FDA AEMS Explained: FAERS Replacement, Dashboard, & Safety Pipelines

On March 11, 2026, the U.S. Food and Drug Administration launched the Adverse Event Monitoring System (AEMS), executing the most extensive technological consolidation of public health surveillance infrastructure in the agency's modern history. AEMS unified seven disparate, legacy postmarket safety databases into a single platform. At launch, AEMS retired and replaced the FDA Adverse Event Reporting System (FAERS) for human drugs, biologics, cosmetics, and color additives, alongside the Vaccine Adverse Event Reporting System (VAERS) (co-managed with the CDC) and the animal-drug and animal-food Adverse Event Reporting System (AERS). By the end of May 2026, the migration expanded to absorb the Manufacturer and User Facility Device Experience (MAUDE) database for medical devices, the Human Foods Complaint System (HFCS) for foods and dietary supplements, and the Center for Tobacco Products Adverse Event Reporting System (CTPAE).

For pharmacovigilance (PV) directors, medical safety officers, regulatory operations managers, and life-sciences data engineers, the transition from FAERS to AEMS is not merely a branding update. It introduces real-time public data publication, deploys automated artificial intelligence (AI) tools for document digitization and Personally Identifiable Information (PII) redaction, and restructures public data access portals. However, widespread misconceptions have emerged regarding what AEMS actually changes. Crucially, mandatory sponsor reporting obligations under 21 CFR 314.80 and 600.80 remain legally identical, spontaneous reporting limitations persist unchanged, and the historical ASCII/XML quarterly data files continue to publish under the renamed AEMS Quarterly Data Extract (QDE) series.

                     [ FDA Adverse Event Monitoring System (AEMS) Migration ]
                                                |
       -----------------------------------------------------------------------------------
      |                                                                                   |
[ Phase 1: March 11, 2026 Launch ]                                  [ Phase 2: May 2026 Completion ]
- FAERS (Drugs, Biologics, Cosmetics)                                - MAUDE (Medical Devices)
- VAERS (Vaccines; co-managed with CDC)                              - HFCS (Human Foods & Dietary Supplements)
- AERS (Animal Drugs & Animal Feeds)                                 - CTPAE (Tobacco Products)

This guide provides a comprehensive technical and operational breakdown of the FDA AEMS platform. Supported by empirical analysis of the pre-AEMS historical repository (spanning 20.3 million reports) and the first post-launch 2026 quarterly data extract, we examine database consolidation timelines, public access mechanisms, AI integration boundaries, SOP remediation requirements, and signal-detection continuity.

---

What is FDA AEMS and which legacy systems did it replace, and when?

Prior to 2026, the FDA operated a fragmented patchwork of seven independent adverse-event collection portals across its various product centers (CDER, CBER, CDRH, CFSAN, CVM, and CTP). Each database maintained distinct data schemas, separate public search front-ends, independent Freedom of Information Act (FOIA) queues, and staggered publication schedules.

FDA leadership characterized this historical environment as outdated and fragmented — in Commissioner Marty Makary's words, "clunky systems" that "wasted millions of taxpayer dollars and created blind spots in our postmarket surveillance" — with the agency noting that maintaining seven disconnected systems cost approximately $37 million annually while processing roughly 6 million adverse-event reports per year. By consolidating these streams into AEMS, the FDA projects $120 million in net operational savings over five years while providing unified cross-product safety intelligence.

                              [ Legacy 7-Database Patchwork ]
             (FAERS, VAERS, AERS, MAUDE, HFCS, CTPAE, Cosmetics Portal)
                                             |
                                  [ March–May 2026 Migration ]
                                             |
                    [ Unified FDA Adverse Event Monitoring System (AEMS) ]
                                             |
       -----------------------------------------------------------------------------
      |                                      |                                      |
[ Real-Time Look-Up Dashboard ]    [ AEMS Quarterly Data Extracts ]       [ Centralized Intake Engine ]
- Instant public search            - ASCII / XML relational files         - Adverse events
- Interactive multi-product filters- Complete case lineage maintained     - Whistleblower reports
- Redacted narrative access        - OpenFDA API bridge                   - Product quality defects

Expanded Surveillance and Centralized Intake

Beyond traditional postmarket adverse-drug-reaction reporting, AEMS was engineered as a comprehensive agency-wide intake portal:

• Product Quality Complaints: Integrates voluntary consumer and healthcare professional complaints regarding packaging defects, contamination, labeling errors, and counterfeit formulations.
• Whistleblower & Misconduct Submissions: Establishes a dedicated, confidential intake channel for regulatory non-compliance, clinical trial fraud, and manufacturing deviations across all FDA centers.
• Cross-Product Signal Surveillance: Allows safety evaluators to query interactions spanning multiple product categories—such as assessing adverse events arising from drug-device combination products (e.g., autoinjectors) or drug-supplement interactions in a single search environment.

