FDA Guidance Intelligence — PQ-CMC Structured Data (FHIR) Submission Requirements

FDA’s Pharmaceutical Quality–CMC structured data program has moved well past the pilot stage — the Stage 2 FHIR Implementation Guide cleared HL7 balloting in early 2025 and is now in active build toward a production release — and the mandatory compliance window for NDA and ANDA filers is no longer a distant policy question. It is a data-systems question your CMC organization needs to answer this quarter, not after FDA sets the deadline.

That sentence is not a forecast dressed up as urgency. FDA’s PQ-CMC structured data initiative began as a voluntary pilot, and it has not stood still: the HL7 FHIR Implementation Guide for Pharmaceutical Quality–Chemistry, Manufacturing and Controls (PQ-CMC) reached Stage 1 as a published Standard for Trial Use, Stage 2 — which adds substantially expanded PQ/CMC data content — was successfully demonstrated at the September 2024 HL7 FHIR Connectathon and entered the January 2025 HL7 ballot cycle, and a further build toward version 2.0.0 is now underway. Under FDA’s PDUFA VII commitments, this is the technical infrastructure a mandatory submission requirement is built on. The companies that understand this are treating the current phase as the last calm window before FDA formally sets a compliance date. The companies that do not are still treating PQ-CMC as a future consideration to revisit once the mandate is official — which is exactly the posture that turns a manageable data-systems project into a submission-blocking scramble.

What FDA’s PQ-CMC Program Has Actually Built — And Why the Absence of a Final Mandate Is Not the Same as Low Risk

The FDA/HL7 Pharmaceutical Quality–CMC program emerged from a foundational recognition at the agency: the information contained in Module 3 CMC submissions is extraordinarily rich in scientific content and extraordinarily difficult to process efficiently because it arrives as unstructured text embedded in Word documents and PDFs. A reviewer evaluating a drug product specification against ICH Q6A criteria must manually locate, read, and compare fields that, in a structured data environment, would be machine-readable and automatically cross-referenced against historical submissions, precedent decisions, and pharmacopeial standards. That processing gap creates review latency, introduces variability across reviewers, and makes systematic quality signal detection across the submission portfolio essentially impossible. The PQ-CMC program exists to close that gap, and FDA has continued to invest in it methodically rather than abandon it — which is itself a signal worth reading correctly.

The technical architecture FDA selected is HL7 FHIR — Fast Healthcare Interoperability Resources — the same interoperability standard driving electronic health record data exchange for the past decade. FHIR is not a pharmaceutical-specific standard; it is a general-purpose healthcare data exchange framework that FDA and HL7’s Biomedical Research and Regulation work group have adapted for CMC content through the PQ-CMC Implementation Guide. The guide defines resource types — MedicinalProductDefinition, Ingredient, ManufacturedItemDefinition, SubstanceDefinition among them — and specifies the data fields, code systems, and value sets required for each CMC domain. What has changed since the pilot period is the maturity of that specification: Stage 2 adds the expanded data content needed to represent drug substance and drug product specifications, stability data, and batch analyses in a way that is no longer experimental.

What this means operationally is that a drug substance specification — which currently lives as a table in a Word document inside a CTD Module 3 section — will, in a compliant PQ-CMC submission, need to exist as a structured FHIR bundle in which each test has a machine-readable name from a defined terminology, each method is referenced by a coded identifier, each acceptance criterion carries an operator (NMT, NLT, equals), a numeric value, and a units code, and each analytical procedure has a validation status field populated from a controlled vocabulary. The narrative prose that CMC regulatory writers have spent careers perfecting does not disappear — it accompanies the structured data — but the structured data layer has to exist and has to validate against the published HL7 FHIR PQ-CMC Implementation Guide once FDA sets the effective date.

