actual effect, good and bad. So many times the “anecdotal evidence” fails to show effect. You can see those results now in data of using hydroxychloroquine or other treatments. We so want it to help, but at the same time we don’t want to cause harm.

I do know that there are trolls and others whose goal is to create chaos and cause discord. Those situations are frustrating and maddening. At the same time, many people have good intentions and want to help a problem, such as the economy. They can be easily influenced and are vulnerable to this manipulated data, which can be dangerous and divisive.

How do we know what is good data and what is not? Big question and not easy to answer. For myself, I read and glean from places I trust, I use the skills I learned to analyze new data, and I look carefully at the motivations of those presenting data outside of the usual research channels. The most respected journals all use “peer reviewed” studies that looks at the validity of the research. I compare and contrast, and I try to stay open to new data as it comes. Be alert to manipulation. If it is too good to be true, it probably isn’t true.

Wash your hands and keep your nose covered.

And finally, my caveat is that this is my experience and my opinions, which are subject to change as more information is available, and not related to the organization I work for. Thanks for reading.