What Is a CGM Monitor? How It Works and Who Needs It

A continuous glucose monitor, or CGM, is a small wearable device that tracks your glucose levels around the clock by reading the sugar concentration in the fluid just beneath your skin. Instead of relying on periodic finger-prick blood tests, a CGM takes a new reading every few minutes and sends the data to a smartphone or receiver, giving you a rolling picture of where your glucose has been, where it is now, and where it appears to be heading. Originally developed for people with type 1 diabetes, CGMs are now used across a much wider range of situations, from type 2 diabetes management to pregnancy to, increasingly, general wellness.

How a CGM Actually Works

A CGM system has three main parts. The first is a tiny sensor filament, thinner than a human hair, that sits just under the skin, usually on the back of the upper arm or the abdomen. This filament is embedded in the interstitial fluid, the thin layer of liquid that surrounds your cells. The sensor contains an enzyme (glucose oxidase in most current devices) that reacts with glucose in the interstitial fluid and generates a small electrical signal proportional to the glucose concentration. A transmitter attached to the sensor on the skin’s surface picks up that signal and sends it wirelessly to a receiver, which is typically a smartphone app or a dedicated handheld device. Most sensors take a reading every one to five minutes, producing up to 288 data points per day.

Because the sensor sits in interstitial fluid rather than directly in your bloodstream, there is a built-in time lag. Glucose has to travel from the blood vessels into the surrounding tissue before the sensor can detect it. Research measuring this transit in healthy adults found the physiological delay is roughly five to six minutes on average.¹ In practice, the lag can stretch a bit longer depending on how rapidly your glucose is changing. When levels are spiking or dropping fast, the CGM reading can trail your actual blood glucose by several minutes, which is why many devices display trend arrows showing whether glucose is rising, falling, or holding steady.

Current consumer CGMs are transcutaneous, meaning the sensor pierces the skin from the outside. Most last 10 to 14 days before needing replacement. Implantable sensors that are placed fully under the skin by a clinician have also been developed, with some lasting up to 180 days.² Researchers are also exploring fully noninvasive approaches that would measure glucose through sweat, tears, or saliva, though none of these have reached the accuracy or reliability of today’s subcutaneous sensors.³

How Accurate Are CGMs?

The standard yardstick for CGM accuracy is something called the mean absolute relative difference, or MARD. In simple terms, MARD tells you the average percentage by which the sensor’s reading differs from a reference blood glucose measurement. A lower MARD means the sensor is closer to the “true” number. Modern CGMs generally report MARD values in the range of about 9 to 14 percent, and the newest flagship sensors push below 9 percent. For context, if your blood glucose is 100 mg/dL and the sensor has a 10 percent MARD, a typical reading might fall somewhere between 90 and 110.

That said, MARD has real limitations. Multiple papers have pointed out that published MARD values are heavily shaped by how the accuracy study was designed, including how many readings were taken, at what glucose ranges, and under what conditions.⁴ One recent critical review went further, arguing that MARD has limited clinical relevance in the current era of sensor technology because its calculation can be skewed so easily by study design choices.⁵ Comparing the MARD numbers printed on two different device boxes, in other words, does not necessarily tell you which sensor will be more accurate on your arm.

Some practical factors also affect accuracy day to day. Sensors tend to be less reliable in their first 12 to 24 hours as the tissue around the filament settles, a period sometimes called the “warm-up.” Readings can also drift during the sensor’s final days. Pressure on the sensor site, such as sleeping on the arm where it is placed, can temporarily compress the tissue and produce falsely low readings. Older-generation sensors were also sensitive to acetaminophen (Tylenol), which could inflate glucose readings, though newer devices have largely addressed this interference.⁶ When in doubt, most manufacturers still recommend confirming a CGM reading with a fingerstick before making treatment decisions, especially if the number does not match how you feel.

What a CGM Tells You That Fingersticks Cannot

A fingerstick gives you a single snapshot: your blood sugar at one moment. A CGM gives you a continuous movie. That difference matters because glucose is always moving, and many of the patterns that influence long-term health happen between fingersticks, overnight, or in the hours after meals. From a CGM’s continuous data stream, several clinically useful metrics emerge.

