I have an unusually large amount of information about my health. I order my own quarterly bloodwork through WellnessFX; I have an annual elective MRI and cardiac ultrasound (first through Health Nucleus, now RadiologyAssist); I have an elective Dexcom G6 CGM; I regularly use a home blood pressure cuff and my Kardia 6L; and more. I am careful to avoid being a burden on the system — all of this is paid for out of pocket, without using insurance — but even then it comes to just a couple thousand dollars a year, which seems reasonable based on how I value my health. It is certainly far less than many pay to have a chronic disease, even covered by insurance.
When I started this several years ago I felt like there was a misconception in the medical community that this additional data, frequently derided as wasteful “overtesting,” was in fact actionable and useful. With the benefit of hindsight, I’m pretty sure now that I was right and this is true. Not being a physician myself, my doctor friends roundly made fun of me when I first showed up with my Dexcom: when I lost 35 pounds by focusing on minimizing the area under the curve, after trying and failing to make any progress the prior six months, they started taking it more seriously. I just got a text from one of them last week with a photo his own elective CGM.
One thing that’s stood out to me is that, unlike many other things, almost all of the data here is useful. It’s not a situation where most of the value comes from a small percent of the effort. Will all this actually improve my healthspan outcome? Of course I don’t know, and this is unknowable; but I do not feel like the effort is wasted, and I can see the impact in traditional metrics doctors know and love.
But with all that said, I do feel like I am reaching the edge of Consensus Medical Information™. While the bloodwork is prima facie informative, no one really follows all these things on a regular basis in large numbers of healthy 30 year olds. For a while there was a clear downward trend in my white blood cell count, with all of the obvious things ruled out, all still within the “green” range, over the timescale of more than a year. Is this normal? Is this concerning? No one knows. This is where we start leaving the territory of Medicine and start engaging in Research. And this is for common tests that everyone already knows.
I ordered a Geneva ION. Why? It seemed like it covered a lot of potentially interesting, if admittedly random, things, and there were a couple analytes in there I was specifically interested in but otherwise wouldn’t have been easily able to assay. The nicely-formatted report I got back showed each result plotted on a reference range with color-coded results. But I was immediately distracted with the obvious question: where did these reference ranges come from? Is there some lab out there that’s actually followed a bunch of people in my demographic to establish that my phosphoethanolamine result of 3.6 μmol/L is a “green” result?
I suspect the answer is that these reference ranges are actually just normalized over the past results this lab has seen; indeed, on the report they call it a “95% reference range,” which makes me wonder. Needless to say, I quickly realized that the informational value of these results — outside of specific analytes I had some other mechanistic understanding of — was low.
But! What if we could figure out how to properly reference arbitrary biomarkers without a substantial existing literature? That would be super powerful, but metabolic variation is immense, not to say anything of other differences.
This is when I realized that there are three people out there who share approximately 50% of my genome, and a bunch more who share either around 25% or 12%. Not only that, but some of them have decades more medical history than I do! This brought me to an idea that I call N-of-5 medicine.
People talk about N-of-1 studies where all of the focus is on one special case, but by bringing the family into it, I have high conviction now that there is a huge opportunity to completely transform how primary care is understood and delivered. These genetic relatives can provide meaningful references for essentially any biomarker I can assay; further, their experiences can help predict things like drug response, impact of lifestyle changes, and a lot more. Doctors have always asked for a cursory family history, but in retrospect, stopping there seems remarkably short-sighted.
These statements still feel, somehow, contrarian. I am sure that posting this to the internet will provoke short-term mockery by primary care professionals who are simply speaking a different language. One of the challenges here is that the results cannot easily speak for themselves, since the feedback cycles are so long. But this is an area I am very interested in and am actively pursuing — I am investor in and advisor to Biograph, launching soon — and would love to meet scientists and clinicians who believe in a future where we make far better use of the possibly available information for prospective precision medicine.