The absence of a finalized Federal Register mandate as of today is not the same signal as low risk. FDA’s PDUFA VII commitments explicitly tie CMC review efficiency performance goals to structured data adoption, which gives the agency a programmatic incentive to move from voluntary to mandatory once the Implementation Guide is production-ready — and Stage 2’s progression through HL7 balloting is the clearest indicator yet that FDA considers the specification close to that point. The pilot program gave the industry years to observe, prepare, and build. The Stage 2 ballot cycle converts that preparation window from open-ended to finite.

The Data Systems That Must Change Before the Mandate Arrives, Not After

The compliance challenge created by PQ-CMC maturation is not a regulatory writing challenge. Your regulatory affairs team cannot solve it by learning FHIR syntax and hand-coding XML files. The challenge is that the structured data FDA will require must come from the systems where CMC data is actually generated, stored, and managed — and most of those systems were not built with FHIR output as a design requirement.

A laboratory information management system that stores test results, method references, and acceptance criteria for drug substance and drug product specifications is the primary data source for PQ-CMC structured submissions. If that LIMS can export structured data in HL7 FHIR format — specifically, if it can generate valid FHIR Bundle resources that conform to the PQ-CMC Implementation Guide — the path to readiness is an implementation and validation project you can start now, on your own timeline. If that LIMS cannot generate FHIR output, or if the data it contains is not structured at the field level the IG requires — test names as free-text strings rather than coded terminology entries, or acceptance criteria stored as a single text field rather than separate operator, value, and units fields — the path requires a system upgrade, a middleware integration layer, or a data remediation effort that takes months to scope and execute, not weeks.

Electronic laboratory notebooks present a parallel challenge, and document management systems are the third critical component. The common gap that emerges when pharmaceutical companies conduct an honest assessment of their current CMC data infrastructure is this: specifications are maintained in Word or PDF documents, not in structured data fields. The test name, method reference, acceptance criterion, and validation status that FDA needs as discrete FHIR fields exist somewhere in those documents — a reviewer can read them — but they do not exist as machine-readable discrete data elements that a FHIR export can capture. Closing that gap is the core work of PQ-CMC readiness, and it cannot be completed in the weeks after FDA announces an effective date.

THE XGENE PQ-CMC READINESS SPRINT

Step 1 — Submission Type Assessment: Identify every IND, NDA, and ANDA in your pipeline with a target submission date within the next 24 months and determine which CMC domains — drug substance specifications, drug product specifications, stability data, batch analyses — are most likely to fall inside FDA’s initial PQ-CMC scope based on the current Stage 2 Implementation Guide content.

Step 2 — Data System Audit: Conduct a structured assessment of your LIMS, ELN, and document management system against the published HL7 FHIR PQ-CMC Implementation Guide (currently advancing toward v2.0.0). For each system, determine whether FHIR export capability exists, what IG version it targets, whether field-level data granularity meets IG requirements, and what vendor roadmap commitments exist.

Step 3 — Gap Remediation Roadmap: Based on the audit findings, assign each identified gap to one of three tracks — vendor software update, middleware integration, or data remediation — each with a timeline, responsible owner, and vendor contact.

Step 4 — First Structured Submission Dry Run: Generate a PQ-CMC FHIR data bundle for one drug substance or drug product specification section and validate it against the published Implementation Guide using available validation tools. This dry run surfaces integration issues while there is still time to correct them — and gives your CMC regulatory team hands-on experience with the structured submission workflow before a mandate makes it non-negotiable.

If your next NDA or ANDA filing has a target submission date within the next 24 months, the question that determines your readiness is simple: can your LIMS generate a valid HL7 FHIR bundle for your drug substance and drug product specifications today? Not once FDA sets a deadline — today. If the answer is no, or if the answer is “we don’t know,” that is the gap FDA’s PQ-CMC mandate will identify the moment it is finalized. Pull your pipeline submission dates, schedule the data system assessment this week, and establish your remediation track while the runway is still measured in quarters, not weeks.

Identify your next NDA or ANDA filing with a target submission date within 24 months and assess today whether your CMC data systems can generate HL7 FHIR-compliant structured output for drug substance and drug product specifications — if the answer is “we don’t know,” that is the gap FDA will identify the moment PQ-CMC becomes mandatory.