CASE FATALITY RATE AND THE FEAR OF FRACTIONS

Numbers can cause nightmares.

I slept fitfully on the eve of my Whipple procedure because in the morning I would undergo a surgery for which I had been cited a 3% operative mortality rate. That percentage had looked reassuringly small when I signed the consent form in a brightly lit clinic but loomed large during a dark night of the soul. Suddenly an approximate one-in-thirty-three chance of death seemed intimidatingly substantial, wracking me with anxious insomnia.

The polysyllables of pancreaticoduodenectomy made no secret of the Whipple’s complexity, with each step of resection carrying its own risk. My surgeon, remarkably skilled as he was, could not rule out the possibility that my digestion would never normalize, that I could incur delayed gastric emptying, a bile leak, or a fistula. I had been told the incidences of each and every one of these complications, but they were all dwarfed by two more numbers: 100% or 0%, the spartan binary math of whether my surgery would prove fatal or not. 

Recently I have been reminded of the subjectivity of statistical interpretation, whereby a number’s meaning can grow or shrink based on the prism through which it is viewed, as the global health community grapples with the threat posed by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). The starkest endpoint from the outbreak has been the rising tally of lives claimed by coronavirus disease 2019 (Covid-19). As the death toll has mounted, it has also been common for the case fatality rate (CFR) to be reported to the public, and I have seen many struggle with how to process that grim fraction. 

As a medical oncologist I am accustomed to framing danger proportionally. In the calculus of difficult decision-making I try to prove to my patients that the well-known side effects of chemotherapy are worth hazarding, presenting a risk/benefit ratio that I hope will not be unacceptably top-heavy. When administering cytotoxic drugs I carry in my head a repository of percentages – a 37% chance of neutropenia with a certain combination regimen versus only a 13% chance when deploying a single agent, for instance – whereas my patients are understandably less empirical. Their choices often stem from their amygdala, their fear center, more than any other part of their brain. And so, for them, it becomes an accountancy of threats. In the ledger of horrors, their natural revulsion to poison has to be superseded by their terror of an unopposed cancer. Every patient's balance sheet looks different. For some, no chance of therapeutic benefit is too slim to lose its seductiveness, still enticing even when hope for a positive outcome looks razor-thin. In my practice I have been astonished at how many patients are willing to accept near-inevitable toxicity for the vanishingly small possibility that they will be one of a select few ‘exceptional responders’.

Physicians are also not immune to cognitive bias in interpreting data, and as a patient I now perceive how impersonal we can sound in discussing such critical matters. Stalin could have been a research-reliant oncologist with his chilling observation that one death is a tragedy, but a million deaths a statistic. Peering through the lens of intellectualization, cancer doctors find themselves mired in the same emotionally neutered numbers he exploited to render the scope of his atrocities inconceivable. But rather than hiding mass graves beneath Siberian snowdrifts, we bury our dead in Kaplan-Meier curves. These graphs can be found writ large in our journals and projected on massive screens at our conferences, but their name sounds rosier than the sober reality they actually depict. If you look closely enough these lines aren’t smooth at all, the difference between a cartographer’s ink and a true coastline. In reality they’re ragged, a series of plummeting cliffs whose topography only gets planed down at a cold remove. Magnified, restored to corrugation, each descending notch on the graph represents a nameless casualty. When these declivities are understood on an individual basis, the “survival” curves become unearthed as steamrolled cemeteries. 

When I entered the operating room with the bicameral mind of a patient-physician, I knew my likelihood of a good outcome was extremely favorable, and indeed I emerged very fortunate. But I learned from that experience not to discount the enormous psychological burden of confronting a small yet appreciable risk of death. As of this writing, the World Health Organization estimated a CFR from Covid-19 of almost exactly 3%, mirroring the odds I faced when going under the knife. 

Attempts to calm the public that the CFR is “only” in the single digits may be well-intentioned but underestimate both apprehension on an individual level and the collective havoc that will result from multiplying even a modest coefficient of mortality across a massive population. While we do not want to stoke fears to a fever pitch, if we are too dispassionate in our messaging we will appear tone-deaf to a looming catastrophe. 

In the information age it is easy to presume medical literacy and numeracy as we invoke epidemiologic terms like Ro to describe the spread of SARS-CoV-2. But digits in isolation are prone to being misunderstood, especially as the zeitgeist tilts towards contemplating worst-case scenarios like a reprise of the 1918 “Spanish flu”. Words matter, and the numbers we disseminate must be couched in language that contextualizes this unprecedented virus against better-known entities like seasonal influenza while providing pragmatic advice to curb its impact. Those of us who are active on social media can transmit verified updates and signpost reliable sources of news, amplifying well-written articles that can counteract false rumors and help patients to parse a deluge of disturbing data. We must also exercise extreme care in choosing the adjectives and adverbs that modify our reports of CFR, lest we dismiss and disrespect the potentially calamitous loss of life. Especially at a time when so many will likely need medical attention, it is crucial that our communication imbues trust in our care and in our compassion as healthcare professionals.

SARS-CoV-2 has made it more important than ever that physicians convey estimates of risk to a troubled populace with accuracy and sensitivity. As stewards of the public health we are now walking a tightrope between preparedness and panic when we broadcast more widely about this emergency of international concern. The famous quote from Mark Twain places mathematical sleight-of-hand deliberately last in the unholy trinity of "lies, damned lies, and statistics," and we cannot hide large sums behind small percentages. While we track the labile denominator – and make every effort to reduce the numerator – of a fearsome fraction, we must be honest with our patients and ourselves that even a low CFR can presage a massive tragedy as Covid-19 menaces us all.