This is the story of a young turkey who lived in a farm. All alone, separated from its parents, it was petrified to see a man approaching its cage. To its surprise, the man merely placed food and water in the cage and left. The next day, when the man approached, the turkey shrank with trepidation, but again the man simply left food and water for it. After a week went by, the turkey, being a very smart one for its species, used probability theory and deduced that the chances of anything bad happening when the man approached, was quite low. As time passed, it became increasingly confident of its deduction as it had more data to work with and lost its natural fear of the man. On the 100th day, it jumped with joy when the man approached the cage, in anticipation of being fed. Alas! It was the day before thanksgiving.

So, what does this story have to do with the pandemic, much less the practice of physiotherapy, one might ask? That requires another story; this time, a true one.

In early January, I had booked a week-long family trip for March 6, to Arizona and California.  At the time, COVID-19 was not on my radar or that of most governments around the world. By early March, doubts began to creep up on the wisdom of travel in an environment of fear and uncertainty. On March 5, I had to take a call - to travel as originally planned or cancel the trip entirely. Friends and family were apprehensive and cautioned me against travel. I had lived through SARS and H1N1. How much worse could this be? I decided that I would not abjectly surrender to paranoia and make my decision based on data.

My home province, Alberta, announced its first case of COVID-19 on March 5 and Arizona had about five cases. California, which was what everyone was concerned about, had about 50 confirmed cases for a population of 40 million. I calculated that my absolute risk of coming in contact with an infected person while in California was 1 in 800,000, which I considered to be low enough to risk travel. Of course, the risk would be somewhat higher at airports, but there was no way to obtain any reliable data to generate a risk profile. In any case, we travelled first to Phoenix and then to San Francisco. Within a day or two of our leaving the country, the rate of infections took a turn for the worse.

On March 13, or thereabouts, California shut down schools and ordered non-essential workers to work from home. On March 15, we returned to Alberta to a two-week mandated self-isolation. Two days later, Alberta declared a state of emergency. On March 27, two days before my self-isolation ended, the Alberta Government shut down all allied health services, except for urgent and emergency care. Needless to say, the exponential increase in infections caught me completely by surprise as I was expecting a linear growth based on the trends in February and early March. In other words, I was the proverbial turkey that assumed that past trends would predict future ones. It seems I am in august company, as even a cursory search of the internet will throw up videos of world leaders who reassured the public that nothing bad was going to happen and that this will all go away.

So, what can we learn from this experience?

Expect the unexpected. Once in a while, life-changing events do happen. Sometimes they are personal events and sometimes they are global. When the risk of adverse events is low, we often behave as though the risk is zero. This illusion of zero risk lulls us into a languid complacency, which often comes back to haunt us. 

The moral of my stories is not that we need to live in perpetual fear of a life-changing event, but rather that we should be prepared for one when it does occur. Are you being a turkey when you say to your patient that the chance of an adverse event is one in a million? Do you have a policy to follow if you cause a pneumothorax or, God forbid, a vertebral artery dissection? Do you have a policy in place for dealing with a college complaint or a malpractice suit?

Complex phenomena are non-linear. By now, anyone who has followed the news on COVID-19 would have seen a histogram or two that shows an exponential curve. One cannot predict the exponential increase in infections using models that assume a linear relationship between cause and effect. In complex systems, which a pandemic is, the whole is greater than the sum of the parts, and the output of the system is disproportionate to input. This is not unlike our own experience with chronic pain. The smallest input, such as the act of picking up a pencil from the floor, could land a chronic pain sufferer in the emergency room. Reductive treatments, such as drugs or injections, which follow from our supposition that we can affect the whole by treating the parts, often fail to make a dent. For far too long, we have idealized equilibrium or homoeostasis as the state to strive for.

However, non-linear systems, including the human body, measurably stray from equilibrium in response to daily stressors. There is emerging evidence that it is, in fact, chaos that helps us adapt to the challenges of life. We now know that physiological variables such as heart rate, hormonal levels, brain activity, and blood pressure fluctuate in a chaotic pattern. A detailed discussion of chaos theory in relation to chronic pain is outside the scope of this blog. Suffice it to say that we are in desperate need of non-linear models to explain complex phenomena like chronic pain, which segues to my next point.

“All models are wrong; some are useful.” - George Box

We use models every day, like the ones on the current pandemic, to understand the world around us. Models are like maps; they help us navigate the maze of complex information that we are confronted with on a daily basis. They are not wrong in the sense of being incorrect, but rather in the sense of being incomplete, simply by virtue of being unembellished. Segmental facilitation, trigger points, and Chan Gunn’s radiculopathy are some examples of models that seek to explain roughly the same clinical phenomena, with slight variations, although the treatments that we derive from those models are vastly different from each other. In chronic pain, which happens to be my area of interest, models abound from the purely biological to some combination of bio-psycho-social factors. I often construct models of my own to explain clinical phenomena to a patient or a student. They are all equally valid or equally wrong, depending on what we are trying to explain. They are merely models, not reality. In the interest of avoiding a profound philosophical discussion on the nature of reality, I will simply state that models are useful, in so far as they help us come up with coherent solutions to a patient’s problem. I have often seen practitioners dogmatically adhere to a single model, even when it is an obvious misfit, simply because they are emotionally and professionally invested in it. There is no model that works in every situation. Dump it if does not work.

In the face of uncertainty, trust your intuition. This is probably the most controversial thing that I am going to say on this blog. I am sure there are some who would advocate relying on data, no matter how poor, for it would still be better than relying on one’s hunch and I would have wholeheartedly agreed with them a few weeks ago.

Friends and family had an intuitive anxiety about travel and advised me against it, yet I chose to rely on poor, or at the very least incomplete, data to make my decision. At the time of my travel, testing was not widespread and there was no reliable data on transmission through asymptomatic individuals or the rate of false negatives with testing. Wait, you might argue, reams have been written about systematic bias resulting from trusting one’s intuition. Isn’t it the role of scientists, which we all purport to be, to arrive at optimal decisions through a rational examination of data? While this is true, little is heard about the successful use of intuition and heuristics for problem solving in the face of uncertainty. There is, in fact, an economic theory called ‘bounded rationality’ that questions the conventional assumption of ‘homo economicus’, the economic man who makes optimal decisions after rationally examining all available data. Don’t get me wrong. When we have reliable data, it still makes sense to base our decisions on calculated probabilities. However, this is seldom the case in clinical practice. Data can be conflicting, misleading, biased, or downright wrong. In the face of such uncertainty, wouldn’t it be better to rely on intuition that is subconsciously informed by years of experience and smart heuristics rather than on poor data? Isn’t there a case for sharpening our heuristics rather than ignoring them?

At some point in our lives, we have all been turkeys. We predict the future using the past and are caught by surprise when unexpected events happen. Avalanches, earthquakes, pneumothoraxes, and strokes do happen. Paradoxically, we are expected to use non-linear thinking in our clinical practice when the research that informs us is dominated by linear models. Few phenomena that we encounter in our professional lives are linear, yet we persist in our linear modeling. The current pandemic is a quintessential example of non-linearity in nature and a lesson in humility to those of us who take pride in our ability to predict the future.

Uday Emani, PT, MClSc, DSc, FCAMPT
udaykirane@hotmail.com

Facebook: Acadia Physical Therapy & Health Centre
Twitter: @Acadiaphysio
Website: acadiapt.com