Adverse drug reactions in drug information databases: does presentation affect interpretation?

Sean M Mcconachie,Christopher A Giuliano,Pramodini B Kale-Pradhan,Insaf Mohammad

doi:10.5195/jmla.2020.748

Abstract

ObjectiveFormatting of adverse drug reaction (ADR) information differs among drug information (DI) resources and may impact clinical decision-making. The objective of this study was to determine whether ADR formatting impacts adverse event interpretation by pharmacy practitioners and students.MethodsParticipants were randomized to receive ADR information in a comparative quantitative (CQUANT), noncomparative quantitative (NQUANT), or noncomparative qualitative (NQUAL) format to interpret 3 clinical vignettes. Vignettes involved patients presenting with adverse events that varied in the extent to which they were associated with a medication. The primary outcome was interpretation of the likelihood of medication-induced adverse events on a 10-point Likert scale. Lower scoring on likelihood (i.e., ADR deemed unlikely) reflected more appropriate interpretation. Linear regression was performed to analyze the effects of ADR information format on the primary outcome.ResultsA total of 108 participants completed the study (39 students and 69 pharmacists). Overall, the CQUANT group had the lowest average likelihood compared to NQUAL (4.0 versus 5.4; p<0.01) and NQUANT (4.0 versus 4.9; p=0.016) groups. There was no difference between NQUAL and NQUANT groups (5.4 versus 4.9; p=0.14). In the final model, at least 2 years of postgraduate training (−1.1; 95% CI: −1.8 to −0.3; p<0.01) and CQUANT formatting (−1.3; 95% CI: −0.9 to −1.7; p<0.01) were associated with reduced likelihood.ConclusionsFormatting impacts pharmacists’ and pharmacy students’ interpretation of ADR information. CQUANT formatting and at least two years of postgraduate training improved interpretation of adverse events.

Highlights

Prescription drug utilization continues to rise in the United States
Vignettes involved patients presenting with adverse events that varied in the extent to which they were associated with a medication
The comparative quantitative (CQUANT) group had the lowest average likelihood compared to noncomparative qualitative (NQUAL) (4.0 versus 5.4; p

Summary

Introduction

Prescription drug utilization continues to rise in the United States. From 1999 to 2011, the prevalence of prescription drug use among adults increased from 51% to 59% and polypharmacy (i.e., 5 or more medications) increased from 8% to 15% [1]. The prevalence of adverse drug reactions (ADRs) may be increasing. Of hospital admissions are complicated by an ADR [3, 4]. Despite their high incidence and associated morbidity, ADRs can be difficult to identify in clinical practice. Most randomized controlled trials that lead to drug approval are not powered to detect rare adverse effects, which has led to a reliance on pharmacovigilance programs such as post-marketing surveillance, patient registries, and patient-level causality assessment to determine the likelihood of ADR occurrences in clinical practice [5,6,7,8]. Studies show that there is often little consensus among experts or assessment tools in identifying and classifying ADRs in real-world settings [9,10,11,12]

Methods

Results

Conclusion