Eight lessons on counting carbs added 2.5 hours a day in glucose range. The study had 30 finishers — and no comparison group

Carbohydrate counting asks a person using mealtime insulin to estimate the carbohydrate in food and match the dose to it. That sounds like arithmetic, but doing it across home cooking, restaurant portions, labels, snacks and exercise takes more than a conversion table. A study published on 31 October 2026 tested a structured, dietitian-led course in adults with type 1 diabetes who already used continuous glucose monitors.

The programme ran for 12 weeks and had eight visits arranged in three levels. Researchers enrolled 38 adults; 30 completed the follow-up and are the people in the analysis. They compared each finisher's most recent 14 days of sensor readings, HbA1c, insulin dose, weight and food intake with the same measures before the course. There was no usual-care or wait-list group.

Time in range — the share of sensor readings from 70 to 180 mg/dL — rose from an average 62.5% to 73.0%. The mean change was 10.4 percentage points, with a 95% confidence interval from 7.1 to 13.8. Spread over a full day, that is about two and a half extra hours in range. Time above 180 mg/dL fell, while time below range did not increase. HbA1c also moved from 7.71% to 7.08%, a drop of 0.63 percentage points.

The encouraging part is what did not rise. The study detected no increase in body weight or total daily insulin, and the extra time in range did not appear to come from trading high readings for more low ones. Glucose variability, measured by the coefficient of variation, fell from 35.7% to 33.6%. But the group already averaged below the recommended 36% variability threshold at the start, and that change stopped being statistically clear in a sensitivity analysis limited to the 22 people who met the study's body-mass-index eligibility rule.

There are two reasons to treat this as a promising practice result rather than proof. First, everyone received the course. A comparison group is needed to separate the lessons from closer attention, repeated measurement, changes people made on their own or regression towards a better average. Second, the researchers designated time in range as the primary outcome only after data collection and analysis. A primary outcome chosen in advance protects against highlighting whichever measure happens to improve most.

The practical message is not to calculate insulin from an online formula. Carb ratios, correction doses, exercise and low-glucose risk are individual clinical decisions. It is that repeated education with a dietitian may add something even for people who already see their glucose continuously. The next useful test is a larger controlled trial that keeps everyone in the analysis and declares its main sensor outcome before the first participant starts.