Actionable Data (Premium)

The original Microsoft Band was a unique and powerful wearable in its day, and it was bristling with sensors at a time when fitness trackers were basic and Apple Watch didn’t even exist yet. But one of its key differentiators wasn’t health or fitness related: it could also integrate with Outlook, and it could provide notifications for things like email, appointments, and other calendar events, plus Facebook and other social networking notifications, text messages, and phone calls.

I found that functionality annoying in use, but I suspect many Band early adopters loved it, and we know now that it was a peek at the future of more pervasive wearables in which we now live. But it was the promise of Microsoft Band that got me hooked: as Microsoft explained to me, the Band’s knowledge of your health and fitness data and your work schedule would lead to improvements in which it could use machine learning to present what it called “actionable” information to its users that would be based on both sets of data.

I found one example I was given to be particularly intriguing. Perhaps a future Band version would notice over time that your heart rate always went up in the half hour before your weekly Monday meeting. And so it would recommend activities like yoga or meditation that might help. And then it could monitor the situation over time to make sure that what they did helped.

This dream was never realized, of course. After an improved Band 2 release a year later in 2015, Microsoft canceled a third version and gave up on the product and the Microsoft Health service that backed it. As I noted in a later post-mortem, both Band versions had their share of missing features and issues, but it is particularly galling that the device never offered proactive functionality like prompting you to stand up or move around, an obvious first step towards that actionable information promise. Microsoft’s promises for Band, alas, were nothing more than science fiction.

Flash forward 8 years and a lot has changed. After a disastrous launch as a luxury accessory for well-off Apple fans, the Apple Watch was reimagined as a health and fitness wearable and now dominates this market. After renaming and redesigning Android Wear several times, Google finally just purchased Fitbit, the market leader in fitness trackers, and then belatedly shipped its first in-house Apple Watch competitor in late 2022. And these companies, plus Samsung, Garmin, and many others long ago caught up to and surpassed the number of useful data-gathering sensors that Microsoft had first pioneered with the Band. The health and fitness industry was worth over $30 billion by the end of 2022, and the smart health and fitness part of that accounted for over $13 billion of the total.

I have been deeply involved with this market and the various products and services that define it probably from the beginning. And I’ve spent innumerable hours reading books and watching documentaries in an effort to understand why I’ve been overweight my entire adult life, and what I can do about it. I’ve learned a lot during this time. I’ve had some successes and some setbacks. And the trackers and other devices I’ve tried have logged an almost infinite amount of data, which I hoped would provide some insight, some answer to my problems. But nothing has worked.

Well, one thing did work, for a time. In December 2016, after years of research, much of which I credit to journalist Gary Taubes, I switched to a ketogenic (keto) diet, which is essentially a low-carb, high-fat diet, and after several months, I had lost about 35 pounds before plateauing as one does on any diet. I wrote about my efforts at that time, though I finally became scared of giving what is essentially health and fitness advice to others because I am not qualified to do so and because of the obvious dangers should anyone follow my advice.

But if I can contribute one essential truth related to the fitness and health field, and I think I can, it’s this: everyone is different and what works for one person will not work for everyone else, and so the future of healthcare and weight loss is some science-based personalized program that will literally vary from person to person. My wife and I can eat exactly the same foods and she will remain thin and I will gain weight. We are made up differently, our bodies work differently, and what works for her does not work for me.

But Taubes was onto something. He shot to fame with a 2002 article in The New York Times called What if It’s All Been a Big Fat Lie? that I distinctly remember reading. And then followed that up with a series of books like Good Calories, Bad Calories (2007) and Why We Get Fat: And What to Do About It (2010) that I devoured. These books greatly expanded on the evidence behind his claims for the health-destructive nature of sugar and carbs and provided me with the foundational knowledge that eventually led to my keto experience.

There was just one problem: while Taubes had identified the problem, he was light on the solution. Each of his books took several years to complete because they are science-based and require extensive research. And I really wanted an action plan: it wasn’t enough to know the problem, I needed to know what to do about it.

(Oddly enough, I’ve thought this way for as long as I can remember. When I was a child—perhaps 10 or 12 years old—I recall my mother, who had struggled with depression and emotional issues for years by that point, bursting into my room with a revelation. “I’ve figured it out!” she cried. “I’m the adult child of an alcoholic!” Unfazed, I looked at her and replied, “That’s great. Now what are you going to do about it?” This even more important step was something she had apparently not yet considered.)

So I looked elsewhere, and I eventually stumbled onto keto, which has actually been around for in some form since Ancient Greece but experienced a modern revival in the late 2010s. Some probably still think of keto as a rehash of the Atkin’s diet from the 1970s, which is considered a fad diet, while others likely believe that keto, too, is just a fad. But keto aligns with the evidence publicized by Taubes. I was sold.

