EMA Medicines API — EPAR, DHPC & Shortage Monitor

Search and monitor four official European Medicines Agency datasets from one Actor:

• medicine pages and EPAR metadata;
• post-authorisation procedures;
• direct healthcare professional communications (DHPCs);
• ongoing and resolved medicine shortages.

EMA refreshes its machine-readable JSON files twice daily. This Actor turns them into compact, filterable records or durable added, updated, and removed change events for AI agents, scheduled alerts, regulatory intelligence, and data pipelines.

Why this Actor

Generic medicine scrapers return a current table. This Actor also answers the operational question: what changed since my last successful run?

• Official EMA JSON sources; no browser scraping or proxy required.
• One schema across four high-value regulatory datasets.
• Search by medicine, active substance, company, status, therapeutic area, or any record field.
• Persistent baselines and stable delivery IDs for scheduled monitoring.
• Field-level diffs for updated records.
• Failed sources never create false removal alerts.
• Monitoring state does not advance when the output budget is exhausted.

Fast starts

Search oncology medicines and safety communications

{
  "sources": ["medicines", "dhpc"],
  "keywords": ["oncology"],
  "keywordMode": "any",
  "categories": ["Human"],
  "mode": "snapshot",
  "maxResults": 200
}

Monitor one active substance quietly

Run this on an Apify schedule. The first run creates a baseline; later runs emit only changes.

{
  "sources": ["medicines", "post_authorisation", "dhpc", "shortages"],
  "activeSubstances": ["semaglutide"],
  "mode": "changes",
  "firstRunMode": "baseline_only",
  "stateNamespace": "semaglutide-watch"
}

Track a company portfolio

{
  "holders": ["AstraZeneca"],
  "mode": "changes",
  "firstRunMode": "emit_current",
  "stateNamespace": "az-portfolio"
}

Output

Snapshot mode emits normalized ema_record rows with the complete official record under data. Change mode emits ema_change rows with:

• changeType: added, updated, or removed;
• changedFields: exact official fields whose values changed;
• before and after: source records for auditability;
• deliveryId: stable idempotency key for webhook and workflow deduplication;
• sourceUrl: direct EMA page.

The Actor writes a run-level SUMMARY record to the default key-value store. It reports source timestamps, source errors, matched and emitted counts, output limits, and whether monitoring state advanced.

Scheduling and alerts

Use Apify Schedules to run after EMA's twice-daily refresh. Connect a dataset webhook or an automation platform to route non-empty change results. For a quiet first run, choose baseline_only.

Each set of filters gets isolated state. Change stateNamespace when you want a fresh or parallel baseline.

Reliability semantics

Change delivery is at least once. Deduplicate downstream with deliveryId.

If a selected EMA source fails, its last known records are preserved and no removals are inferred from the failed fetch. If a run cannot emit all candidate changes because of maxResults or the Actor spending limit, state does not advance; the same stable change IDs may appear again on the next run.

Official source

The data comes from the EMA machine-readable website data files. EMA states that these JSON files update at 06:00 and 18:00 Amsterdam time.

This Actor republishes descriptive public regulatory metadata. It does not provide medical, legal, investment, or regulatory advice.