Through standardization and automation, insulin error rates were reduced and insulin timeliness was improved in a pediatric hospital; the insulin error rate reduction was sustainable.


Adverse drug events (ADEs) during hospitalization are common. Insulin-related events, specifically, are frequent and preventable. At a tertiary children’s hospital, we sought to reduce insulin-related ADEs by decreasing the median event rate of hyper- and hypoglycemia over a 12-month period.


Using Lean 6 σ methodology, we instituted a house-wide process change from a single-order ordering process to a pro re nata (PRN) standing order process. The standardized process included parameters for administration and intervention, enabling physician and nursing providers to practice at top of licensure. Automated technology during dose calculation promoted patient safety during dual verification processes. Control charts tracked rates of insulin-related ADEs, defined as hyperglycemia (glucose level >250 mg/dL) or hypoglycemia (glucose level <65 mg/dL). Events were standardized according to use rates of insulin on each nursing unit. The rates of appropriately timed insulin doses (within 30 minutes of a blood sugar check) were assessed.


Baseline median house-wide frequencies of hyperglycemic and hypoglycemic episodes were 55 and 6.9 events (per 100 rapid-acting insulin days), respectively. The median time to insulin administration was 32 minutes. The implementation of the PRN process reduced the median frequencies of hyperglycemic and hypoglycemic episodes to 45 and 3.8 events, respectively. The median time to insulin administration decreased to 18 minutes.


A PRN ordering process and education decreased insulin-associated ADEs and the time to insulin dosing compared with single-entry processes. Engaging bedside providers was instrumental in reducing insulin-related ADEs. Strategies that decrease the time from patient assessment to drug administration should be studied for other high-risk drugs.

You do not currently have access to this content.