Glycemic control immediately after hospital admission is difficult. This study aimed to develop an algorithm-based approach to initiate insulin therapy on admission.

Patients with history of diabetes mellitus admitted at UC Davis medical center, with any blood glucose (BG) value ≥ 180 mg/dL, or who received any insulin within the first 24 h of hospitalization were selected for a retrospective chart review.

Total of 315 patient records were studied. Patients prescribed insulin prior to admission had higher 24-hour average BG and higher corrected total daily dose of insulin (CxTDD), compared with the patients who were not prescribed insulin prior to admission. For the patients not receiving home insulin and not given new glucocorticoids, first BG upon presentation correlated with the risk of first 24-hour average BG > 180 mg/dL. Factors associated with CxTDD were first BG, weight, oral intake, and glucocorticoid dose. Home insulin daily dose, opiate/intravenous pain medication and systemic inflammatory response syndrome were associated with CxTDD only in the patients receiving home insulin.

A subgroup of patients can be given correction insulin as a sole initial treatment on admission. For patients requiring basal-bolus insulin, several factors associated with the initial insulin requirements are identified.