---

What changed for public data access: dashboard, quarterly extract files, APIs, and FOIA?

The launch of AEMS significantly altered how external researchers, biopharma intelligence teams, and the general public access postmarket safety data:

                            [ AEMS Public Data Access Channels ]
                                              |
      ---------------------------------------------------------------------------------
     |                         |                       |                               |
[ Real-Time Public Dashboard ] [ AEMS Quarterly Data Files ] [ OpenFDA & Custom APIs ]  [ FOIA Workflows ]
- Replaced FAERS Public        - Relational ASCII / XML - REST API endpoints          - Drastic reduction
  Dashboard (fis.fda.gov)        archive continued        feeding AEMS datasets;         in routine narrative
- Near real-time case indexing - 2026 Q1 extract          enhanced APIs promised        document requests via
- Multi-product query engine     verified post-launch     by FDA in coming months       instant redaction

1. The AEMS Public Dashboard

The legacy FAERS Public Dashboard hosted on fis.fda.gov was decommissioned and redirected to the new AEMS Public Dashboard:

• Real-Time Indexing: Under FAERS, public dashboard data refreshed on the quarterly extract cadence. AEMS dashboard data is updated daily (the agency describes it as near real-time) once initial automated redaction and validation are complete.
• Search Capabilities: Users can search simultaneously across products using brand or generic names and reaction terms coded in the Medical Dictionary for Regulatory Activities (MedDRA), filtered by year, quarter, outcomes, and reporter geography — with up to five products or reactions per query.
• Export Restrictions: The public web dashboard allows export of a limited set of search data (charts, tables, and the "Listing of Cases" view) to Excel. Comprehensive case-level data still requires the quarterly data extract files or a FOIA request, and dashboard counts may differ from the extract files due to deduplication and case-version handling.

2. Proof of Data Continuity: The AEMS Quarterly Data Files

A critical concern for industry data-science teams was whether the standardized FAERS Quarterly Data Extract (QDE) files—the foundational data source for academic pharmacoepidemiology, quantitative signal detection, and competitive intelligence—would be discontinued.

Official FDA dashboard documentation explicitly confirms: "FDA will still continue to provide the AEMS Latest Quarterly Data Files online." The download page was retitled from FAERS to AEMS Latest Quarterly Data Files, maintaining the historical relational schema:

                                [ AEMS Relational Data Structure ]
                                                |
       -----------------------------------------------------------------------------------
      |                     |                      |                     |                |
[ DEMO (Demographics) ] [ DRUG (Medications) ] [ REAC (Adverse Rx) ] [ OUTC (Outcomes) ] [ RPSR (Sources) ]
- Primary Link:         - Suspect / Concom     - MedDRA PT & LLT     - Death, Hospital,   - Reporter Type
  safetyreportid          Primary / Secondary    Reaction Term Codes   Life-Threatening     (HCP / Consumer)

Empirical Verification: Pre-AEMS vs. Post-AEMS Extract Analysis

To verify data continuity, our analysis examined both the pre-AEMS historical snapshot (spanning 1968 through early 2026) and the first official post-AEMS 2026 quarterly extract covering reports received in the first quarter (exported July 22, 2026):

                        [ Historical FAERS Cumulative Volume: 20,328,575 Reports ]
                                                    |
       ---------------------------------------------------------------------------------------------
      |                     |                     |                     |                     |
 [ 2022 ]              [ 2023 ]              [ 2024 ]              [ 2025 ]              [ 2026 Q1 ]
 1,523,665 reports     1,368,573 reports     1,319,104 reports     1,307,332 reports     310,100 reports

This empirical comparison confirms that data schemas, field enumerations, and demographic proportions in the post-AEMS extracts align seamlessly with the historical FAERS architecture, allowing automated ETL pipelines to operate without structural breaking changes.

