Background: Diabetic Ketoacidosis (DK)A at our center has historically been managed in the intensive care unit (ICU) utilizing low-dose insulin infusion. According to the American Diabetes Association (ADA) and a 2016 Cochrane review, cases of mild to moderate DKA can be safely managed on general medical floors utilizing subcutaneous short-acting insulin. Prior studies have demonstrated statistically insignificant differences in the total amount (units) of insulin used, time to resolution of DKA, patient outcomes, and time to discharge when compared to insulin infusion therapy.
Purpose: We share our experiences and identify challenges associated with managing mild to moderate diabetic ketoacidosis on the medical floor using subcutaneous insulin in a tertiary level hospital. As hospitals across the nations are facing unprecedented bed shortage both at critical care units and hospitals floors, we initiated a quality improvement project to manage patients with mild to moderate diabetic ketoacidosis with a protocol-based regimen on the medical floor using subcutaneous insulin with a focus on patient safety, quality of care, efficiency, and proper utilization of resources. This project was approved by our departmental quality improvement project committee.
Description: Patients diagnosed with DKA at the emergency department are triaged to either floor appropriate or intensive care unit appropriate based on severity and exclusion criteria. Those who are floor appropriate are started on protocol-based intravenous fluid and subcutaneous insulin with periodic (every two-hour glucose and every four-hour basic metabolic panel) laboratory monitoring. Once the anion gap closes patients are started on diet and insulin is transitioned to their home dose. We compared process outcomes and patient outcomes before and after the implementation of the project. Our findings show A) Localization of patients to a dedicated floor was more after implementation (44% vs 8%, p< 0.001) B) Basic metabolic panels were done as specified more frequently while patients are in ICU as compared to when they are on the floor (100% vs 74%, p=0.34), however, it was not statistically significant. C) Hypoglycemic events were more while patients were on intravenous insulin as compared to subcutaneous insulin (65% vs 62%, p < 0.001) which was statistically significant. D) Average time in hours for closure of anion gap was shorter in patients managed on the floor as compared to those on ICU (15 vs 31.1, p< 0.001). This could be due to patients on the floor being less sick than in ICU. E) There was no statistically significant difference in time for bicarbonate to be > 18mmol/l in either group (ICU 17.6 hrs vs floor 14.5 hrs, p=0.46). F) The total length of stay in the hospital was slightly more for patients managed in ICU as compared to patients managed on the floor (4.5 vs 4.1, p =0.82) which was not statistically significant. In addition, we noticed increasing confusion and hesitancy amongst the prescribers on proper utilization of protocolized order sets, execution of orders by nursing and pharmacy staff, especially in the early implementation phase.
Conclusions: Safe and effective management of selected patients with mild to moderate DKA with subcutaneous insulin on the medical floor using protocolized order sets is possible. However, it requires investment from the leadership, frequent education and reinforcement, and ongoing monitoring.