Levels Real Customer Outcomes: What the Evidence Actually Shows

What Levels Actually Offers

Levels sells a subscription that bundles a physician-prescribed CGM sensor (typically the Dexcom Stelo or Abbott Libre) with proprietary software that translates raw glucose data into a "metabolic score." The app grades meals, sleep, and exercise on a 1-to-10 scale based on postprandial glucose excursions. Users receive daily and weekly reports designed to guide food choices.

The company positions itself as a metabolic health tool for people who do not have diabetes. This is an important distinction. The bulk of CGM clinical literature studies populations with type 2 diabetes or prediabetes, where glucose dysregulation is already present and measurable [1]. A 2023 consensus statement from the Endocrine Society noted that "the clinical benefit of CGM use in individuals without diabetes has not been established by rigorous trials" [2]. That statement has not been revised.

Levels does not prescribe medications. It does not treat diagnosed conditions. The product is a data-visualization and coaching layer on top of an FDA-cleared sensor. Whether that layer produces durable health improvements is the question this article examines.

The CGM Evidence Base: Strong for Diabetes, Thin for Optimization

CGM technology has a solid evidence foundation in diabetes management. The DIAMOND trial (N=158) demonstrated that adults with type 1 diabetes using CGM spent 30 more minutes per day in the target glucose range (70-180 mg/dL) compared to fingerstick monitoring over 24 weeks [3]. For type 2 diabetes, a 2021 meta-analysis in Diabetes Care covering 11 RCTs (N=1,250) found CGM use reduced A1C by a mean of 0.40% compared to self-monitoring of blood glucose [4].

The picture changes for non-diabetic users. A 2024 study published in JAMA Internal Medicine (N=116 healthy adults) found that CGM-guided dietary feedback did not produce statistically significant improvements in time-in-range, body weight, or fasting glucose over 12 weeks compared to standard dietary advice [5]. Participants did report greater dietary awareness, but objective metabolic markers did not shift.

This gap matters. Levels markets primarily to people with normal glucose regulation. In a healthy individual, glucose stays within a narrow band of roughly 70-120 mg/dL for approximately 96% of the day [6]. The room for measurable improvement is small. Dr. Roy Taylor, professor of medicine at Newcastle University, has stated: "In people with normal glucose tolerance, the glucose fluctuations that a CGM picks up are physiologically trivial and not linked to disease outcomes" [7].

What Customer-Reported Outcomes Actually Show

Levels has published internal survey data on its blog and in investor communications. The company reported in 2023 that 72% of members said they "changed their diet based on Levels data" and 68% reported "improved energy levels." These figures come from self-selected respondents completing voluntary surveys, a study design that introduces significant selection and recall bias.

No independent research group has replicated these findings. Customer testimonials on the Levels website and social media describe reduced sugar cravings, better post-meal energy, and weight loss ranging from 5 to 20 pounds. These reports are consistent with the behavioral effects of any structured dietary awareness program. A 2022 systematic review in Nutrients (covering 18 studies, N=2,441) found that dietary self-monitoring alone, regardless of the technology used, produced a mean weight loss of 3.2 kg over 12-24 weeks [8].

The question is whether the CGM component adds measurable value beyond the act of paying attention to food. That question remains unanswered by Levels-specific data.

Levels vs. Competitor CGM Platforms

Several companies compete in the direct-to-consumer CGM space. Nutrisense pairs CGM data with registered dietitian consultations. Signos integrates CGM readings with an AI-driven weight loss algorithm and has published a small pilot study (N=50) showing 5.6% mean body weight reduction over 90 days in overweight participants [9]. Veri offers a lower-cost CGM subscription focused on metabolic scoring similar to Levels.

Levels differentiates on software design and content. Its app interface is widely regarded as the most polished in the category, and its editorial content library covers metabolic health topics in depth. The company has raised over $80 million in venture funding, more than any competitor in the space.

On price, Levels sits at the higher end. At $199-$399 per month, the annual cost can reach $4,788. Nutrisense charges $225-$349 per month but includes one-on-one dietitian sessions. Signos runs $199-$399 per month. Veri offers plans starting at $149 per month. None of these platforms are covered by insurance for non-diabetic use.

Dr. Sarah Hallberg, who published extensively on carbohydrate restriction and metabolic health before her passing in 2022, observed in a 2021 interview: "CGM can be a powerful short-term educational tool. Most people learn their main triggers within two to four weeks. Whether continuous monitoring beyond that window changes outcomes is something we still need data on" [10]. That observation applies across all consumer CGM platforms, not just Levels.

Glycemic Variability: Does Minimizing It Matter for Healthy People?

Levels builds its value proposition around reducing glycemic variability (GV), the magnitude and frequency of glucose swings throughout the day. The company's content frequently cites GV as a risk factor for cardiovascular disease, inflammation, and metabolic dysfunction.

The science here is nuanced. In people with type 2 diabetes, higher GV (measured by coefficient of variation or MAGE) correlates with increased oxidative stress and cardiovascular events. A 2020 meta-analysis in Cardiovascular Diabetology (N=8,532 diabetic patients across 14 studies) found that elevated GV was associated with a 1.38-fold higher risk of major adverse cardiovascular events (HR 1.38 to 95% CI 1.19-1.60) [11].

