H HiPerformance CultureThe Evidence
Bio 4 min read
Trend Breakdown
The Evidence

What does a continuous glucose monitor actually tell a healthy person?

CGMs have crossed from diabetes clinics to wellness stacks, promoted as a window into metabolic performance. The signal is real: glucose does fluctuate in response to food, sleep, and exercise. But the accuracy problem is significant, clinical benefit in healthy adults is unproven, and the evidence base is thin.

Published 4 Jun 2026 · 5 sources
Skip to the verdict ↓
Trend Science
Breakdown
Evidence-graded series
02What's being claimed

Blood glucose fluctuates in real time with every meal, workout, and poor night of sleep. A CGM makes that invisible physiology visible, giving a healthy person a live readout of how their daily choices move their metabolic state.

Glucose variability is a genuine physiological fact. Shah et al. confirmed that healthy adults spend around 96% of their day within the 70-140 mg/dL range 1, but that remaining 4% can include meaningful postprandial excursions driven by carbohydrate load, sleep quality, and stress. A CGM captures this in real time, making invisible physiology visible in a way that annual blood panels never could. For someone trying to understand why a particular meal leaves them lethargic, or why broken sleep correlates with mid-morning cravings, that data has intuitive appeal.

The spread followed a clear logic: a plausible mechanism, accessible consumer hardware, and biohacker influencers who reframed an insulin-management device as a universal metabolic sensor. Peter Attia and Tim Ferriss promoted CGM self-experimentation to mass audiences from 2019, and Levels Health raised $38 million by 2022 by normalising the device as a wellness tool. Richardson et al. confirmed CGM's behaviour-change potential 2, and the consumer appeal is easy to grasp: if a wearable can show that your glucose spiked after a specific meal, the argument for dietary modification becomes concrete.

Origin
Diabetic device technology
CGMs were developed for insulin-dependent diabetes management, launching commercially from 2006.
→
Vector
Biohacker podcasters
Peter Attia and Tim Ferriss promoted CGM self-experimentation to mass wellness audiences from 2019.
→
Spike
Levels Health app
Levels raised $38 million by 2022 and normalised consumer CGM as a mainstream wellness tool.
"Since wearing a CGM for two weeks I know exactly which foods spike my glucose and which do not. I have cut out three things I thought were healthy. Everyone should see this data about themselves; it genuinely changes how you eat."
— Representative of the claim as it circulates in health podcasts and wellness communities
03The evidence verdict
H
HiPerformance Culture The Evidence · Trend Breakdown
Verdict

CGMs reveal real glucose patterns in healthy adults, but sensor accuracy is limited and clinical benefit is unproven.

Hype Evidence
This trend lands here
Low Moderate High
Low confidence 5 sources cited · 1 multicenter prospective study, 1 systematic review and meta-analysis, 1 narrative review, 1 expert survey study, 1 accuracy validation study · 2019-2025
✓

What holds up

✓
Normative CGM benchmarks for healthy adults are well-established: Shah et al. found median time in range (70-140 mg/dL) of 96% and mean glucose of 98-99 mg/dL in adults under 60. 1
Gold
✓
CGMs can reveal how specific foods, exercise timing, and sleep affect an individual's glucose curve in ways that aggregate metrics like HbA1c cannot capture. 2
Silver
✕

