Why literature monitoring is a PV obligation
Published articles are one of the sources of individual case safety reports (ICSRs). EU Good Pharmacovigilance Practices (GVP) Module VI asks marketing authorisation holders to screen widely used reference databases at least once a week, and to cover local journals in the countries where their products are authorised. A report found in an article is handled like any other spontaneous or solicited report: when it meets the minimum criteria for a valid case, the reporting clock starts on the day the company first became aware of it (day zero).
Doing this by hand means running the same searches every week, reading hundreds of abstracts that turn out to be irrelevant, ordering full texts, and keying the valid cases into the safety database. It is careful work, but most of it is reading and typing.
How Edriza screens the literature
Edriza covers three routes into the literature, and each article follows the same steps afterwards.
- Live Scan runs your search strategy for each product against published literature such as PubMed/Medline, on the schedule your SOP sets, and keeps every run with its query and its results.
- Local Journal monitors the local and regional journal sources you subscribe to, so that local literature is covered as well as the international databases.
- Document Scan reads full-text articles and other documents you upload, for example a paper sent by a partner or a full text ordered after an abstract looked relevant.
For every article, Edriza checks whether it concerns one of your products, whether it describes an adverse event in an identifiable patient, and whether the four minimum criteria for a valid ICSR are present: an identifiable reporter, an identifiable patient, a suspect product and an adverse reaction. It then assesses seriousness and causality, checks for duplicates against the cases you already hold, and enters the case data. Articles that are not relevant are not simply dropped: each one is recorded with the reason, so that you can show an inspector what was screened and why it was not a case.
From article to E2B(R3) case
A valid literature case continues through the same flow as every other source in Edriza: data entry, MedDRA coding of the reactions and medical history, a narrative, quality control, and an E2B(R3) XML that is checked against the E2B(R3) schema (XSD) before submission. The literature reference is kept with the case, and day zero is recorded from the date the article was first seen. If the full text is needed to complete the case, the request and its result are tracked with the case as well. See ICSR case processing and E2B(R3) submission for the steps after screening.
Two AI agents, a QC agent and your experts
Every article is read by two independent AI agents. Where they agree and the case is clear, the work moves on; a quality-control agent then checks the result against the source. Where the agents disagree, where information is missing, or where your SOP asks for a person, the article goes to one of your experts with the evidence highlighted. This is what makes the workflow fully automated with human oversight: the reading and data entry are automated, and the decisions that need judgement stay with your team. Read more in pharmacovigilance automation with AI.
What you can show at an audit
- The search strategy per product and every run, with the date and the number of results.
- Every article screened, with the decision (case, not a case, needs follow-up) and the reason.
- Who or which agent decided each step, and when.
- The link from each literature case to its E2B(R3) report and its submission.
Part of one platform
Literature monitoring in Edriza is not a separate tool: the cases it finds go straight into the same case processing, QC, E2B(R3) generation and submission, with one audit trail and no extra subscription. See the platform modules or contact us to discuss your products and sources.