reporting-adverse-events

Reporting adverse events into FAERS / ICH E2B(R3)

A pharmacovigilance case starts as free-text narrative ("68 yo on warfarin developed GI bleed, hospitalized"). To make it reportable you must structure it into the ICH E2B(R3) data elements that the FDA's FAERS (and EMA's EudraVigilance) expect: a suspect drug, one or more reactions coded to MedDRA Preferred Terms, seriousness criteria, and a reaction outcome.

OpenMed extracts the drug and condition spans on-device; this skill turns those spans plus the narrative into the E2B(R3) skeleton. The reaction coding step needs MedDRA, which is licensed by the MSSO and user-supplied — it is never bundled with OpenMed and must be loaded from the user's own subscription.

When to use

• A narrative names a drug and an adverse reaction and you need an ICSR (Individual Case Safety Report) shell with the right E2B(R3) fields.
• You must classify seriousness (E2B sections C.1.7 / E.i.3) — death, life-threatening, hospitalization/prolongation, disability, congenital anomaly, or "other medically important condition".
• You need to characterize each drug as suspect / concomitant / interacting (the drugcharacterization axis FAERS uses).
• You are pre-filling a 3500A / FAERS electronic submission or staging cases for a safety database.

This skill produces a structured draft for human safety review — it does not file reports or perform causality assessment autonomously.

Quick start

import openmed

narrative = (
    "68-year-old patient on warfarin 5 mg daily developed a gastrointestinal "
    "hemorrhage and was hospitalized. Warfarin was discontinued; the patient "
    "recovered."
)

# 1) Extract drug spans (Pharmaceutical category) on-device.
drugs = openmed.analyze_text(
    narrative,
    model_name="pharma_detection_superclinical",
    output_format="dict",
)["entities"]

# 2) Extract condition / reaction spans (Disease category).
conditions = openmed.analyze_text(
    narrative,
    model_name="disease_detection_superclinical",
    output_format="dict",
)["entities"]

# 3) Assemble an E2B(R3)-shaped ICSR skeleton (reaction PTs filled later via MedDRA).
icsr = {
    "patient": {"age": None, "sex": None},          # from de-identified demographics
    "drugs": [
        {
            "name": e["text"],
            "drugcharacterization": 1,              # 1=suspect 2=concomitant 3=interacting
            "action": None,                          # e.g. drug withdrawn / dose reduced
        }
        for e in drugs
    ],
    "reactions": [
        {
            "verbatim": e["text"],                   # narrative term, pre-MedDRA
            "meddra_pt": None,                       # coded with user's MedDRA dict
            "outcome": None,                         # E2B reaction outcome code
        }
        for e in conditions
    ],
    "seriousness": {
        "serious": None, "death": False, "lifeThreatening": False,
        "hospitalization": True, "disability": False, "congenitalAnomaly": False,
        "otherMedicallyImportant": False,
    },
}