BACKGROUND: The Pharmacovigilance Programme of India (PvPI) is responsible for collecting reports of adverse drug reactions (ADRs) to assess the association between particular drugs and ADRs. The aim of the present study was to apply statistical tools to determine associations between drugs and ADRs for signal detection in the PvPI. METHODS: Four methods were proposed for quantitative signal detection: one was based on Bayesian inference and others on classical inference procedures. The effectiveness of the proposed methods was assessed by applying them to 4 drug-ADR combinations. RESULTS: The proposed methods were easy to apply and relevant to the Indian context. In selected methods, the information component value was more specific, whereas the proportional relative risk was more sensitive. CONCLUSIONS: The proposed methods may help in the identification of new signals in Indian individual case safety reports.
BACKGROUND: The Pharmacovigilance Programme of India (PvPI) is responsible for collecting reports of adverse drug reactions (ADRs) to assess the association between particular drugs and ADRs. The aim of the present study was to apply statistical tools to determine associations between drugs and ADRs for signal detection in the PvPI. METHODS: Four methods were proposed for quantitative signal detection: one was based on Bayesian inference and others on classical inference procedures. The effectiveness of the proposed methods was assessed by applying them to 4 drug-ADR combinations. RESULTS: The proposed methods were easy to apply and relevant to the Indian context. In selected methods, the information component value was more specific, whereas the proportional relative risk was more sensitive. CONCLUSIONS: The proposed methods may help in the identification of new signals in Indian individual case safety reports.
Keywords:
Bayesian; pharmacovigilance; signal detection