Long‐Term Impact of a Clinical Decision Support System on Potassium Prescription Safety in a Brazilian Tertiary‐Care Hospital: A Retrospective Observational Study (2006–2023)
Gustavo Santos Paiva Laender Moura, Allana Campos Alves, Fernando Fávero, Lucas Chaves, Ajith Kumar Sankarankutty, Eduardo Barbosa Coelho, Julieta Mieko Ueta, Antônio Pazin‐Filho, Fabio CarmonaABSTRACT
Rationale
Limited evidence exists on the long‐term impact of Clinical Decision Support Systems (CDSS) on potassium safety in middle‐income countries, with most studies focusing on short‐term outcomes in resource‐rich settings.
Aims and Objectives
This study evaluates the long‐term effectiveness of a CDSS in reducing potassium‐related adverse events in a Brazilian tertiary‐care hospital from 2006 to 2023.
Methods
This retrospective observational study analysed all potassium‐containing prescriptions (oral and intravenous) at a Brazilian tertiary‐care teaching hospital. The study compared the pre‐implementation (2006–2014) and post‐implementation (2014–2023) phases of a CDSS that monitors and blocks prescriptions exceeding predefined safety limits for potassium concentration, infusion rate, and daily dose. Near misses were retrospectively identified by applying the same CDSS‐defined rules to completed prescriptions—both before and after CDSS implementation. Trends in adverse events and blocked prescription attempts were analysed using Prais‐Winsten regression. Data were obtained directly from hospital databases, without coding or transformation. The study adhered to the RECORD statement and received ethical approval.
Results
Of 9,308 drug‐related adverse events recorded between 2013 and 2023, 64 (0.7%) involved potassium salts. Despite an increase in overall drug‐related events ( p < 0.001), the proportion involving potassium decreased from 2.8% in 2013% to 0.2% in 2023 ( p = 0.008). After CDSS implementation, 90,741 potassium prescriptions were submitted, of which 38,961 (42.9%) triggered CDSS blocked attempts, showing a consistent upward trend ( β = 20.4, p < 0.001). Most blocked attempts were due to violations of intravenous concentration (60%) and infusion rate (18%) limits. Near‐miss prescriptions remained proportionally high [median 27.4% (IQR: 17.5–32.2)] throughout the post‐CDSS phase.
Conclusion
This study demonstrates the long‐term effectiveness of a potassium‐specific CDSS in reducing adverse drug events in a middle‐income setting. However, persistent increases in blocked attempts and near misses suggest potential issues related to alert fatigue, overreliance on CDSS, and override practices, highlighting the need for ongoing system refinement and staff training.