Hospitalized patients with complex dietary restrictions often suffer from hyperglycemia, a condition in which blood glucose levels are high. This occurs in approximately one-quarter to one-half of these patients and can lead to serious complications, especially in those who also have diabetes. Managing blood glucose levels in the hospital is challenging for a variety of reasons, including inconsistent caloric intake, changes in kidney and liver function, surgery, infections and limitations in the need for laborious blood glucose monitoring and insulin dosing.
To address these challenges, UCSF endocrinologist Robert J. Rushakoff, MD, MSc, and his team of nurses, pharmacists, physicians, and programmers developed and implemented the Self-Tuning Subcutaneous Insulin Algorithm (SQIA) for use at three UCSF San Francisco hospitals. A new study evaluating SQIA shows lower incidence of severe hyperglycemia and hypoglycemia (low blood sugar) compared with traditional insulin administration.
Rushakoff will report blood glucose control data from the study's first three years at the American Diabetes Association's annual scientific sessions in Orlando on June 22.
SQIA is an integrated calculator built into the Medication Administration Record (MAR) of the electronic medical record. During the first three years after full implementation (September 2020 to September 2023), SQIA was used for thousands of hospitalized patients with nutritional restrictions in one of three categories: no feeding by mouth (NPO), continuous tube feeding (TF), or total parenteral nutrition (TPN).
When physicians ordered rapid-acting insulin for patients who fell into one of these nutritional categories, they were given the option to use SQIA or proceed with conventional insulin (CI) dosing orders. With SQIA, physicians only entered the initial starting insulin dose, which was then automatically adjusted by the algorithm. CI, on the other hand, required physicians to manually enter new insulin doses as needed.
When it's time to administer insulin, the nurse enters the patient's current blood glucose level into the MAR, and SQIA automatically calculates the new insulin dose using the previous insulin dose, previous blood glucose level, and current blood glucose level. Based on ongoing monitoring and feedback from nurses, pharmacists, and physicians, the algorithm and calculator interface have been adjusted to improve titration of the appropriate insulin dose for patients.
Using this process, the researchers showed that SQIA reduced the number of insulin prescriptions doctors wrote for a given patient by more than 12-fold compared with CI administration.
Compared to conventional insulin regimens, use of SQIA reduced hyperglycemia rates and further reduced already low hospital-wide hypoglycemia rates., This increases physician efficiency by largely eliminating the need for physicians to create new orders or make adjustments in SQIA order sets.”
Robert J. Rashakoff, senior author, professor of medicine at UCSF and director of inpatient diabetes medicine at UCSF
SQIA increased insulin doses administered with NPO and TPN diets, reduced the incidence of severe hyperglycemia, and did not increase hypoglycemia, suggesting that physician-initiated CI orders may be undertreating patients. Furthermore, the incidence of severe hyperglycemia with SQIA gradually decreased over the study period, suggesting that continued development of SQIA may provide increasing benefit to patients over time.
“Our findings suggest that the typical insulin inertia observed in insulin dose adjustments in many centers could be overcome by automated algorithms such as SQIA, reducing physician workload,” said Rashakoff.
SQIA is now the primary method of prescribing insulin to hospitalized patients across UCSF hospitals and is chosen by physicians for approximately 80% of eligible hospitalized patients.
This initiative builds on previous innovations in inpatient diabetes management at UCSF. In 2013, the Virtual Glucose Management Service (vGMS) was conceived and implemented. Each morning, the vGMS automatically generates a report of all hospitalized patients with uncontrolled glucose levels. A diabetologist remotely reviews this report along with insulin glucose charts and enters insulin dosing recommendations into each patient's EMR. These recommendations are available to clinicians for review by 6:30 a.m. each day. Since implementing vGMS, UCSF has seen a 50% reduction in the number of hospitalized patients with daily hyperglycemia reports and consistently lower rates of hypoglycemia. Publications about the success of vGMS have led to medical centers around the world implementing local versions of vGMS.