But in looking up some of the topics for this post, I was both surprised and delighted to learn that Taubes has, in more recent years, taken that next step that I had always wanted him to take. Following years of research and a newspaper article-like short-form publication, he finally published The Case for Keto: Rethinking Weight Control and the Science and Practice of Low-Carb/High-Fat Eating in late 2020. I don’t know how I missed this book, and I just purchased it as soon as I discovered it, so I’ve not yet read it. But the book appears to be the prescriptive effort I’ve long wanted, by following up the why with the how. Fingers crossed.

Anyway, it’s been six years since my keto experience. Since then, the COVID-19 pandemic happened, triggering some unhealthy food and drink decisions, and, as bad on my part, a heightened aversion to visiting the doctor. But this year, finally, I resolved to turn things around, and a key part of this was to get caught up on doctor’s appointments and health tests. I wrote about this in a Premium newsletter editorial a few weeks ago, noting that I wanted to be proactive about my health and, among other things, get a continuous blood glucose monitor because my blood sugar has been borderline high for at least 20 years, as I was curious to use this advance to see how different foods impacted me. The goal, of course, is to isolate those foods to which my body reacts poorly—by spiking my blood sugar—and be healthier, and, hopefully, lose weight.

But if you read that editorial, you know that I ran into an issue that everyone in my country will run into again and again: our healthcare system is broken, and it’s centered not on preventative care, which would save all of us money in the long run, but rather on treating problems after they occur. And in most cases, that treatment takes the form of medicine that treats the symptoms rather than curing the problem. For example, a diabetic can take insulin to regulate their blood sugar, but this won’t cure the diabetes, just ensure that you spend the rest of your life taking drugs and managing it.

I don’t have diabetes (yet?), but I’ve long straddled the line between normal blood glucose (which is a fasting blood sugar level of under 100 mg/dL) and impaired glucose or pre-diabetes (100 to 125 mg/dL). If you surpass 125 mg/dL, you’re considered diabetic. And so I want to do what I can do to stabilize or lower my blood glucose level and stave off diabetes.

Unfortunately, my doctor also has to deal with our broken healthcare system, and she explained to me that insurance would not pay for a continuous glucose monitor. I could pay for it myself, at great expense, or I could start off by getting an over-the-counter glucose test system, which involves a lot of tiny needles and test strips, and continually pricking yourself in the finger throughout the day. It’s not an ideal system—besides being manual slices of time and not a continuous rating, it caused my fingertips to get numb as I moved from one to the other, pricking each multiple times—but I kept at it for two weeks.

I learned a few small things and verified some assumptions, but it wasn’t really what I was looking for. And so, prompted by a helpful comment to that editorial, I started investigating what continuous glucose monitor options were available online after discussing the cost with my wife. And while this is not an endorsement per se, I ended up with Veri, which looks like one of several good choices.

This has been a revelation.

For those unfamiliar, a continuous glucose monitor is a small sensor disc that you apply to and keep on your body, typically on the back, beefier part of the upper arm. It’s waterproof and will work for about two weeks, and you can apply stick-on covers over the sensor to prevent it from getting caught on clothing or whatever, something I quickly realized I’d want. When you apply it, a tiny needle inserts a little filament into the arm—it’s literally painless—and then you transfer its stored data from the sensor to your smartphone and thus to the (Veri, in my case) service. You do so by hovering the phone over the sensor like an NFC wand.

As implied by its name, the sensor continuously collects your glucose data, a marked improvement over the occasional measurements I was getting when I was pricking my fingers throughout the day, providing a more complete picture of the situation. You also use the (Veri) app to track what you eat, and it connects to other related services, and I connected it to Fitbit, so it could collect my sleep and fitness data too. And then all of that is placed in a live chart that includes my glucose level over time. Each meal is scored based on how it impacts the glucose level, and there are interesting daily and weekly scores and summaries.

But that’s not the revelation. That stuff is all welcome and very useful, but it was also expected. No, the thing that has blown me away about the continuous glucose monitor is that I’m starting to get a new and unexpected understanding of my problem. What was hinted at by the manual finger pricks has emerged as a clear pattern with the continuous glucose monitor.

And it’s this: my overall average glucose level isn’t as bad as I thought it was. Yes, it’s still near 100 mg/dL, but it’s often below that, not above it. (Excepting food-triggered spikes, of course. I have been experimenting with different types of food, and the carb-heavy stuff, especially, sends my glucose level into the stratosphere.) But there’s one major exception: after hitting the expected low glucose levels while I sleep—I can get down into the 60s—my glucose spikes, consistently, every single day at around 5:30 am before I get up. And so I get up with an elevated glucose level.

Here’s why that’s interesting: when you get your glucose levels measured at the doctor, they always ask you to come in first thing in the morning in a fasting state where you’re allowed coffee or tea but no milk or other drinks or foods. But my glucose spikes at this time, and so my reading has always been high. It goes down throughout the day, assuming I don’t sabotage myself with carbs. And as I write this at 2:30 pm today, and having eaten a lunch of eggs and bacon after no breakfast, my glucose level is 83. Perfectly normal.