The most important is time in range, or TIR, which measures the percentage of the day your glucose stays within a target window, typically 70 to 180 mg/dL for most people with diabetes. An international consensus panel recommended that adults with type 1 or type 2 diabetes aim for a TIR above 70 percent, spend less than 4 percent of the day below 70 mg/dL, and less than 1 percent below 54 mg/dL.⁷ TIR is not just a convenience metric. Analysis of data from a landmark diabetes trial found that every 10 percent drop in TIR was associated with a 64 percent higher rate of retinopathy progression and a 40 percent higher rate of kidney-related complications.⁸

Beyond TIR, CGMs also reveal glucose variability, the magnitude and frequency of swings throughout the day. Two people can have identical average glucose levels but very different variability, and emerging research suggests that large, frequent spikes and dips may carry their own risks independent of the average. CGM data also captures overnight patterns, dawn-effect rises, and delayed responses to specific foods in a way that no number of fingersticks reasonably could.

Who Benefits from a CGM

The clearest and longest-established case is for people with type 1 diabetes. Because their bodies produce no insulin at all, managing glucose is a constant balancing act. CGM allows tracking of dynamic glucose fluctuations and trends over time, enabling better therapy adjustments and helping prevent dangerous lows.⁹ A meta-analysis of 21 randomized trials found that CGM lowered HbA1c (a measure of average blood sugar over about three months) compared with traditional fingerstick monitoring, with especially large improvements in people whose starting HbA1c was above 8 percent.¹⁰

For older adults with type 1 diabetes, who face particular dangers from hypoglycemia, a randomized trial found that CGM cut time spent in low glucose by about half an hour per day compared with fingerstick monitoring alone, while also improving HbA1c.¹¹ For someone prone to overnight lows they might sleep through, that difference is not trivial.

Type 2 Diabetes, Including Those Not on Insulin

The evidence for CGM in type 2 diabetes has grown rapidly. Even among people with type 2 diabetes who do not use insulin, a meta-analysis of randomized trials found that CGM was associated with a meaningful drop in HbA1c, roughly a third of a percentage point, along with about nine percentage points more time spent in range and eight percentage points less time spent above range.¹² Those numbers may sound modest, but over months and years they translate to meaningfully better glucose control.

One trial of CGM in people with type 2 diabetes not on insulin reported benefits that went beyond glucose. Participants saw improvements in BMI, blood pressure, triglycerides, and total cholesterol, along with a lower predicted ten-year cardiovascular risk. Their average glucose dropped from 184 to 147 mg/dL, and time in range jumped from about 58 percent to 83 percent.¹³ The likely explanation is that real-time glucose feedback motivated better food choices, more physical activity, and more aggressive medication adjustments by clinicians.

Gestational Diabetes

Pregnancy adds another dimension to glucose management because high maternal blood sugar increases the risk of complications for both mother and baby. A systematic review of trials in women with gestational diabetes found that CGM use was linked to lower HbA1c at the end of pregnancy, less gestational weight gain, and lower birth weight in newborns.¹⁴ A large recent trial, the GRACE trial, found that real-time CGM use in women with gestational diabetes reduced the rate of large-for-gestational-age births from 10 percent in the fingerstick monitoring group to 4 percent in the CGM group.¹⁵ Another trial found lower rates of unscheduled cesarean sections, preterm deliveries, and neonatal intensive care admissions in the CGM group.¹⁶

CGMs for People Without Diabetes

Over the past few years, a growing number of people without diabetes have started wearing CGMs as a wellness or performance tool. Companies market subscriptions that pair a CGM with an app to help users see how specific meals, exercise routines, and sleep patterns affect their glucose. The appeal is understandable: you get real-time biofeedback that was previously invisible. A review of postprandial glucose behavior in non-diabetic individuals noted that while post-meal spikes are well understood in diabetes, their significance in healthy people is less defined.¹⁷

What does “normal” CGM data actually look like? A multicenter study of healthy, non-diabetic participants found that average 24-hour glucose was about 99 mg/dL, with roughly 96 percent of the day spent between 70 and 140 mg/dL. Time above 140 was a median of about 30 minutes per day, and time below 70 was about 15 minutes per day. Adults over 60 tended to run slightly higher, averaging about 104 mg/dL.¹⁸ A larger reference dataset from over 7,000 non-diabetic individuals aged 40 to 70 has also been published, providing benchmarks for what typical glucose patterns look like across a range of ages and body compositions.¹⁹

Endurance athletes represent another niche audience. They use CGMs to monitor fueling during training and racing, trying to keep glucose in a window that avoids bonking (running out of glycogen) without overloading on gels. The evidence that CGM-guided fueling actually improves athletic performance is still thin, but the technology does offer athletes a way to observe how their body responds to different carbohydrate loads and timing strategies in real time.

