Insulin Clinics / Monitoring and safety

Monitoring blood glucose

A glucose number on its own is not the goal. The goal is patterns: where you sit most of the time, where you spike, and where you go low.

Last updated: October 2026 · Based on ADA Standards of Care, FDA prescribing information, and NIDDK · General information, not medical advice — see our editorial policy

Fingerstick meters

A blood glucose meter measures glucose in a drop of blood from a fingertip. The hardware is mature and inexpensive. Test strips are the consumable cost.

How to get an accurate reading:

• Wash your hands with soap and water and dry them. Residue from food (especially fruit) will give falsely high readings; alcohol wipes can give falsely low ones if not fully dry.
• Use the side of the fingertip rather than the pad — fewer nerves there.
• Rotate fingers; do not always use the same one.
• Use a fresh strip from a sealed vial. Strips that have been exposed to air for weeks can read inaccurately.
• Some meters allow alternate-site testing (forearm, palm). These sites lag behind fingertips when glucose is changing fast, so use fingertips when symptoms suggest a low or after a meal.

Meters sold today must meet accuracy standards (roughly within 15 percent of a lab value for most readings), which is good enough for treatment decisions. If a reading does not match how you feel, recheck before acting on it.

Continuous glucose monitors (CGMs)

A CGM is a small sensor — usually worn on the back of the upper arm or the abdomen, depending on the device — with a tiny filament under the skin that measures glucose in interstitial fluid (the fluid between cells). It reads continuously and sends a number to a phone, smartwatch, receiver, or insulin pump every 1–5 minutes.

How CGMs work

Interstitial glucose lags behind blood glucose by about 5–15 minutes, depending on how fast glucose is changing. When glucose is steady, the two values are very close. When glucose is rising or falling fast (after a meal, after rapid insulin, during exercise), the CGM reads slightly behind the actual blood value. Modern CGMs use software to predict where glucose is heading and partially correct for this lag.

Major systems available now include the Dexcom G7 and G7 15 Day, Abbott FreeStyle Libre 3 Plus and Libre 2 Plus (which replaced the original Libre 2 and Libre 3 sensors in the U.S. in 2025), the Medtronic Guardian 4 and Simplera sensors, and the implantable Eversense 365. Sensors worn on the skin last about 7 to 15 days depending on the device; the implantable sensor lasts up to a year. Most current sensors do not require fingerstick calibration in normal use.

What CGMs add

• Trend arrows — knowing whether glucose is steady, rising, or falling — guides whether to act now or wait.
• Alarms — for highs, lows, and predicted lows. Especially valuable for overnight safety and for people with hypoglycemia unawareness.
• Time in range data — much more clinically useful than a single A1C number.
• Shareable data — partners, parents, and care teams can see real-time and historical readings (with permission).

Access

CGM coverage in the United States has expanded significantly. As of 2026, Medicare covers CGMs for people who use any insulin, and for some people not on insulin who have problematic hypoglycemia; many commercial plans cover CGMs for people on insulin, though rules vary by plan. Some over-the-counter sensors are sold without a prescription but are intended for people who do not use insulin — if you take insulin, use a prescription CGM that is cleared for insulin dosing decisions.

A1C: what it actually measures

A1C (also called HbA1c or glycated hemoglobin) is a blood test that measures the percentage of hemoglobin molecules in red blood cells that have glucose stuck to them. Because red blood cells live about 3 months, A1C reflects average glucose over roughly the previous 8–12 weeks, with more weight on the most recent month.

What A1C does and does not show

A1C is excellent at summarizing whether you are running high or low on average over months. It is the strongest predictor of long-term complication risk, and most clinical guidelines target an A1C below 7 percent for many adults — sometimes higher (7.5–8 percent) for older adults or people at high risk of severe hypoglycemia, sometimes lower for younger adults or in pregnancy.

What A1C does not show:

• Variability. Two people with an A1C of 7 percent can have very different lives — one steady around 154, one swinging from 50 to 250.
• Hypoglycemia. Frequent lows lower the average and can mask poor control.
• Conditions that affect red blood cells. Sickle cell trait or disease, recent transfusion, iron-deficiency anemia, hemolytic anemia, kidney disease on erythropoietin, and pregnancy can all distort A1C. In those situations, CGM-derived metrics are more reliable.

Time in range

Time in range (TIR) is the percentage of time CGM readings sit within a target glucose range — for most adults, 70 to 180 mg/dL (3.9 to 10.0 mmol/L).

Standard targets endorsed by the international consensus on time in range:

• Time in range (70–180 mg/dL): 70 percent or more for most adults with diabetes.
• Time below range (under 70): less than 4 percent.
• Time well below range (under 54): less than 1 percent.
• Time above range (over 180): less than 25 percent.
• Time well above range (over 250): less than 5 percent.

For older adults at higher risk of low blood sugar, slightly looser targets are common (more than 50 percent in range, less than 1 percent below 70, and less than 10 percent above 250). For pregnancy with type 1 diabetes, targets are tighter (70 percent in 63–140 mg/dL, less than 4 percent below 63).

Time in range is increasingly used by clinicians because it captures both the average and the variability — a 70 percent TIR usually means a steadier life than the same A1C with worse range data. Most CGM platforms generate a standard report (the Ambulatory Glucose Profile, or AGP) that shows TIR alongside a typical 24-hour pattern.

How often to check

The right number depends on the regimen:

Type 2 diabetes on basal insulin only

• Fasting glucose daily (or most days), to guide basal dose.
• Occasional pre- and post-meal pairs to check if mealtime insulin will eventually be needed.
• Some people use a 14-day CGM session every few months instead of fingersticks.

Type 2 diabetes on basal-bolus, or any type 1 diabetes

• If using fingersticks: often 6–10 times a day — before meals and snacks, at bedtime, before, during, and after exercise, any time symptoms suggest a low, after treating a low, and before driving.
• If using a CGM: continuous, with calibration fingersticks only as the device requires (most no longer require any).

Special times to check more

• When sick, more often — every 2–4 hours, plus ketones if glucose is over 250 mg/dL (or the level in your sick-day plan).
• Before, during, and after exercise.
• Before driving, especially if you have had recent lows.
• When trying a new food, a new schedule, or a new dose.
• Overnight, when investigating a possible nighttime low or basal dose problem.

What patterns to look for

The most useful question is rarely "is this number good?" — it is "is this happening repeatedly?"

• Fasting highs most mornings → basal dose may be too low, or evening eating patterns are pushing morning glucose up.
• Fasting lows or near-lows several mornings a week → basal dose is too high, especially with older long-acting insulins.
• Post-breakfast spikes that come down by lunch → carb ratio at breakfast is off (breakfast often needs more insulin per gram of carbohydrate than other meals).
• Mid-afternoon lows → lunch dose may be too high or absorbing too fast.
• Bedtime highs that drop overnight → dinner bolus timing or basal pattern.
• "Dawn phenomenon" → glucose rising in the early morning hours before waking due to natural hormone shifts; sometimes addressed by adjusting basal timing or, on a pump, programming a higher rate before dawn.

Most patterns become visible only when you look at 7–14 days at once. CGM reports do this automatically; if using a meter, an honest week of logging is more useful than scattered checks over a month.

Frequently asked questions