That’s Probably Not a Zebra
September 21, 2020 | Letters
***This letter is not intended to provide medical advice or interpretation of epidemiological data for seasonal influenza (flu) or COVID-19. For information on these viral diseases, please see https://www.cdc.gov/flu/index.htm or https://www.cdc.gov/coronavirus/2019-ncov/index.html.
Simply stated, “When you hear hoof beats, look for horses, not zebras” means that when someone turns up ill, it is much more likely to be a common disease than an exceedingly rare disease only seen in med school textbooks. Now, in the middle of a globally devastating pandemic is an odd time to be pondering this notion of common versus novel disease risk. We are over 6 months into the COVID-19 pandemic that has changed every one of our lives, and the positive percentage of those tested remains a critical metric for managing and detecting hot spots.
In many places, the majority of tests was/is/are conducted on people showing symptoms of illness. This is not a random sample of people, thus results likely cannot be projected over the whole (symptom-free) population. If we set aside the obviously relevant but presently distracting debates about testing availability in the U.S., this means that when 10% of those with symptoms are testing positive, then 90% of those symptomatic people tested have symptoms caused by something else. At this point, we’ve all had the ‘is this COVID, allergies or a cold scare’, and we’ve been reminded repeatedly that the pandemic illness toll is in addition to and not in replacement of other illnesses and ailments (aka “Excess Deaths Associated with COVID-19”). People are still falling ill with any number of viruses, bacterial infections and an assortment of other ailments. Having said that, Australia had a COVID-19 silver lining. COVID-19 precautions such as social distancing, mask wearing and hand washing resulted in lower numbers of flu cases. Unfortunately, it is debatable whether the U.S. will be so lucky, so get your flu shot (see CDC flu shot 2020 recommendations here).
Percent positive rates rising or trending upward is taken as a very serious warning sign by public health officials. But probabilities and percentages are wrought with interpretation challenges. We as humans have a very hard time internalizing risk and/or probabilities, and interpretation is fraught with a number of fallacies. For example, we link together probabilities that aren’t related, or we fail to recognize linked outcomes when they do exist.
Now, in consumer-related research as it pertains to agri-food businesses…
Why have consumers selected food item X over food item Y? Well, it could be that consumers do not like the processes employed on food item X and they are skeptical about the safety of that item. Perhaps they heard untrue rumors about issues in a production facility. It could be a lot of things. Or it could be that food item Y simply tastes good, and that’s all the thought most people put into it.
In a business-to-business example, when you lose a customer to a competitor, it could indeed be an elaborate scheme by the other company to target your customers with a unique product-specific discount with personalized, targeted marketing that took into account their bundled needs with another provider of a complimentary product. Or it could be that your customer saw a promotion on a billboard. Or it could be – very simply – that your customer has asked you for something that you didn’t deliver, and thus, they left.
What is the point here?
Probabilistically speaking, in any given situation, my money is on the horse. But roll these dice enough times and you’ll find zebras. 2020 has delivered a herd of novel zebras to comingle with already known ones. We concoct all sorts of ideas to help ourselves process probabilities, but most of them are stories we tell ourselves rather than mathematical/statistical truths. Do the odds of finding (avoiding) the proverbial zebra go up because you recently saw (didn’t see) one? Could it be that zebra sightings are related or driven by underlying factors (including biological processes) that are changing the probabilities in related ways? Sure. Is this likely? It depends on the situation.
All we are asking is that you maintain the simple explanation in your option set when answering the question, “What is going on here?”.
Disclaimer: COVID-19 may have been a zebra, but a novel and potentially deadly new kind that kicks you in potentially novel ways. (If you’ve seen National Geographic, you know zebras kick lions and wildebeests, so being rare doesn’t make something not dangerous.) So, if you think you have allergies, you might, but it’s best to seek testing and professional medical advice.