Background: Errors in medication reconciliation frequently occur at transitions of care. Patients discharged to skilled nursing facilities (SNFs) are particularly vulnerable to the consequences of these mistakes. An interprofessional team at UVA Health implemented a longitudinal quality improvement (QI) project to reduce medication reconciliation errors for patients discharging from acute care medicine services to SNF. Before this QI project, 60% discharges contained at least one medication error. After its implementation, discharges with a medication error were reduced to an average of 34%. To better understand reasons for ongoing errors and opportunities for improvement, we analyzed whether the concordance of a patient’s discharging unit and a team’s “home unit” affected discharge medication error rates.
Methods: A “geographic” discharge is defined as a patient discharge from the designated “home unit” of their medicine team. A “non-geographic” discharge occurs when the patient discharges from any other unit. Data from 1099 discharges were collected and analyzed from September 2020 to June 2022. An error in medication reconciliation was defined as a deviation from the best possible medication list at discharge as determined by a clinical pharmacist’s review of medication documentation. Errors were categorized by the Agency for Healthcare Research & Quality medication error classification tools. Errors were classified as severe if they had the potential to result in hospitalization or more significant harm. Medication error rates were compared between geographic and non-geographic discharges.
Results: Fewer medication reconciliation errors occurred for geographic compared to non-geographic discharges. These error rates (errors/discharge) were 0.809 and 0.937 (p< 0.00001) respectively. Severe error rates were also lower for geographic discharges than for non-geographic discharges with error rates of 0.0297 and 0.036 (p=0.74).
Conclusions: Our analysis of a longitudinal QI project demonstrates that geographic alignment of patients and their care team correlates with fewer medication reconciliation errors at discharge to SNFs. One possible explanation for this finding is better adherence to a key countermeasure implemented in the QI project, an interprofessional safety huddle to review discharge medications. When nurses, physicians and pharmacists share a geographic unit there is likely enhanced participation in the huddle process. In addition, unit medical directors conduct weekly quality rounds with geographic teams to highlight local QI efforts and reinforce patient safety best practices, including medication reconciliation efforts. Our results support the role for geographic patient alignment in improving patient safety, though further study is needed to identify the specific aspect of geography that resulted in fewer medication errors. We hope that the results of this study can be used to advocate for increased geography for care teams in a hospital setting.