Implementation of a Best Practice Advisory to Alert Inpatient Providers of Necessary Discharge Prescriptions for Insulin and Supplies for Patients with Diabetes

Objective:
Prescribing errors or omissions on hospital discharge are common, avoidable, and a potential source of patient harm. A 2016 study at our institution found that one-third of patients who were new to insulin were not getting the necessary diabetes-related prescriptions on discharge. In 2019, a dot-phrase in the discharge summary improved correct diabetes prescriptions up to 92% in this cohort. We aim to improve insulin and diabetes-supply prescription accuracy for all patients seen by diabetes education using a BestPractice Advisory (BPA) to alert discharging providers of diabetes recommendations for discharge.

Methods:
A BPA was created in the discharge navigator for all patients seen by a Diabetes Care and Education Specialist (DCES) during their admission. This alert notifies the primary provider of necessary diabetes supplies on discharge. The recommendations from the DCES (customized to the patient’s insurance, with Epic codes provided) open on the side panel with a single click. The ambulatory diabetes insulin/supplies orderset link is also provided. The BPA clears only once the provider acknowledges that insulin/supply orders have been placed. This BPA rolled out enterprise-wide on 2/1/2022, but only in use at our institution's Ohio hospital locations where DCES is trained with dot-phrases that activate the BPA. In this quality improvement project, we analyzed BPA usage from 2/1/22-12/31/22. To review the effectiveness of this intervention, chart review was conducted on 100 eligible patients hospitalized consecutively in 12/2022.

Results:
The alert fired 6,714 times for 3,049 patient encounters in 11 months. Of 100 patients sampled (58% male, mean age 57±15.7 years and A1c 8.7±2.6%), 76% had type 2 diabetes, 11% had type 1 diabetes, and 13% had other diabetes. Patients were hospitalized for mean 8 days; primary teams were 77% medical and 23% surgical. Ninety-nine patients (99%) had the correct type and formulation (pen/vial) of insulin prescribed on discharge, or did not require this prescription. Eighty-eight patients (88%) had all recommended supplies prescribed (needles, lancets, and test strips), yet, only 1 patient (1%) called in after discharge in need of supplies. This suggests some supplies may have been omitted intentionally, for example, if the patient already had sufficient supplies at home.

Discussion/Conclusion:
Using a BPA built into the Discharge Navigator, we provided a hard stop with timely communication of DCES recommendations to increase the number of patients with diabetes who are discharged with the appropriate insulin and supplies.