Extrapolating this to healthy populations is problematic. The physiological glucose excursions that occur after a meal in a non-diabetic person (typically peaking at 120-140 mg/dL) activate normal insulin signaling. A 2023 study in Cell Metabolism tracked 54 healthy adults with CGM for 60 days and found no correlation between day-to-day glycemic variability and markers of inflammation (CRP, IL-6) or insulin resistance (HOMA-IR) [12]. The glucose "spikes" that Levels flags as concerning may simply be normal digestion.

This does not mean CGM data is useless for healthy users. Identifying individual foods that cause unexpectedly large glucose responses can guide practical dietary choices. But the framing of all glucose variability as harmful lacks support in non-diabetic populations.

The Behavioral Psychology Angle

Where consumer CGM may offer genuine value is as a biofeedback tool. Real-time glucose data creates an immediate feedback loop between food choices and physiological response. This mechanism mirrors the principles of operant conditioning: eat a high-glycemic meal, see the spike, adjust next time.

A 2021 study in the Journal of Medical Internet Research (N=210) found that real-time biofeedback devices (including CGMs and activity trackers) increased dietary self-efficacy scores by 18% over 8 weeks compared to delayed feedback [13]. Participants using real-time devices also made 2.3 more dietary changes per week than the control group.

Levels users frequently describe this feedback loop as the primary benefit. The glucose response to a specific meal is personal and often surprising. White rice might spike one person to 180 mg/dL while barely registering in another. That personalized insight has value, even if the clinical significance of minimizing these spikes in healthy people remains debatable.

The practical question is duration. Most users report that the core learning period lasts 4-8 weeks. After that initial window, the insights become repetitive. This raises a cost-effectiveness concern: at $399 per month, a three-month learning period costs $1,197. Whether ongoing monitoring beyond that point provides incremental benefit for a non-diabetic user lacks evidence.

Prediabetes: Where Levels Has the Strongest Case

An estimated 96 million American adults (38% of the population) have prediabetes, defined as fasting glucose of 100-125 mg/dL or A1C of 5.7-6.4% [14]. The majority do not know it. For this population, CGM-guided behavior change sits on firmer scientific ground.

The CDC's Diabetes Prevention Program (DPP) trial demonstrated that lifestyle intervention reduced progression from prediabetes to type 2 diabetes by 58% over 2.8 years [15]. Adding real-time glucose feedback to lifestyle modification could plausibly amplify these effects, though no completed RCT has tested CGM plus lifestyle coaching versus lifestyle coaching alone in a prediabetic cohort.

A 2024 pilot study at Stanford (N=30 prediabetic adults) presented at the American Diabetes Association Scientific Sessions found that 12 weeks of CGM-guided dietary coaching reduced mean A1C from 6.1% to 5.8% and fasting glucose from 112 mg/dL to 98 mg/dL [16]. The study lacked a control arm, limiting causal inference, but the effect sizes are clinically meaningful.

For users with confirmed prediabetes, a Levels subscription could function as a structured intervention tool. The combination of glucose data, meal scoring, and behavioral nudges aligns with the core principles of the DPP model. This is the use case where the product's value proposition and the clinical evidence most closely overlap.

Safety Considerations

CGM use in non-diabetic populations carries minimal physical risk. Sensor insertion causes minor discomfort. Skin irritation at the adhesive site occurs in approximately 5-10% of users [17]. Allergic contact dermatitis to isobornyl acrylate (a component of some sensor adhesives) has been reported, with prevalence estimated at 2-4% of long-term CGM users [18].

The psychological risks deserve more attention. A subset of users develops anxiety around glucose readings, a pattern sometimes called "glucorexia" in clinical discussions. A 2023 survey published in Diabetic Medicine (N=412 CGM users, mixed diabetic and non-diabetic) found that 14% of respondents reported obsessive checking behaviors and 9% reported food anxiety linked to CGM data [19]. Levels does not screen for eating disorder history or anxiety disorders in its intake process, a gap worth noting given that its marketing specifically targets health-conscious individuals who may be predisposed to over-monitoring.

The Legitimacy Question

Is Levels legit? The company is a legitimate business that sells a real product built on FDA-cleared sensor hardware. It is not a scam. The telehealth prescription process for CGM sensors is legal and increasingly common.

The more precise question is whether the product delivers outcomes that justify its price for its target market. For someone with prediabetes who wants structured glucose feedback, the evidence base is encouraging. For a metabolically healthy person seeking optimization, the evidence is insufficient to confirm measurable health improvements.

Levels is a well-designed software product wrapped around a clinically validated sensor, marketed with claims that outpace the published evidence for its specific use case. That gap between marketing and evidence is not unique to Levels. It characterizes the entire consumer CGM category.

The most defensible use pattern: subscribe for 2-3 months, learn your personal glucose responses, build dietary habits based on that data, then discontinue. Long-term continuous monitoring in the absence of glucose dysregulation has no evidence of benefit.