Presentation (CTI)

OBJECTIVES: To evaluate variability across commonly used drug-drug interaction databases in detecting and classifying DDIs in oncology prescriptions and to assess the potential implications of this variability on clinical decision support in Indian oncology practice.
METHODS: A cross‑sectional audit of 600 prescriptions from oncology outpatient and inpatient departments at a tertiary care hospital in India was conducted over six months (April-September 2025). Prescriptions containing ≥2 systemic anticancer or supportive care medications were screened. Identified combinations were evaluated for DDIs using three widely referenced DDIDBs (e.g., Lexicomp®, Micromedex®, and a publicly accessible database). For each DDI pair, presence/absence, severity, and documentation levels were recorded. Inter‑database agreement was analyzed using Cohen’s kappa (κ) statistics. The potential clinical impact was interpreted with reference to oncology‑specific guidelines and expert clinical pharmacology review.
RESULTS: Out of 600 prescriptions, 348 (58%) contained ≥2 interacting drug pairs. The number of DDIs identified varied significantly across databases: Database A detected 842 interactions, Database B 612, and Database C 975. Agreement between databases ranged from fair to moderate (κ = 0.32-0.55). Discrepancies were most pronounced in supportive care drug pairs (e.g., antiemetics, antifungals) and pharmacokinetic interactions involving CYP450 modulators. Severity categorization diverged in 44% of overlapping pairs, potentially altering clinical action plans. Expert review indicated that reliance on a single DDIDB would have missed 18-27% of clinically significant interactions.
CONCLUSIONS: There is considerable inter‑database variability in DDI identification and severity classification in oncology prescriptions. Such inconsistency may undermine the accuracy and reliability of CDSS alerts, potentially affecting clinical decisions and patient safety. These findings underscore the need for standardization and context‑specific calibration of DDIDBs before integration into oncology CDSS in India. Future work should focus on development of an oncology‑centric DDI reference, consensus frameworks, and evaluation of real‑world clinical impact.