For deeper context on historical reporting trends, consult our 20 million FAERS reports analysis and our guide on how to read a FAERS report.

---

What exactly does the AI do in AEMS (and what does it not do)?

Media coverage following the March 2026 launch frequently described AEMS as an "AI-powered" safety system that autonomously discovers drug safety signals. FDA's own program documentation describes a far more specific, bounded role. The agency's exact language: AEMS is designed to "strengthen safety surveillance capabilities through advanced case processing workflows, AI-based redaction and digitization tools, enhanced analytics, and comprehensive cross-product surveillance." That is the full extent of the AI description — redaction and digitization, deployed inside case intake.

                            [ Artificial Intelligence Boundaries in AEMS ]
                                                  |
       -----------------------------------------------------------------------------------
      |                                                                                   |
[ What AI DOES in AEMS (Operational Plumbing) ]                     [ What AI DOES NOT DO (Human Regulatory Core) ]
- AI-based redaction of personally identifiable                     - Autonomous signal adjudication
  information from case narratives                                  - Causality determination or liability assignment
- AI-based digitization of paper and PDF submissions                - Mandatory label warning or boxed warning decisions
- Enhanced analytics and cross-product query support                - Automated clinical hold or market recall triggers

1. Document Digitization

Historically, voluntary MedWatch Form FDA 3500 submissions arrived via postal mail, facsimile, or unstructured PDF attachments, requiring manual handling before the case entered a reviewable database. FDA's description of "AI-based... digitization tools" points to automating this intake step. The agency has not publicly detailed the underlying model architecture, so pharmacovigilance teams evaluating AEMS-derived timelines should not assume any particular technology beyond what FDA has stated.

2. Automated PII Redaction

Real-time public publication is only possible if identifying information is removed before a case becomes publicly visible. FDA ties its "AI-based redaction" tools directly to this obligation, noting AEMS will publish reports in real time "consistent with meeting agency obligations not to release individually identifiable patient or consumer information." Automated redaction is therefore the operational enabler of the new publication cadence — and the reason case narratives appearing in the public dashboard can be released within days rather than quarters.

3. Enhanced Analytics and Cross-Product Surveillance

Beyond intake, FDA promises "enhanced analytics, and comprehensive cross-product surveillance" — the ability to query adverse events spanning product categories in one environment. For combination products such as drug-eluting stents or autoinjectors, an event previously split between a device database (MAUDE) and a drug database (FAERS) can now be analyzed in a single surveillance view.

What AI Does NOT Do: Preserving Clinical Oversight

FDA Chief AI Officer Jeremy Walsh, who called AEMS "the biggest technical transformation in agency history," described an execution and delivery feat — not a delegation of scientific judgment. FDA's own documentation is explicit that the human regulatory core is unchanged:

• No Automated Causality Decisions: AI tools do not evaluate whether a drug caused an adverse event. FDA's AEMS materials reiterate that the presence of a report "does not mean that the drug or biologic caused the adverse event," and causality assessment remains the domain of the agency's clinical reviewers in CDER and CBER, supported where needed by linked data systems such as the Sentinel System.
• No Autonomous Regulatory Actions: Labeling changes, Boxed Warning additions, Dear Healthcare Provider letters, REMS modifications, and product withdrawals continue to require clinical review of primary records, epidemiological evidence, and formal deliberation.

---

What must pharmacovigilance SOPs, publications, and pipelines update?

While underlying statutory reporting obligations under 21 CFR are unchanged, the launch of AEMS necessitates immediate operational, documentation, and technical updates across biopharma safety organizations:

                            [ Biopharma AEMS Remediation Checklist ]
                                               |
      ---------------------------------------------------------------------------------
     |                         |                       |                               |
[ 1. SOP Nomenclature ] [ 2. Literature & CSRs ] [ 3. ETL Data Feeds ]          [ 4. Cross-Product Screening ]
Update references to    Revise citations to     Verify ASCII/XML parsing       Expand safety surveillance
"FAERS database" across FAERS; cite "FDA AEMS"  endpoints; update automated    to capture device-drug and
all PV & QA SOPs        in manuscripts/dossiers data-pull cron jobs            combination AEMS records

