Are continuous glucose monitors worth it if you don't have diabetes?

Continuous glucose monitors are now available over the counter, cost about eighty dollars a fortnight, and are being marketed hard to people whose glucose is entirely normal. I have opinions about this, and they are more mixed than either camp would like.

What a CGM genuinely gives a non-diabetic person

The honest answer is: a feedback loop, and almost nothing else. It will not diagnose anything — an A1c and a fasting insulin will tell you more about your metabolic health than a fortnight of sensor data. What it does is collapse the delay between an action and its consequence from months to ninety minutes, and that is not a small thing for a behaviour change.

The most useful outcomes I have seen from client CGM trials are all of that shape. One discovered that his 6am oatmeal — which he had been eating for a decade on the grounds that it was healthy — put him at 180 mg/dL and dropped him into a slump at 10am, and that adding eggs and walnuts to the same oatmeal flattened it. Another found that a twelve-minute walk after dinner cut her post-meal peak by roughly 30 points, reliably, every time she tested it. Neither of those facts is exotic. Both were far more persuasive as a line on a phone than as advice from me.

Three things people misread

1. 1A spike is not damage. Glucose rising after eating carbohydrate is your metabolism working correctly. What matters is the height, the duration, and whether it comes back down — not the existence of a curve.
2. 2The sensor is not accurate to the digit. Consumer CGMs read interstitial fluid, not blood, with a lag of 5 to 15 minutes and a mean absolute relative difference around 8 to 10 percent. A reading of 141 might be 128. Read shapes and trends, not points.
3. 3Compression lows are real and terrifying. Sleep on the sensor and you will see a fake 55 at 3am. Every time. It is the arm, not the pancreas.

How to run one usefully

If you are going to do it, do it as an experiment with a question rather than as ambient surveillance. Two weeks, one sensor, and a plan written before it goes on.

• Days 1–4: change nothing at all. Eat exactly as you normally do. This is your baseline and it is the part everyone skips.
• Days 5–10: test specific swaps. Same breakfast with and without protein added. Same dinner with and without a post-meal walk. One variable at a time, repeated at least twice, because day-to-day variation is larger than you would guess.
• Days 11–14: implement the two changes that showed the largest effect, and confirm they hold.
• Then take it off. Write down what you learned. The learning is the deliverable, not the ongoing data stream.

My actual position

For a person with prediabetes, a family history, or a fasting insulin that has caught my attention, a two-week CGM trial is one of the better eighty dollars they will spend, because it converts an abstraction into something they can see. For a metabolically healthy 34-year-old who saw an advertisement, it is a device that will mostly teach them to be frightened of fruit.

And in every case: two weeks, a question, then off. The people who wear one indefinitely are not, in my experience, learning anything after the first month. They are just watching.