Keywords
How to Cite
Abstract
Introduction: Children hospitalized with pneumonia are usually treated with more than one drug, and potential drug–drug interactions (DDIs) are a plausible contributor to poor outcome. A retrospective pilot of 180 children found no association between documented DDI level and treatment outcome, but its interval excluded only moderate or larger associations, interactions were mostly absent or minor (none severe), DDI level was a global clinician rating, and an ordinal outcome was analyzed with a linear model. This protocol describes a prospective study designed to address these limitations.
Methods and analysis: A prospective, multicentre, observational cohort study will be conducted in several tertiary care hospitals in Peshawar, Pakistan. Consecutive children aged 2 months to 12 years admitted with clinically diagnosed pneumonia and receiving two or more concurrent drugs will be enrolled (proposed target 500). Every concurrent drug pair will be screened against a single interaction reference by two pharmacists blinded to outcome, giving pair-level exposure data (highest severity, number of interacting pairs, days of exposure). The primary outcome is a four-level ordinal outcome at discharge (recovered, improved, ICU transfer, death). The primary analysis is a proportional-odds regression adjusted for comorbidity, age, illness severity, number of concurrent drugs, antibiotic regimen and hospital. Secondary analyses use a time-to-event model with time-varying exposure. The sample size gives about 83% power to detect a small effect (Cohen’s f2 = 0.02).
Ethics and dissemination: Approval will be sought from the institutional ethical review committee and the participating hospitals, with written informed consent from parents or guardians. Results will be reported following STROBE and published in a peer-reviewed journal.