Investigation of Clinically Significant Drug-Drug Interactions in Outpatient Prescriptions and Analysis of Associated Factors Using Bayesian Model Averaging (BMA) and Logistic Regression at Da Nang Hospital

Abstract

This study aimed to investigate patient characteristics, patterns of medication use, and to analyze clinically significant drug–drug interactions (DDIs) and their associated factors in outpatient prescriptions at Da Nang Hospital. A retrospective cross-sectional design was applied to outpatient prescriptions collected from the Outpatient Department of Da Nang Hospital between May 2024 and June 2024. DDIs were assessed by cross-referencing four databases (Drugs.com, Medscape, Micromedex, and Decision No. 5948/QD-BYT of the Ministry of Health). An interaction was considered clinically significant when it appeared in at least two of the four databases and met the severity criteria defined by the European Medicines Agency (EMA). Associated factors were analyzed using Bayesian Model Averaging (BMA) and
multivariable logistic regression. A total of 885 prescriptions were analyzed. The mean patient age was 64.9 ± 10.5 years (range: 20–96 years), and 89.6% of patients had two or more comorbidities. The average number of medications per prescription was 4.4 ± 1.3. The prevalence of prescriptions containing clinically significant DDIs was 47.8%, and the frequency of interactions at each severity level differed significantly across the databases. Forty-eight clinically significant DDI pairs were identified, of which the most common were amlodipine–metformin, insulin–metformin, and clopidogrel–esomeprazole. BMA identified four factors strongly associated with the risk of clinically significant DDIs: number of medications, use of cardiovascular drugs, diuretics, and hormonal/endocrine agents. Multivariable logistic regression showed that all these factors significantly increased the risk of clinically significant DDIs (p < 0.001). The study demonstrated that the prevalence of clinically significant DDIs in outpatient prescriptions increased with a higher number of medications and greater use of drug groups such as cardiovascular agents, diuretics, and hormonal/endocrine therapies (p < 0.001).