When your blood sugar is even, the glucose sensor value deviates from the actual blood sugar level by approximately ±10%. The magnitude roughly correlates with self-performed blood sugar tests in general. Therefore, the glucose sensor reading can be used as a basis of the bolus insulin dose. That said, if the reading materially deviates in either direction considering your condition and the circumstances, you should validate the result with a finger-prick test.
It is also good to bear in mind that the glucose level of tissue fluid indicated by the sensor does not precisely correspond with the glucose level of blood taken from a capillary in the fingertip. When your blood sugar rises or drops, there is a delay of roughly 5–10 minutes before the sensor reading catches up.
For adults, the insulin-to-carbohydrate ratio is normally 0.5–2 units per 10 g of carbohydrate, i.e. if you consume 10 grams of carbohydrates, you need to take 0.5–2 units of rapid-acting insulin. You can estimate the correct ratio, duration of action of the rapid-acting insulin and correct timing of the injection based on the post-meal glucose graph. The precondition is that the amount of carbohydrates per meal, the bolus insulin dose measured out on the basis of it, any potential correction doses as well as physical exercise are stored in the system several days in a row:
If, 3–4 hours after the start of the meal, the glucose level is higher than to begin with, the rapid-acting insulin dose is probably too small.
If the glucose level is below 10 mmol/l 1–2 hours after the start of the meal but lower than to begin with 3–4 hours after the start of the meal, the rapid-acting insulin dose is probably too big.
If the glucose level rises too high 1–2 hours after the start of the meal but is too low 3–4 hours later, the effect of the rapid-acting insulin is delayed. Most probably, you should time the injection earlier, for example, 10–20 minutes before the start of the meal.