Of course, it’s not always normal, and that is still very much a problem. And so I will use the continuous glucose monitor to track my levels over time, observing how different foods trigger (or don’t trigger) spikes, and I will continue to experiment until I find myself in a place where most of what I eat is healthy from a glucose perspective. I have little doubt that this will essentially be a keto diet, but I am aiming for what I think of as “healthy keto,” which is more varied and adds fiber. I feel like this will work for me.

But I need to figure out how to prevent that early morning spike. And to do that, I need to first understand what might be causing it. As it turns out, an early morning glucose spike is quite common, but there are several reasons why it happens, and it could even be considered normal in some cases. But I have my eye on one cause in particular (with the caveat that I’m not done investigating): poor-quality sleep can trigger a glucose spike just before you wake up because you would otherwise not have enough energy to wake up and get on with the day. As is the case with plateauing when trying to lose weight, your body is doing something perfectly reasonable to help, but it has side effects that are harmful.

This potential cause set off alarm bells in my brain because I have sleep apnea, and even when I was young and skinny, I had a horrific snoring problem. I don’t know exactly when I developed sleep apnea, but I was diagnosed in 2012 and received a CPAP (continuous positive airway pressure) at that time. And I use it to sleep every night. I can sleep without it, but it’s not great.

What I’m wondering is whether the worsening quality of my sleep coincided with my rising early morning glucose levels. Is that the cause?

I know what you’re thinking: “That’s great. Now what are you going to do about it?” And fair enough. But in this case, I may need to identify the cause to (hopefully) correct the problem. And it is perhaps coincidental but interesting (to me) regardless that while the CPAP works for me on some level, it’s never really solved the problem. That is, I rarely wake up feeling refreshed and alert. Fortunately, I also discussed this with my doctor when I asked about the glucose monitor, and she set me up with a new sleep study appointment to see what’s up there. Unfortunately, that won’t happen until early September (again, our healthcare system sucks), and so I will spend the intervening time trying to chip away at this on my own, using the data-based information provided by the glucose monitor as a guide.

And that’s where the term “actionable” reenters the story. I mentioned my Microsoft Band experience upfront because I always remember the exciting possibilities that were never realized with that device, and I’ve spent the intervening years looking for some other health/fitness solution that might finally fulfill that promise. There have been minor advances that skim close to this area, like the Apple Watch that tells you to stand up each hour or the Fitbit that notifies you to move more, but that’s more proactive than actionable. And I do like some of the gamification advances, like Apple’s system of “closing the rings.”

Is any of that truly actionable? No. But with the continuous glucose monitor, I feel like I’m seeing the first real push in this direction. For example, in one experiment, I ate a lunch of eggs, bacon, and two slices of what I call “keto bread” (toasted), which is just 6 grams of net carbs per slice. About two hours after each meal, Veri will provide me with a score for that meal, on a scale from 1 to 10, that’s based on how it impacted your glucose levels over that time. This meal delivered a lower-than-expected score of 6. But when I tried it without the bread, I got a 10, with my glucose levels actually falling after the meal. So even that keto bread had a negative impact, triggering a glucose spike.

The continuous glucose monitor has also confirmed some of my previous educated guesses, and I’m happy about that. For example, even when I was doing keto, I would eat corn on the cob in the summer with the theory that corn mostly passes through your body undigested, so even though it’s a carb, it’s probably fine. And it is, at least for me: I’ve had corn with various types of meat for dinner several times with no spikes.

I likewise figured that the vinegar in sushi rice would help offset some of the glycemic hit of the white rice, which is high in carbs. But the jury is still out on this one: we eat the same basic sushi each Wednesday, but for the first experiment, I stacked the deck by eating mostly riceless sashimi and received a high score (8) as expected. Then, last week, I went back to our normal meal of rice-laden rolls but also had fried pork dumplings, and the score was just 2, my lowest yet. I will try that meal again without the dumplings this week.

Veri also provides daily scores, weekly reports, and an ongoing metabolic health span, which calculates your metabolic health and compares it to the average. I’m below average on that one, with high glucose variability, and oscillation, and a high morning fasting glucose. It’s a lot of information, but I like that it doesn’t just hang there in isolation. I can see what happened and why. And for the most part, I know what to do about it.

Overall, I’m not happy where I am, health-wise, and I have a terrible history of failing to fix it. I know that the continuous glucose monitor isn’t the answer so to speak, but I had hoped that it would provide some insights that might put me on the right path, and it has so far exceeded my expectations. The next steps are, of course, the most important. But the big difference between what I’m doing now and my keto experience of 2017 is that I’ll be using data to guide the way. We’ll see what happens, but I do feel pretty good about that.