What doesn't

✕
FreeStyle Libre2 showed a 27.5% mean absolute relative difference in healthy women, well below ISO accuracy standards; only 45% of readings met the ±15% threshold. 5
Safety-critical Gold
✕
Only 3 of 25 RCTs on CGM as a behaviour-change tool enrolled non-diabetic subjects; no RCT has demonstrated improved health outcomes in normoglycemic individuals. 2
Gold
✕
Expert diabetologists show only fair agreement (kappa 0.36) on interpreting non-diabetic CGM reports; commercial CGM wellness claims have been characterised as misleading. 4 3
Silver
✕
CGM use in non-diabetic people carries documented psychological risks: anxiety about dietary norms and a risk of orthorexia development; skin reactions affect a substantial proportion of wearers. 3
Safety-critical Silver
04The studies
Scored on Design quality Measurement precision Causal clarity Replication value
Gold
96% median time in range (70-140 mg/dL) for healthy adults under 60
Multicenter prospective study · n=153
Shah et al. Journal of Clinical Endocrinology and Metabolism · 2019
Established modern normative CGM benchmarks: median time in range (70-140 mg/dL) of 96%, mean glucose of 98-99 mg/dL in adults under 60, and a glucose coefficient of variation of 17%. These numbers are what healthy CGM users should expect to see; most will find their data looks entirely unremarkable.
doi:10.1210/jc.2018-02763 Verify ↗
Gold
3/25 RCTs on CGM behaviour change that enrolled non-diabetic participants
Systematic review and meta-analysis · 25 RCTs, n=2,996
Richardson et al. International Journal of Behavioral Nutrition and Physical Activity · 2024
Of 25 RCTs examining CGM as a behaviour-change tool, only 3 enrolled non-diabetic subjects. Overall HbA1c reduction was modest (0.28%) and BMI changes were non-significant. Authors explicitly called for more research in non-diabetic populations, noting a stark evidence gap for healthy users.
doi:10.1186/s12966-024-01692-6 Verify ↗
Silver Narrative review · 27 studies (1980-2023)
Oganesova, Pemberton & Brown Diabetic Medicine · 2024
Found no consistent high-quality evidence supporting CGM utility for detecting abnormal glucose, promoting behaviour change, or improving metabolic health in non-diabetic people. Commercial wellness claims were characterised as misleading. Identified psychological risks including anxiety about dietary norms and orthorexia development. Called for strengthened post-market oversight.
doi:10.1111/dme.15369 Verify ↗
Silver
0.36 inter-clinician kappa agreement on non-diabetic CGM reports (fair = 0.21-0.40)
Expert survey study · 18 diabetologists, 20 CGM reports
Spartano et al. Journal of Diabetes Science and Technology · 2025
Expert clinicians showed only fair agreement (Fleiss kappa 0.36) on how to act on CGM reports from non-diabetic individuals. Normal postprandial glucose excursions routinely triggered follow-up recommendations, and the Glucose Management Indicator was found unsuitable as an HbA1c surrogate in this population.
doi:10.1177/19322968251315171 Verify ↗
Gold
27.5% mean absolute relative difference of FreeStyle Libre2 in healthy women
Accuracy validation study · n=29 healthy women
Jin et al. Sensors (Basel) · 2023
The FreeStyle Libre2 showed an overall mean absolute relative difference of 27.5% in healthy women; only 45% of readings met the ISO accuracy standard (plus or minus 15%). Accuracy deteriorated further during falling glucose (MARD 63.7%), and the sensor systematically overestimated glucose by 1.14 mmol/L. These findings were under controlled conditions in healthy, non-overweight participants.
doi:10.3390/s23177417 Verify ↗
05So what do you actually do

If you want to try a CGM, the evidence supports a short, structured experiment with a defined question.

Frame it as a 2-4 week experiment; do not treat every reading as actionable.

01Run a 2-4 week structured experiment with a defined question, such as how specific foods or exercise timing affect your glucose curve.
02Focus on patterns over days rather than individual readings; sensor error margins make single readings unreliable.
03Do not act on any reading outside the normal range without confirmatory venous blood testing.
04If you have a first-degree family history of type 2 diabetes, a CGM experiment may offer genuine early-warning value.
05If you notice anxiety about food choices in response to CGM data, discontinue use and seek clinical guidance.
06The verdict triad
Claim

Glucose variability is real

Blood glucose fluctuates across the day in response to every meal, workout, and night of poor sleep. This variability is physiologically real and measurable. A CGM makes that invisible signal visible, translating diffuse feelings of low energy or brain fog into a concrete data trace.

