Continuous glucose monitoring made me continuously crazy
MAHA is obsessed with these wearables — for all the wrong reasons.
By Victoria Song
Photography by Amelia Holowaty Krales / The Verge
Apr 8, 2026, 11:00 AM UTC
Victoria Song
is a senior reporter and author of the Optimizer (https://www.theverge.com/optimizer-newsletter) newsletter. She has more than 13 years of experience reporting on wearables, health tech, and more. Before coming to The Verge, she worked for Gizmodo and PC Magazine.
Daily life is different when you’re tracking glucose. A little over a year ago, I was on my way to a conference. My bags were packed, the Uber was on its way, but there was one last thing to do before I could head to the airport. Tearing open a small isopropyl alcohol wipe, I cleaned the skin on the back of my arm. After that, I applied a small applicator to the clean skin, doing my best to ignore the visible needle inside. I squeezed my eyes shut and pressed a button. It made a ka-thunk. I repeated the process on the other arm.
In my right arm, I now had a Dexcom Stelo. In the left, an Abbott Lingo. Both were over-the-counter continuous glucose monitors (CGM) that would monitor the rise and fall of my glucose levels. Opening my phone, I checked both the Dexcom and Abbott apps to make sure the CGMs were transmitting data. I made a mental note to check how high altitudes might impact readings. It crossed my mind that, to my surprise, I’d felt zero pain.
There was no urgent medical reason why I needed to track my glucose. I’m not a diabetic. My A1C levels — the metric that measures long-term blood sugar — have always been good. But glucose tracking isn’t just for diabetics anymore. On social media, you can see doctors, wellness influencers, biohackers, and athletes talking about CGM use. I just happen to test health tech, so I thought I’d give it a whirl for a few weeks and see if there was any benefit for a non-diabetic like myself using this tech.
Instead, I ended up spending over a year testing the devices, reading up on studies, speaking with researchers, and falling down rabbit holes. I bounced from doctor to doctor trying to figure out if there was actually something wrong with me — or the devices I was using.
The first “professional use” CGM was cleared by the Food and Drug Administration in 1999. Most people think these devices are used to track blood sugar, but that’s not entirely correct. Technically, they provide real-time glucose measurements from the interstitial fluid between your cells, just underneath your skin. Compared to traditional finger-stick tests, which directly measure blood sugar, CGMs can track glucose trends over an extended period of time.
Until 2024, CGMs required a prescription and were devices primarily used by Type 1 diabetics — people who produce little to no insulin. Now, both Dexcom and Abbott sell CGMs targeted at non-diabetics, prediabetics, and Type 2 diabetics who don’t rely on insulin. To differentiate, sometimes you’ll see companies market over-the-counter devices as “glucose biosensors.”
The benefits of using CGMs for prediabetics and Type 2 diabetics are clear. Unlike Type 1 diabetes, prediabetes and Type 2 tend to develop over time as the body becomes more resistant to insulin. If caught early enough, it can be “reversed” with lifestyle interventions like changes to diet and exercise. People with pre- and Type 2 diabetes also make up the overwhelming majority of cases in the US. According to the American Diabetes Association (https://diabetes.org/about-diabetes/statistics/about-diabetes), as of 2021, Type 2 diabetics make up about 95 percent — or roughly 36 million — of the estimated 38.4 million Americans with diabetes. Meanwhile, about 98 million were estimated to have prediabetes. Put all that together, and a significant number of people could potentially learn a lot from using CGMs about how their dietary habits impact a legitimate metabolic condition.
Less clear is whether this technology is useful for non-diabetics. But that hasn’t stopped the push for this tech, from both the CGM makers and the government. If Health Secretary RFK Jr. has his way, in four years, everyone might be wearing one of these (https://www.theverge.com/analysis/756994/rfk-jr-wearables-maha-health-wearables-disordered-eating). Underscoring this, President Donald Trump’s controversial surgeon general nominee Casey Means is also the cofounder of Levels, a CGM startup aimed at non-diabetics. In her book Good Energy, she calls out the technology as a useful tool in fixing metabolic dysfunction — something she claims is at the root of every possible chronic ailment today (https://www.theverge.com/column/889922/optimizer-casey-means-wellness-influencer-playbook). However, several medical experts have publicly questioned this logic, stating the evidence for non-diabetics simply isn’t there (https://publichealth.jhu.edu/2026/is-glucose-monitoring-useful-for-non-diabetics).
But we’ve now entered a new era where wearable technology offers a tantalizing promise. It’s not just about flagging potential illnesses. It’s about optimizing your body’s biometric data to live the longest, healthiest life possible. In the context of the US’s terribly flawed healthcare system, this tech has often been positioned as giving back a degree of control to the average person. Wear this device, track your health, learn more about yourself, and subsequently make better choices.
