From the 2019 HVPAA National Conference
Dr. Taylor J. Roberts (Parkland Health & Hospital System), Dr. Nimish N. Shah (Parkland Health & Hospital System), Dr. Brian P. Duffy (Parkland Health & Hospital System), Dr. Emmanuel T. Johnson (Parkland Health & Hospital System), Dr. Timothy J. Brown (University of Texas Southwestern), Dr. Nainesh Shah (University of Texas Southwestern)
Background
Medication errors are prevalent at discharge and contribute to harmful adverse drug events with potentially avoidable healthcare utilization and cost. Documentation of medications at discharge is a potential source of errors, which may occur in discharge summaries (DS), discharge instructions (DI), or electronically generated discharge medication lists (eDML). Within Epic, our institution’s electronic health record (EHR), each of these document types may contain unique and independent discharge medication lists, creating opportunities for discrepancies. Providers could traditionally generate their DS and DI with statically-imported medication lists via Smartlinks; however, the most accurate medication list is the dynamic eDML.
Objective
Identify the frequency of discharge medication documentation errors, defined as discrepancies between DS, DI, and eDML medication lists, and reduce these errors to zero within one year by eliminating the use of Smarklinks that import medication lists.
Methods
This quality improvement initiative was conducted at Parkland Memorial Hospital (PMH), a large tertiary-care county hospital in Dallas, Texas. We performed a randomized chart review of 60 discharge encounters from the Internal Medicine, Cardiology, and Surgery departments between January 1, 2017 and March 30, 2017. We compared medication names, number of medications, dosages, and sigs among the DS, DI, and eDML for each encounter. After approval by the EHR governance committee, the Smartlinks that import discharge medication lists were disabled at PMH in April 2017. We performed a subsequent randomized chart review of 60 discharge encounters from each department between July 1, 2017 and September 30, 2017. Error frequencies before and after the intervention were then compared. We also reviewed 60 discharge encounters from the Internal Medicine, Cardiology, Cardiothoracic Surgery, Pulmonology, and Obstetrics & Gynecology departments at Clements University Hospital (CUH), a large tertiary-care private hospital within the same academic medical center, between January 1, 2017 and March 30, 2017; however, no intervention was performed.
Results
At PMH prior to the intervention, 76% of DS included a medication list, and 21% of DS contained errors. Post-intervention, these rates were reduced to 45% and 10%, respectively. Prior to the intervention, 40% of DI included a medication list, and 12% of DI contained errors. Post-intervention, these rates were reduced to 18% and 6%, respectively. Findings were similar across departments. The DS or DI that contained a medication list post-intervention utilized alternative Smartlinks that had not been disabled. At CUH, 95% of DS included a medication list, 17% of DS contained errors, 43% of DI included a medication list, and 8% of DI contained errors.
Discussion
We demonstrate that discharge medication lists imported into multiple discharge documents are frequently associated with error, and that modifying EHR functionality to curtail medication list importation can reduce these errors.
Clinical Implications
Medication errors at discharge may be reduced by disabling EHR functions that import redundant but static medication lists into discharge documents. Hospitals should consider similar interventions in order to eliminate preventable adverse events due to medication error in their efforts to improve patient safety.
