The Importance of Anecdata

My wife will tell you that one of my more annoying traits is my unwavering desire to always be a contrarian. In my youth, this was probably more quirky than annoying: I opted for the Dell DJ instead of the iPod, watched Seinfeld when my peers were watching Nickelodeon, and always took the unpopular position in the in-school debates (think debating as George W. Bush versus the rest of the class who chose John Kerry in the 8th grade presidential debate).

Nowadays, in the world we live in, ever so divisive, it is probably less quirky and more annoying, but if it’s mainstream, I’m skeptical, and if it’s fringe, I’m intrigued.

Take for instance, the Carnivore Diet. I tell people at dinner parties that I’m “Carnivore curious.” That is, I am intrigued about trying a diet that only includes meat, seafood, and eggs. And the reason for my intrigue isn’t grounded in science, per se, but rather what I would call an abundance of anecdata.

You see, all over the internet these days, I see person after person detailing how, after they went on a Carnivore Diet, they not only got in the best physical shape of their lives, but they have never been more mentally “locked in,” and many physical ailments plaguing them have disappeared. And I’m not talking the tweaked knee or bummed shoulder, but more serious inflammatory and auto-immune disorders. Medically, it makes some sense: it is an elimination diet, devoid of processed foods, gluten, and other pro-inflammatory substances that may be intertwined with the rise in conditions like IBD, Crohn’s, Colitis, RA, and even MS.

A recent doctor friend of mine and I got into it recently, as I suggested that some of their patients with difficulties losing weight no matter what diet they follow should consider carnivore. I was mocked, hailed as some charlatan who was no better than the snake oil salesmen of late-night TV. Why? Because “I had no evidence.” And, when I pointed to the near infinitum case studies I’ve seen on social media, these were brushed aside as purely anecdata. (Admittedly, I realize on the scale from MAD Magazine to the New England Journal of Medicine, Twitter anecdotes are like SI for Kids in terms of impact factor).

Now, to me, my stance is, the plural of anecdotes is observational data, and it is enough to at least be intrigued. Hence my dinner party stance (which I will say is certainly more rooted in trying to get a rise out of my doctor friends than it is actual interest in trying the Carnivore Diet, which I don’t think I would be able to properly adhere to). But I think my point more broadly, outside of fad diets, is that anecdotes, and anecdata, should be considered from an evidentiary perspective with more seriousness, at least if we are ever going to live in a world where we actually attempt to practice precision medicine.

That is what makes an n of 1 study, or any time a drug is used off-label or in a novel combination, or even Compassionate Use (Single Patient INDs, intermediate Expanded Access protocols) so fascinating.

We can learn a lot from these anecdotes that may better elucidate when to use the right drug, at the right time, for the right patient.

At xCures, we have run a number of Single Patient INDs and Intermediate Expanded Access programs that have yielded many unique results. Take for instance, this case study, in which we reported an individual diagnosed with stage III BRAF D594G-mutant melanoma who experienced an extraordinary response to the ERK1/2 inhibitor ulixertinib as fourth-line therapy. Without this protocol, it is unlikely this individual would have been able to access ulixertinib. And now, in the future, other individual cases who may present similarly now have some evidence about how to elicit an exceptional response.

This isn’t a unique case, either. At xCures, our dataset is full of extraordinary anecdata, with many outliers and exceptional responders to off-label or unique combination therapies.

The key for researchers is to identify this anecdata, the specific hallmarks of the case that could be applied more broadly to patients that may also benefit, and most critically, disseminate that information to physicians who see patients. Because too often, the incredible anecdata does not make its way to the treating physicians. And if they do not know about the exceptional responders and the unique presentation of the cases, they won’t know how to apply that information to the next patient who comes in who may benefit.

I do not think anyone has truly figured out the best way yet to apply anecdata for the deployment of precision medicine. But, in the developing landscape of AI and big data, novel Clinical Decision Support Systems may one day ingest large datasets full of anecdata, and then recognize certain patients who fit patterns consistent with those case studies, and recommend to the physician a contrarian treatment recommendation which might just work.??