1. Standard Operating Procedure (SOP) Remediation

All Quality Management System (QMS) and PV SOPs referencing external safety surveillance must be audited and revised:

• Surveillance Protocols: Update procedures governing "Routine FAERS Signal Detection" to "Routine FDA AEMS Database Screening."
• Reference Document Citations: Replace links to decommissioned fis.fda.gov dashboards with the active fda.gov/safety/fda-adverse-event-monitoring-system-aems endpoints.
• Auditing & Inspection Readiness: Ensure regulatory audit documentation reflects that safety teams are monitoring the real-time AEMS portal alongside monthly or quarterly QDE data extracts.

2. Medical Writing and Scientific Publications

Manuscripts, Periodic Safety Update Reports (PSURs), Periodic Benefit-Risk Evaluation Reports (PBRERs), and Common Technical Document (CTD) dossiers must reflect accurate naming:

• Methodology Citations: When citing postmarket data covering periods through March 2026, cite "FDA Adverse Event Reporting System (FAERS), transitioned to the FDA Adverse Event Monitoring System (AEMS) in March 2026."
• Post-March 2026 Queries: Explicitly reference "FDA Adverse Event Monitoring System (AEMS) Public Dashboard / Quarterly Data Extract files."

3. Data Engineering and ETL Pipeline Verification

Bioinformatics and data engineering teams must ensure data pipelines remain resilient:

• Download Automation: Update automated scripts scraping or pulling quarterly data files to target the new AEMS URL path (/drugs/fda-adverse-event-monitoring-system-aems/fda-adverse-event-monitoring-system-aems-latest-quarterly-data-files).
• File Format Continuity: FDA's download page continues to publish the relational ASCII and XML extracts organized into the familiar case-level tables (DEMO, DRUG, REAC, OUTC, RPSR, THER, INDI). Inspection of the July 22, 2026 quarterly extract confirms the same case-level fields — seriousness flags, reporter type, MedDRA-coded reactions, and outcome categories — so legacy ETL logic runs unchanged.
• OpenFDA API Bridges: Verify that API queries targeting api.fda.gov/drug/event.json correctly map existing fields to the unified AEMS repository.

For related class-specific surveillance workflows, review our analyses across therapeutic areas:

• GLP-1 receptor agonist adverse events in FAERS/AEMS
• MAUDE drug-delivery device adverse events by the numbers
• SGLT2 inhibitor adverse events by the numbers
• TNF inhibitor adverse events in postmarket surveillance
• FDA drug recall and triage workflow

---

Which FAERS data limitations carry over unchanged?

A dangerous fallacy among commercial analysts and trade observers is the assumption that a modernized, real-time database eliminates the methodological limitations inherent in spontaneous reporting. The FDA's official AEMS documentation prominently reiterates the identical statutory disclaimers governing the legacy FAERS repository:

                  [ Fundamental Limitations of AEMS / Spontaneous Reporting ]
                                               |
    ---------------------------------------------------------------------------------------
   |                           |                           |                               |
[ 1. No Causality ]     [ 2. Underreporting ]       [ 3. Reporting Biases ]         [ 4. No Incidence Rates ]
Presence of a report    Only 1%–10% of real-world   Stimulated by news,             Total exposed patient
does NOT prove drug     adverse events are ever     litigation, and social media    denominator is unknown;
caused the event        spontaneously reported      (Weber Effect)                  cannot compute true risk

1. Spontaneous Reporting and the Lack of a Denominator

AEMS remains a passive, spontaneous surveillance system. It captures numerator data (the number of reported adverse events) without capturing denominator data (the total number of patients prescribed or exposed to the product). Consequently:

• True Incidence Cannot Be Calculated: A higher raw report count for Drug A compared to Drug B does not prove Drug A is more dangerous; it may simply reflect higher market share, recent launch enthusiasm, or extensive direct-to-consumer advertising.
• Underreporting Bias: Medical literature estimates that fewer than 10% (and often less than 1%) of all real-world adverse drug reactions are spontaneously submitted to regulatory authorities.