Optimization of metabolic health has been a rising trend for years. Most commonly, this has come in the form of food logging. However, CGM-related metabolism tracking has gained traction due to the extra layer of data it provides. January AI (https://www.january.ai/) was originally a CGM startup that then pivoted to providing glucose spike prediction within a meal logging app based on the data it collected. Meanwhile, Oura has since partnered with Dexcom (https://www.theverge.com/news/661069/oura-dexcom-stelo-meals-glucose-metabolic-health-wearables) to provide glucose readings, meal logging, and AI interpretations inside its own smart ring app. Abbott also recently partnered with Withings (https://www.theverge.com/tech/853413/abbott-lingo-withings-cgm-wearable-health-ces-2026), another wearables maker, for a similar integration. When you factor in that GLP-1 use is increasing, alongside the return of 2000s-era, ultra-skinny diet culture (https://www.elle.com/uk/life-and-culture/culture/a65463685/diet-culture-back/), it seems like CGM use is being positioned as the next evolution of fitness tracking.
Why not try biohacking my nutrition to manage my diabetes risk, gain insight into PCOS, and be a better athlete?
“It’s predicted one in two Americans will have obesity by the year 2030,” says Dr. Thomas Grace, a diabetes clinician with Type 1 diabetes who consulted with Dexcom on the Stelo CGM. “I think the most exciting thing for myself, for my patients, and for people using Stelo is the instant reward they get from understanding how food, activity, stress, and sleep affect their overall glucose health.”
I’m the exact sort of non-diabetic that Dexcom, Abbott, and other CGM startups are targeting. I’ve got a family history of Type 2 diabetes and high cholesterol. I’ve been diagnosed with polycystic ovary syndrome (PCOS) — a chronic condition that numerous doctors have told me means I likely either have insulin resistance or chronic inflammation that makes me predisposed to diabetes. As a runner, I’ve struggled with energy levels while training for long-distance races despite following standard fueling practices and carb loading. Why not try biohacking my nutrition to manage my diabetes risk, gain insight into PCOS, and be a better athlete?
Wearing a CGM 24/7 is invisible until it isn’t. I went days without remembering they were in my arms. Then, they’d inevitably snag on my shirt sleeve, or I’d brush into a doorframe, the contact popping the CGM out of my arm. Now, even when I don’t wear CGMs, I’m aware of the back of my arms and the cut of all my shirts. That hypervigilance comes with being constantly monitored — even if you’re the one doing the monitoring.
At first, I’d review my data every morning, after each workout, and a few hours after each meal. Most of what I saw was pretty normal. A bowl of pasta? Glucose spike. Roasted salmon and a side salad? Minimal increase. Carb loading before a long run? Mondo spike, followed by a sharp decrease. (I found CGM use too cumbersome for mid-run fueling, as there’s a five-minute delay between readings.)
Nevertheless, reviewing my data multiple times a day began spiking my anxiety. Both Dexcom and Abbott’s apps have educational articles about what the ideal glucose range is for healthy nondiabetics — 70 to 140mg/dL. Fasting glucose levels — like when you’re sleeping — ought to sit in the range of 70 to 99mg/dL. Those first few months, I woke up well beyond 100mg/dL every day, even without late-night snacks. Sometimes, I’d wake up to see the Dexcom app had alerted me to glucose spikes while I slept. (Abbott doesn’t send spike notifications for Lingo.)
A year prior, I had tested Nutrisense (https://www.theverge.com/23844553/glucose-monitor-wearables-nutrisense-review) — which also makes use of CGMs — for two weeks and never had elevated morning glucose levels. Clearly, diabetes had finally come for me.
Elevated morning glucose is a concern because of the Dawn Phenomenon (https://my.clevelandclinic.org/health/diseases/24553-dawn-phenomenon). To help prepare your body to wake up, you produce hormones like cortisol and growth hormone. That signals to your liver to release glucose, giving you energy for the day. In diabetics and insulin-resistant people, however, it leads to overly high blood sugar levels during the morning. Something, perhaps, like what I was seeing.
Cue anxiety, scheduling a flurry of doctors’ appointments, and hyperfixating on my diet and exercise. At the doctor’s office, I got shrugs, quizzical brow raises, and a reluctant acquiescence to run blood tests. (“Do they really have non-diabetics wearing those?” a nurse asked while taking my blood pressure.) My A1C, a gauge for your average glucose levels over two to three months, was perfectly normal. No diabetes here. I did, however, have elevated liver enzymes and cholesterol levels.
Out of an abundance of caution, my doctor ordered an ultrasound. I had to wait several weeks for an appointment. Later, once the goo was slathered over my abdomen, I watched nervously out of the corner of my eye as an ultrasound technician muttered to herself. After another few days, I got a call from the doctor. I was diagnosed with a “nothing to worry about right now” case of non-alcoholic fatty liver — a condition that commonly occurs alongside PCOS. Maybe, my doctor suggested, I should cut out all alcohol (I rarely drink) and lose body fat (I’ve been trying for 10 years). Come back in about a year, they said, and we’ll see where you’re at.
