The goal of this feasibility trial is to test whether a novel insulin dosing calculator can improve blood glucose management for patients with type 2 diabetes mellitus on hospital admission. The main questions it aims to answer are:

• Whether using this calculator leads to no excessive hypoglycemia?
• Whether using this calculator leads to a decrease of hyperglycemia?

Participants will receive initial insulin doses by the calculator on hospital admission, and the rest of care will follow standard of care. Participants will be compared with patients in the UW observational cohort who received initial insulin doses by standard of care.

Inclusion Criteria:

• Patients who are admitted to University of Washington Medical Centers
• With history of type 2 and/or steroid-induced diabetes mellitus
• Receiving insulin therapy prior to admission

Exclusion Criteria:

• Patients who are admitted to ICU or obstetrics ward
• Patients who are admitted for elective surgery or procedure
• Patients who present with diabetic ketone acidosis, hyperosmolar hyperglycemic state or require intravenous insulin infusion
• Patients who have no meal intake for 24 hours prior to admission, or planned nothing per oral (NPO) during the first 24 hours after admission
• Patients who report low appetite (25% or less) on admission or have a significantly decreased level of consciousness that study team does not think they are going to eat right after admission
• Patients who receive enteral feeding after admission
• Patients who develop severe acute kidney injury needing dialysis therapy

Data sharing will be considered only on a collaborative basis with PIs, after evaluation of the proposed study protocol and statistical analysis plan.

from hospital admission until 48 hours after admission

• Single-arm feasibility study which does not include a control arm
• Relatively small sample size to detect a rare outcome or adverse event
• Single center study