2. The Weber Effect and Stimulated Reporting

Reporting velocity in AEMS is heavily confounded by external publicity:

• The Weber Effect: Adverse event reporting typically peaks within the first two years following commercial launch or label expansion, subsequently declining regardless of true clinical event rates.
• Media and Litigation Artifacts: High-profile news coverage, class-action legal advertising, and regulatory safety alerts trigger massive waves of consumer submissions (consumers filed 49.5% of reports in the 2026 first-quarter extract).

3. Confounding by Indication and Multi-Drug Regimens

In complex disease areas like oncology, rheumatology, and cardiology, patients routinely receive multi-agent regimens alongside severe baseline disease morbidity. A reported fatal outcome in an oncology patient receiving a CELMoD, an anti-CD38 antibody, and dexamethasone may reflect progressive malignant disease, opportunistic infection, or background frailty rather than direct drug toxicity.

---

Frequently Asked Questions (FAQ)

Is FAERS still available after the AEMS launch?

FAERS as a standalone system has been superseded by the FDA Adverse Event Monitoring System (AEMS), and the legacy dashboard at fis.fda.gov now leads to the AEMS Public Dashboard. FDA is migrating historical adverse event data — more than 20 million cases from 1968 through 2026 — into the new platform, where records remain accessible through the dashboard and the quarterly data files.

Do the quarterly data extract files still exist under AEMS?

Yes. The FDA continues to publish the standard ASCII and XML relational database files under the renamed title AEMS Latest Quarterly Data Files. Empirical verification of the post-launch July 22, 2026 extract confirms that the table structure, primary keys, and MedDRA coding conventions remain backwards-compatible with legacy FAERS ETL pipelines.

Did mandatory manufacturer reporting requirements change with AEMS?

No. Biopharma sponsor reporting obligations under 21 CFR 314.80 (drugs), 21 CFR 600.80 (biologics), and 21 CFR 803 (medical devices) remain legally identical. Manufacturers must continue to submit 15-day expedited alert reports for serious, unexpected events and routine periodic safety reports via standard electronic E2B(R3) transmission gateways.

Does an adverse event report in AEMS prove that a drug caused the reaction?

No. The FDA explicitly disclaims that the presence of an adverse event report in AEMS does not establish causality. Reports frequently represent unverified spontaneous submissions that lack clinical detail, autopsy verification, or medical confirmation, and may be confounded by concomitant medications or underlying disease progression.

---

Sources

1. U.S. Food and Drug Administration (FDA): FDA Launches New Adverse Event Look-Up Tool: The Adverse Event Monitoring System (AEMS). FDA Press Announcement (Mar 11, 2026). Available at: https://www.fda.gov/news-events/press-announcements/fda-launches-new-adverse-event-look-tool
2. U.S. FDA Center for Drug Evaluation and Research: FDA Adverse Event Monitoring System (AEMS) Program Overview and System Architecture. Content current as of March 11, 2026. Available at: https://www.fda.gov/safety/fda-adverse-event-monitoring-system-aems
3. FDA AEMS Public Dashboard: Interactive Search Portal and FAQ Documentation. fis.fda.gov / FDA CDER Informatics. Available at: https://www.fda.gov/drugs/fda-adverse-event-monitoring-system-aems/fda-adverse-event-monitoring-system-aems-public-dashboard
4. FDA CDER / CBER Data Management: AEMS Latest Quarterly Data Files (Formerly FAERS QDE). Relational ASCII and XML Data Releases (Exported July 2026). Available at: https://www.fda.gov/drugs/fda-adverse-event-monitoring-system-aems/fda-adverse-event-monitoring-system-aems-latest-quarterly-data-files
5. Center for Infectious Disease Research and Policy (CIDRAP): FDA Announces AEMS, New Adverse-Event Database to Replace VAERS and FAERS. News Analysis (Mar 2026). University of Minnesota. Available at: https://www.cidrap.umn.edu/
6. National Center for Biotechnology Information (NCBI): FAERS Essentials: Methodology, Signal Detection Algorithms, and Regulatory Surveillance Principles. PMC12393772. National Institutes of Health. Available at: https://pmc.ncbi.nlm.nih.gov/articles/PMC12393772/
7. U.S. Code of Federal Regulations (eCFR): Title 21 CFR 314.80 and 21 CFR 600.80: Postmarketing Reporting of Adverse Drug and Biological Product Experiences. National Archives and Records Administration. Available at: https://www.ecfr.gov/