Read one way, this could be a success story for CGMs. The devices flagged something had changed, and though it wasn’t diabetes, I had at least one new official diagnosis. Glass half full: Control over my health had been handed back to me.
A more skeptical take is that elevated liver enzymes and high cholesterol had been present in previous blood work. A former doctor had also suspected fatty liver disease, but said an ultrasound wasn’t necessary unless my levels worsened. Were my levels worse at this CGM-inspired blood test? Not particularly. The advice I’d been given was the same as it had been in years past. Glass half empty: I learned nothing new.
It wasn’t until much later, after consulting with Dexcom and Abbott, that I learned side sleeping could lead to inaccurate overnight glucose readings. When you’re on your side, the CGM can get compressed. That, in turn, could lead to readings lower or higher than your actual glucose levels. I tried swapping arms for the Lingo and Stelo to test for this, but it was impossible to account for how I shifted positions each night. Either way, I continued getting high overnight and morning glucose readings from both sensors for months.
The worst part was that I couldn’t tell whether the data was inaccurate or my doctor had missed something.
I started running into issues after the first month of continuous wear. For starters, it can be pricey. Dexcom and Abbott both gave me several test units, but for the average person, it can cost around $100 a month. A single sensor lasts about 15 days if everything goes well. Sometimes, they malfunction or get ripped out after getting caught on clothing and other objects. The adhesive also leaves stubborn residue that takes weeks of showering to fully remove. Partly to stretch out my supply, partly to give my skin a break, I started testing two weeks of every month. After six months, I pared down to once a quarter.
But regardless of how often a person uses CGMs, every body is different. Foods that cause a spike for me may do nothing for you. The only way to know for sure is through diligent logging and experimentation. The conundrum is that this creates a mountain of data, and as I learned, interpreting that data without adequate context can lead to unnecessary anxiety.
Compounding the issue? CGM makers have differing approaches on how to present that data. The Dexcom Stelo app, for example, will give you spike alerts after about an hour or two. Abbott’s Lingo CGM eschews such alerts altogether. Instead, Abbott opts for a Lingo Score that tries to simplify raw data into a digestible number that signifies how well you did on average at keeping your glucose within a healthy range of 70 to 140mg/dL.
If you wear two CGMs simultaneously, as I did, it’s possible to get different numbers and not know which one is truly correct
Using a CGM to stay within that range sounds relatively simple and easy. In reality, it’s complicated.
“The one study that did look at follow-up outcomes found that people who spent more time above that range were more likely to get diabetes, but this was in a pretty small population,” says Nicole Spartano, assistant professor at Boston University’s Chobanian and Avedisian School of Medicine. “They may have already had prediabetes and are sort of on their way already. I think we’re really at a point where we don’t have a lot of information from a research standpoint.”
For example, Spartano notes that there’s often a lack of context for how CGM numbers could relate to an individual’s health. On the one hand, doctors often see a fasting blood glucose level of over 100mg/dL as a cause for concern. CGMs, however, do not measure blood glucose. They measure interstitial glucose, which can, at times, be higher than blood glucose. Spartano says more research is needed to determine what the differences between blood and interstitial glucose could be, both broadly and on an individual level.
“We assume that they’re completely accurate, but they’re not exactly accurate. There’s a certain level of accuracy that the FDA requires, but that still leaves some wiggle room,” says Dr. David Klonoff, medical director at the Diabetes Research Institute at Mills-Peninsula Medical Center and editor-in-chief of the Journal of Diabetes Science and Technology.
Basically, if you wear two CGMs simultaneously, as I did, it’s possible to get different numbers and not know which one is truly correct. And while most people would only wear one at a time, there could also be a difference between one CGM you wear and the next, based on whether they were applied correctly and several other factors.
But say everything is assumed to be accurate. What is the best way to interpret CGM data for non-diabetics? To try and find out, Spartano ran a clinical study (https://pubmed.ncbi.nlm.nih.gov/39936548/) in which 18 endocrinologists were asked to evaluate data from non-diabetics.
“We gave them 20 different glucose monitor reports and asked them, ‘If someone came in with this report, would you suggest they have a follow-up screening for this?’” says Spartano. She says some experts viewed peaks as a normal part of physiology. Others saw elevated levels as a sign someone might need further testing.
“Essentially, there was no consensus. Even clinicians who read CGM data all day, every day don’t know what to do with this data,” says Spartano. “One of the problems is we don’t have an ideal way of analyzing the information yet. We know if someone is doing really poorly or completely normally, but the people that seem to be headed for trouble? We can only say, ‘It doesn’t look normal, but it’s not abnormal. It’s something in the middle,’” agrees Klonoff.
“Essentially, there was no consensus. Even clinicians who read CGM data all day, every day don’t know what to do with this data.”
According to Klonoff and Spartano, there isn’t a reference set of data that’