For the average healthy person, the honest question is whether the data changes anything. Seeing that a bagel spikes your glucose more than an omelet is interesting the first few times, but for someone whose body is regulating glucose normally, those spikes resolve on their own. There is currently no strong evidence that wearing a CGM improves long-term health outcomes in people without diabetes or prediabetes. It can, however, be a useful educational window: a few weeks of CGM wear can reveal whether you have unrecognized glucose patterns that warrant a conversation with your doctor, especially if you have a family history of diabetes or other metabolic risk factors.

The Closed-Loop Connection

One of the most consequential uses of CGM technology is as the sensing half of an automated insulin delivery system, often called a closed-loop system or artificial pancreas. These systems pair a CGM with an insulin pump and a control algorithm. The algorithm reads the CGM data in real time and adjusts insulin delivery automatically, increasing it when glucose is climbing and reducing or suspending it when glucose is falling.²⁰ The person still has to change the sensor, refill the pump, and enter carbohydrate estimates for meals, but the minute-to-minute insulin dosing is handled by the algorithm.

These systems have been shown to improve both glucose control and quality of life, partially relieving people with type 1 diabetes of the constant mental burden of managing their insulin.²¹ Several commercially available systems are now in use, and the technology continues to evolve.²² Without CGM, none of this would be possible. The sensor is the system’s eyes.

Living with a CGM Day to Day

Wearing a CGM is not quite as effortless as the marketing images suggest. The sensor applicator uses a spring-loaded needle to insert the filament under the skin, a process most people describe as a brief pinch rather than real pain. Once in place, the small adhesive patch and transmitter stay on for the sensor’s lifespan, through showers, exercise, and sleep. But the adhesive can cause irritation, and for some people it causes genuine allergic contact dermatitis. The most commonly identified culprit is a chemical called isobornyl acrylate (IBOA) used in certain sensor adhesives, which has sensitized a large number of published cases over the past several years.²³ Other allergens identified include epoxy resin, colophonium, and nickel, and the repetitive occlusion and moisture under the patch can worsen skin reactions even without a true allergy.²⁴ Barrier films and skin-prep wipes can help, but persistent reactions sometimes force people to switch brands or sensors.

The other lived-reality issue is alarm fatigue. CGMs can be set to alert you when glucose goes too high or too low, and in theory that sounds perfect. In practice, frequent alerts, especially overnight, can become exhausting. A review of this phenomenon in children and adolescents with type 1 diabetes found that alarm fatigue can negatively affect daily life for both patients and caregivers, and is a documented reason some people abandon CGM use altogether.²⁵ Adjusting alarm thresholds and using “do not disturb” windows strategically can help, but it remains a balancing act between safety and sanity.

Access, Cost, and Disparities

CGM technology has improved dramatically over the past decade, and retail prices have dropped as competition has increased. In the United States, most private insurance plans and Medicare now cover CGM for people with diabetes who use insulin, though coverage for people with type 2 diabetes who are not on insulin is more variable and often requires prior authorization. Out of pocket, sensor costs typically run in the range of $75 to $150 per month without insurance, depending on the brand and pharmacy.

Even where CGMs are clinically justified and cost-effective, uptake is uneven. A review of the evidence noted persistent disparities by age, race and ethnicity, insurance coverage, and care setting, with certain groups far less likely to be offered or to use CGM despite standing to benefit from it.²⁶ This is not purely a cost issue. Provider awareness, patient education, and structural factors in healthcare delivery all play a role.

Data Privacy and Interoperability

A CGM generates an enormous amount of personal health data, and most of it flows through manufacturer-controlled cloud platforms and smartphone apps. An analysis of the privacy landscape found that patients’ ability to control how their CGM information is collected, stored, and used is virtually nonexistent, and that current data protections are not robust enough to address the risks.²⁷ If you are using a wellness-oriented CGM subscription, this is worth considering: your glucose responses to meals, your sleep patterns, and your activity levels are being captured continuously.

On a related front, getting CGM data to talk to other systems, like your electronic health record or a different brand’s insulin pump, remains frustratingly difficult. A commentary on the interoperability landscape noted that standardized data-sharing protocols are still lacking, and that regulatory and liability questions around device interoperability add further complexity.²⁸ For the person wearing the device, this can mean manually exporting reports, logging into multiple portals, and occasionally finding that their CGM data simply does not flow where it would be most useful. The technology on the sensor side has leapt ahead; the data infrastructure surrounding it has not quite kept pace.