Large differences observed among insulin algorithms at US academic hospitals
Click Here to Manage Email Alerts
Key takeaways:
- Of 20 hospitals providing insulin algorithms, 60% stated which patients should initiate basal or nutritional insulin.
- Correction dose insulin recommendations were included in 25% of algorithms.
Algorithms for initiating insulin among noncritically ill adults varied widely across 20 U.S. academic hospitals, according to a brief report published in Diabetes Technology & Therapeutics.
“We keep hearing from patients how poorly their diabetes is controlled when they are admitted to the hospital,” Irl B. Hirsch, MD, professor of medicine at the University of Washington School of Medicine in Seattle, told Healio. “We thought we would take a closer look at this, in addition to the fact this is an area we hope to do more research on. The survey focused on type 2 diabetes for several reasons, including the fact we didn’t want to complicate the answers, with pump therapy becoming so popular now with type 1 diabetes.”
Hirsch and colleagues sent emails to the leaders of inpatient glycemic services at the top 20 hospitals and top 30 endocrinology programs in the U.S. News & World Report’s 2022 Best Hospitals ranking. The emails asked for each hospital’s algorithm for starting insulin therapy among noncritically ill patients. Researchers used the provided algorithms to examine which patient should start basal or nutritional insulin, which should start correction dose insulin, how an initial dose should be calculated, how to choose the correction dose algorithm and whether algorithms were included in electronic medical order systems or as separate written protocols.
Variability among algorithms
Twenty hospitals provided insulin algorithms. Of those, 60% had instructions specifying which patients should start basal or nutritional insulin. Of those hospitals, nine said people with type 1 diabetes should always receive basal insulin, two said basal or nutritional insulin should be administered to all or most people who need it, and one said people using basal insulin at home should also receive it in the hospital. Recommendations for basal or nutritional insulin in type 2 diabetes were inconsistent across the hospitals.
Only 25% of hospitals had instructions on which patients should receive correction dose insulin. Three recommended correction dose insulin for people without a history of diabetes, whereas two said people with well-controlled diabetes using noninsulin medication could receive correction dose insulin.
For calculating insulin dose, 65% of hospitals recommended both weight-based and home dose-based calculations, 25% suggested only a weight-based calculation and 10% recommended only a home dose-based calculation.
Most hospitals had high, medium and low levels for correction insulin dosing. There were three hospitals with two levels and two with four levels. Forty percent of hospitals considered daily and basal insulin when choosing a correction algorithm, 30% considered daily and basal insulin dose plus BMI or body habitus and 20% considered BMI or body habitus alone.
Of the participating hospitals, eight had their algorithm in their electronic medical order system, 11 had it included as a separate written protocol, and one included the algorithm in teaching materials.
There were 10 hospitals that mentioned oral glucose-lowering medications in their algorithm. Of those hospitals, seven recommended discontinuing medication at admission. Of six hospitals that mentioned GLP-1 receptor agonists, three recommended discontinuing them on admission and three recommended withholding them on the day of surgery.
Evidence-based algorithms needed
Hirsch said the findings were not surprising due to the lack of evidence on insulin algorithms in the hospitals. One of the biggest challenges Hirsch said hospitals face with algorithms is how people with diabetes can have large differences in their baseline glycemic control, baseline insulin deficiency and the stress of their illness, even if their glucose levels are similar.
“The problem is patients are so heterogeneous,” Hirsch said. “It seems clear to me that we can’t put everyone on the same insulin regimen, given the variability and complexity of our patients.”
The brief report’s findings revealed the need for more standardized algorithms in hospitals, according to Hirsch.
“We need to develop evidenced-based algorithms that uses all of the factors (weight/BMI, home glucose control, use of steroids, severity of illness, home diabetes meds, etc) to come up with initial and subsequent insulin dosing,” Hirsch said. “Ideally, this needs to be developed so that it can be implemented in any hospital. The other thing to keep in mind, especially with the introduction of over-the-counter continuous glucose monitors, is more patients will be admitted already wearing these sensors, so each hospital will need to decide how they will use these sensors until we learn from FDA if they will be approved for inpatient use.”
For more information:
Irl B. Hirsch, MD, can be reached at ihirsch@uw.edu.
Collapse
Click Here to Manage Email Alerts