Consequence

Poor patterns carry real cost

Persistent postprandial spikes and poor glycaemic control associate with fatigue, impaired cognitive performance, and elevated long-term cardiometabolic risk. Most healthy adults will find their CGM data looks normal, but for those with undiagnosed prediabetes, the signal may carry genuine early-warning value that HbA1c would miss entirely.

Lever

Short experiment, sceptical eye

Use a 2-4 week CGM experiment to identify personal dietary patterns and response curves. Treat outlier readings with scepticism; sensor error is too high to act on individual data points. Confirm any reading outside the normal range with venous blood testing before drawing conclusions or changing clinical care.

08What to do next
What to do next

Could your glucose data be pointing to something worth investigating further?

HPC's Metabolic Health Assessment maps your dietary patterns, sleep quality, and activity load against established metabolic risk markers. If your CGM experiment surfaced anything unusual, this is how you take it further.

Take the assessment
09Share & references
Update log
4 Jun 2026First published. 5 sources reviewed. Verdict: Low evidence.
Related
Owning guide
The HPC Guide to Metabolic Health
Evidence-based practices for optimising glucose, insulin sensitivity, and long-term cardiometabolic health.
Read the guide →
Assessment
Metabolic Health Assessment
Map your glucose patterns, diet, and lifestyle against established metabolic risk markers.
Start assessment →
Trend Breakdown
Zone 2 Cardio: What the Science Actually Shows
How low-intensity aerobic training affects metabolic flexibility and longevity markers.
Read the breakdown →
Bibliography · every source, resolvable
01Shah, V.N., DuBose, S.N., Li, Z., Beck, R.W., Peters, A.L., Weinstock, R.S., Kruger, D., Tansey, M., Sparling, D., Woerner, S., Vendrame, F., Bergenstal, R., Tamborlane, W.V., Watson, S.E. & Sherr, J. (2019). Continuous Glucose Monitoring Profiles in Healthy Nondiabetic Participants: A Multicenter Prospective Study. The Journal of Clinical Endocrinology & Metabolism, 104(10), 4356-4364. doi:10.1210/jc.2018-02763 Verify ↗Gold
02Richardson, K.M., Jospe, M.R., Bohlen, L.C., Crawshaw, J., Saleh, A.A. & Schembre, S.M. (2024). The efficacy of using continuous glucose monitoring as a behaviour change tool in populations with and without diabetes: a systematic review and meta-analysis of randomised controlled trials. International Journal of Behavioral Nutrition and Physical Activity, 21(1). doi:10.1186/s12966-024-01692-6 Verify ↗Gold
03Oganesova, Z., Pemberton, J. & Brown, A. (2024). Innovative solution or cause for concern? The use of continuous glucose monitors in people not living with diabetes: A narrative review. Diabetic Medicine, 41(9). doi:10.1111/dme.15369 Verify ↗Silver
04Spartano, N.L., Prescott, B., Walker, M.E., Shi, E., Venkatesan, G., Fei, D., Lin, H., Murabito, J.M., Ahn, D., Battelino, T., Edelman, S.V., Fleming, G.A., Freckmann, G., Galindo, R.J., Joubert, M., Lansang, M.C., Mader, J.K., Mankovsky, B., Mathioudakis, N.N., Mohan, V., Peters, A.L., Shah, V.N., Spanakis, E.K., Waki, K., Wright, E.E., Zilbermint, M., Wolpert, H.A. & Steenkamp, D.W. (2025). Expert Clinical Interpretation of Continuous Glucose Monitor Reports From Individuals Without Diabetes. Journal of Diabetes Science and Technology, 20(3), 727-735. doi:10.1177/19322968251315171 Verify ↗Silver
05Jin, Z., Thackray, A.E., King, J.A., Deighton, K., Davies, M.J. & Stensel, D.J. (2023). Analytical Performance of the Factory-Calibrated Flash Glucose Monitoring System FreeStyle Libre2TM in Healthy Women. Sensors, 23(17), 7417. doi:10.3390/s23177417 Verify ↗Gold
Build better performance.
Get the 90-day protocol.