Computerized intravenous insulin algorithm virtually eliminates severe hypoglycemia across 19 hospitals, analysis suggests
Presenter: Joseph Aloi, MD, Advocate Health, Wake Forest Baptist Hospital, Winston-Salem, NC
Virtual elimination of severe hypoglycemia utilizing a computerized insulin dosing algorithm for intravenous insulin infusions. Presented April 23, 2026.
A computerized intravenous insulin dosing algorithm was associated with a 98% reduction in severe hypoglycemia compared with a standard paper protocol in hospitalized patients receiving IV insulin, according to a 3-year analysis presented at AACE 2026.
The retrospective study evaluated 17,838 patients treated according to Endotool IV (ETIV; Monarch Medical Technologies) insulin-dosing software at 19 hospitals within the Advocate Health system between January 2023 and December 2025, including a prespecified subgroup analysis of patients with severe renal insufficiency (estimated glomerular filtration rate [eGFR] < 15 mL/min/1.73 m2).
“This large database really provides providers with a lot of confidence in the fidelity of the system and the consistency of data across different patient populations and different hospital settings,” Dr. Aloi said in an interview.
Severe hypoglycemia in hospitalized insulin-treated patients is rare but can lead to death, and hospitals now face mandatory reporting of severe hypoglycemia to the Centers for Medicare and Medicaid Services, Dr. Aloi said.
Paper-based insulin protocols, many now transcribed into the electronic medical record, titrate infusions primarily from the rate of change in blood glucose, Dr. Aloi said. By contrast, computerized algorithms such as ETIV pull data directly from the electronic medical record to estimate insulin sensitivity in near real time, calculate a drip rate, and factor in the estimated residual insulin in the patient.
This approach “holds the promise to virtually eliminate severe hypoglycemia, a severe consequence of giving patients insulin, particularly in this model of intravenous insulin,” Dr. Aloi said in the interview.
Hypoglycemia outcomes
In the study, severe hypoglycemia was defined as a blood glucose value below 40 mg/dL. All blood glucose readings obtained while patients received intravenous insulin were included, and outcomes were compared with a pre-ETIV cohort managed under a standard paper protocol.
Across the ETIV-treated population, 362,901 blood glucose values were obtained. Severe hypoglycemia occurred in 16 readings (0.005%), compared with 0.23% under the prior paper protocol, a 98% reduction, Dr. Aloi said. Among patients with diabetic ketoacidosis, the severe hypoglycemia rate was 0.007%.
Renal subgroup
In the renal subgroup, 677 patients with eGFR less than 15 mL/min/1.73 m2 had a severe hypoglycemia incidence of 0.007% (1 reading), similar to the 0.005% (7 readings) observed in 7,197 patients with eGFR greater than 60 mL/min/1.73 m2.
Dr. Aloi attributed the performance in patients with advanced renal disease to the algorithm’s estimate of residual insulin, noting that reduced kidney function prolongs insulin half-life, and that patients with end-stage kidney disease tend to have little glycogen reserve to buffer a falling glucose.
“I think that’s why the patients with end-stage kidney disease do particularly well with this system, because that’s a patient population where how long the insulin lasts can be very variable,” Dr. Aloi said.
ETIV recommends a specific prophylactic carbohydrate dose when the algorithm estimates that blood glucose will fall below 70 mg/dL. This predictive feature was credited with averting an estimated 2,213 episodes of hypoglycemia.
The incidence of severe hyperglycemia remained low at 0.17%, authors reported.
Consistency across academic and community sites
A key contribution of the current analysis is that the results span academic and community centers. Much of the existing literature on computerized insulin dosing has come from single academic centers, Dr. Aloi said, while similar work done in community hospitals has often not reached peer-reviewed journals.
The present analysis spans 19 hospitals across different community areas, including two major academic hospitals, all of which had similar hypoglycemia outcomes with ETIV: “There wasn’t a single hospital that really stood out. This was across the board,” Dr. Aloi said.
Based on their findings, authors concluded that the implementation of ETIV across these diverse hospital settings was associated with the near-elimination of severe hypoglycemia while maintaining control of hyperglycemia.
Moreover, they said, the findings suggest a need for more centers to consider the use of a computerized insulin dosing algorithm to improve glycemic outcomes.
A larger, dedicated analysis of the severe renal insufficiency subgroup is planned, according to Dr. Aloi.
Disclosures
Dr. Aloi had no disclosures relevant to the presentation.
References
Aloi J, Chidester P. Virtual elimination of severe hypoglycemia utilizing a computerized insulin dosing algorithm for intravenous insulin infusions. Endocrine Pract 2026; 32(suppl). https://doi.org/10.1016/j.eprac.2026.03.035