A month’s worth of blood sugar data

I tracked my glucose levels for twenty-eight days using a continuous glucose monitor (CGM). Instead of traditional twice-a-day finger pricks, CGMs check glucose at regular intervals via a small probe embedded under the skin. I wore mine on the arm for two two-week periods, totaling 2,629 measurements. A startup called NutriSense helped me obtain the CGM and parse the data (I paid full price for the program). NutriSense taught me not only about GCMs and glucose tracking, but also about myself.

Why did I do this?

CGMs are typically prescribed for diabetics and paid for by health insurers. I do not suffer from diabetes. I’m fortunate to be in good health, able to prioritize physical activity and watch my diet.

I signed up for NutriSense mostly out of curiosity. I also like the company’s founders. One could also argue I’m a metrics geek; I’ve previously tracked sleep, steps, strength training, and diet. Another biometric quest seemed a good reason to stick a needle in my arm!

About the device.

It’s amazing. While the CGM has its downsides, I am floored this technology exists and works. The device does its job without charging, maintenance, or any of the usual user hassles of gadgets.

No ambiguity – the CGM felt like a medical device from the moment I took it out of the packaging. There’s no risk of mistaking this for a fun consumer-y product like the Fitbit. That said, putting it on wasn’t actually that big of a deal. I envisioned a needle stuck in my arm for weeks. But the needle isn’t much of a needle, really. It’s more like a very thin, flexible tube. I think there’s a needle in the applicator device, but I never saw it.

It didn’t hurt. The idea of needles turns many away so let me be clear: I felt zero application pain. There was only a little arm soreness afterward for the first 1-2 days. Most of the time I wasn’t even aware of the device on my arm. The only times I felt the device was when (usually around meal time) I became obsessive about reading its glucose measurements.

The calibration isn’t perfect. When I switched sensors after two weeks, the replacement read 10-15 units higher on a regular basis. NutriSense’s tech team looked into it and concluded slight miscalibration is a drawback inherent to CGMs, fixed only by actually drawing blood to create baseline comparisons. I definitely am not interested in drawing blood, so have to accept this.

Continuous is not really continuous. The device seems to sample once every 15 minutes. This is still very good fidelity for any non-medical use case.

Short data storage window. The CGM’s 8-hour device memory, coupled with my usual 8 hours of sleep, means I often missed data.

What I learned about my glucose.

I first learned how quickly my blood sugar level jumps around! A little research informed me this is quite common. Our bodies are constantly re-balancing, responding to changes in available resources (food), state (sleep, wake, exercise) and environment (stress).

The other “whoa” lesson was how different foods uniquely affect my blood sugar. Duh, right? More interesting was how different food combinations mitigated glucose spikes. Adding nuts or oil to my daily overdose of chocolate mellowed my spikes. The combo of apple + peanut butter was pretty good too. And despite a hopeful hypothesis, I found adding fudge ice cream to a diet did little to battle blood glucose.

My usual big salad proved a champ for keeping blood sugar in check. Despite the gigantic portion size, my body produced almost no reaction.

We don’t eat gluten (wheat flour) in the house because of allergy. Instead, we use a variety of alternative flours for pasta – think chickpeas, lentil, quinoa. The CGM showed alt-flour pasta has little effect on my glucose. Even when I hit it hard, my spike wasn’t pronounced.

I did learn, however, that white rice is my glucose-spiking nemesis. I first saw this when eating Indian food packed with basmati. After that first spike, I noticed the clear and immediate impact every time I ate rice. Purple sweet potatoes also seemed to have a strong effect on my glucose, while regular sweet potato did not.

One fun surprise was learning red wine didn’t spike my glucose. I expected a pretty strong reaction, but received none. Oat milk, on the other hand, as an alternative to half & half, had a larger-than-expected impact on my blood sugar. It sure is delicious, though!

What I’d like to see

It would be neat to pair NutriSense with a specified menu, a more thorough process for testing foods, or both. While I learned a few helpful nuggets about the link between my diet and my blood sugar, I definitely didn’t get all I could have out of the month.

Better normalization of data from one device to another is also needed. I lost too much confidence in the data once the second CGM showed a 15% higher base level glucose. Even though I knew this normalization issue was more psychological than physical, my lost confidence showed in my dedication to logging meals.

I’m curious about the interplay of stress and blood glucose. How one might accurately measure stress levels to compare against CGM data? Much of the variation in my blood glucose came from diet, yes, but I also know stress plays a role.

Longer onboard sensor storage would improve user experience. Having to frequently open the app and scan created a weird neurosis over my blood sugar. I obsessed about my levels more than I wanted, in part owed to the necessity of constantly opening the app.

An easy way to compare my body’s differing reactions to the same food stimulus would also be helpful. Nowadays, I drink coffee with oat milk pretty frequently. It’d be cool to see multiple, regular snapshots of my glucose after drinking that oat milk coffee.

I still want to understand how to determine my optimal blood sugar range. 70 to 140 is often tossed around as a generic target range, but that doesn’t really hit home for me. The same urge that drew me to CGMs also demands to know the source of any metric. I’d like a more thorough basis for knowing whether my diet is helping or hurting my body. I want the data! NutriSense’s CGM is a great starting line; I’m curious to see what comes next.