Healthcare professionals must better understand the evolving algorithms behind automated insulin delivery systems to optimise personalised diabetes care, experts say, as device options and patient needs become more varied.
Healthcare professionals need a firmer grasp of automated insulin delivery algorithms as diabetes technology becomes more varied and more patient-specific, Erin Henry, a nurse practitioner at North Atlanta Endocrinology and Diabetes in Georgia, said during a presentation at the Association of Diabetes Care & Education Specialists annual meeting. Speaking to Healio, Henry said newer systems have made algorithm knowledge more important because patients now choose devices for the way they operate, not just for their appearance or hardware.
Henry drew a clear line between total daily dose systems and adaptive basal systems. In the first group, the pump effectively sets the basal rate itself, so changing the programmed basal schedule may have little or no effect. In the second, the algorithm responds to an existing basal pattern, so clinicians can still influence glucose control by adjusting that background delivery. She said the distinction matters in everyday care, particularly when treating dawn phenomenon or reviewing why a change has or has not helped.
Her comments align with broader guidance in the diabetes literature. A review in PubMed and a consensus paper on open-source automated insulin delivery both stress that clinicians should understand how different systems calculate insulin delivery if they are to support shared decision-making and troubleshoot problems effectively. A separate overview in Diabetes Spectrum also notes that algorithms differ in design, with model predictive control, proportional-integral-derivative and fuzzy logic among the main approaches used in current systems.
Henry said there is no single best algorithm for every patient. Preferences around tubing, button-pushing, continuous glucose monitor compatibility and other device features can all shape the right choice. She also pointed to practical extras such as temporary basal settings, exercise modes and extended boluses, saying these features can be crucial in real-world situations, from long car journeys to hiking.
The approach also changes depending on the person’s type of diabetes. Henry said she offers pump therapy to many people using multiple daily injections, but tends to look for larger reservoirs in type 2 diabetes because insulin needs are often higher. In type 1 diabetes, where insulin sensitivity is usually greater, she said dose changes should generally be smaller, and site failure education is especially important because ketoacidosis can develop faster when there is no insulin reserve.
She also warned that starting a glucagon-like peptide-1 drug while using automated insulin delivery calls for rapid insulin reassessment. Henry said insulin requirements can fall sharply, by about 30 per cent in type 1 diabetes and roughly 20 per cent in type 2 diabetes, and early dose reductions are often needed. For total daily dose systems, she said the main levers are the correction factor and carbohydrate-to-insulin ratio, and patients should have rescue medicine available because of the risk of severe hypoglycaemia.
When it comes to follow-up, Henry said well-controlled patients can usually be reviewed at routine three-month visits, while those with poor control may need attention every four to six weeks. Pregnant patients, she added, may need weekly review. The most useful measures, in her view, are time spent in automated mode, how often and how patients bolus, total daily dose and the split between basal and bolus insulin. Reviewing at least a week of daily reports alongside the summary view, she said, helps pinpoint whether the problem is under-bolusing, excessive basal delivery or another issue altogether.
Disclaimer: This content is for informational purposes only and is not intended to be a substitute for professional medical judgment, advice, diagnosis, or treatment.
