DOI: 10.2337/db23-168-or ISSN: 0012-1797

168-OR: Marked Variability of Starting Insulin Algorithms for Noncritical Illness—A Survey of 32 Academic Hospitals in the United States

  • Endocrinology, Diabetes and Metabolism
  • Internal Medicine

Glycemic control immediately after hospitalization is difficult, with blood glucose within 24 hours after admission often above recommended targets. Further, it takes time to adjust and determine individual insulin requirements. Endocrine Society guidelines suggest starting scheduled insulin therapy at 0.2–0.5 units/kg body weight as the total daily dose, but there has been no rigorous study to support this guideline and other institutional algorithms. To understand the variability of current practice, we conducted a survey of starting insulin algorithms for non-critically ill patients among top-ranking hospitals and endocrinology programs in the US.

Among the 32 hospitals surveyed, replies were received from 25. Of these, 19 provided their algorithms, 5 stated they did not have such, and one used a proprietary (commercial) algorithm. The figure details the number of algorithms specifying basal/nutritional insulin regimens versus only correction insulin, approaches to calculating the basal/nutritional dose, and guides to choosing correction algorithms.

In summary, amongst US academic hospitals there is great variability in methods for determining starting insulin dosing on admission for non-critically ill patients. This inconsistency suggests that a clinical trial is required to determine the best approach(es) to estimating initial insulin requirements.


H.Chiang: None. S.E.Kahn: Advisory Panel; Anji Pharmaceuticals, Bayer Inc., Boehringer Ingelheim Inc., Eli Lilly and Company, Merck & Co., Inc., Other Relationship; Novo Nordisk. I.B.Hirsch: Consultant; Abbott Diabetes, Lifecare, Inc., Hagar, Research Support; Beta Bionics, Inc., Insulet Corporation, Dexcom, Inc.


National Institutes of Health (5T32DK007247-45)

More from our Archive