GUDID: FDA Global Unique Device Identification Database
Understand the US medical-device reference database, the information behind a DI record, and the human decisions around submission and change.

On this page · 10 sections
- 1What GUDID is—and the operational question it answers
- 2US medical-device scope and the labeler responsibility check
- 3GUDID, UDI, DI, PI and AccessGUDID compared
- 4DI record content, PI flags and source-to-field mapping
- 5Ownership, human approval and the boundary of software support
- 6Submission routes, integrations and recovery from uncertain results
- 7Lifecycle management: assess changes before editing a record
- 8Worked cases: production changes, obsolete terminology and uncertain delivery
- 9Common data failures, their causes and targeted controls
- 10Implementation checklist and evidence for an internal readiness review
How does GUDID apply to your shop floor?
Pick your industry and scale — Ask V5 rewrites the definition in your context, gives a worked example, and shows what V5 does on day one.
01What GUDID is—and the operational question it answers
GUDID (Global Unique Device Identification Database) is the US FDA-administered reference catalog for medical devices with a unique device identifier (UDI). FDA’s GUDID overview describes this identification role. This article focuses on US medical devices, not a general registration system for pharmaceutical, food or chemical products.
For a device labeler, the practical question is: how can an identifier on a device label connect to an internally approved, consistently maintained reference record? A useful operating frame follows the information across these connected objects:
- The label: the identification information associated with the device being described.
- The source product information: controlled internal information used to prepare and review the reference data.
- The submitted record: the representation that the labeler compares with its approved source information.
As an internal control, reviewers should be able to trace a submitted value back to its approved source rather than treat a catalog entry as the source of truth for every downstream decision. This is a practical data-management approach, not an additional FDA requirement.
Keep identification, market-authorization assessment and product release as separate organizational decisions. A catalog match should not serve as internal authorization to market a device or approval to release product. Where an electronic release record is used, its human approval remains a separate decision. The point of the GUDID workflow is therefore consistent reference information—not replacing regulatory assessment or release judgment.
02US medical-device scope and the labeler responsibility check
For US medical devices, check the current text of 21 CFR Part 830 for scope before deciding applicability or making a binding-duty determination. Keep that legal-text assessment distinct from FDA website explanations and the operational review practices below.
Identify the labeler by activity
For US medical devices, FDA’s UDI overview describes submitting device information to GUDID among the labeler’s UDI activities. This is FDA explanatory material; use the legal-text review above to determine binding duties.
For US medical devices, FDA’s labeler explanation includes anyone who causes a device label to be applied or modified, with the intent that the device will be commercially distributed without subsequent label replacement or modification. For US medical devices, FDA also explains that the labeler may be a specification developer, single-use device reprocessor, convenience kit assembler, repackager or relabeler.
As an operational practice, document who causes labeling actions and the intended distribution arrangement. Avoid assigning responsibility solely from a company’s commercial description as manufacturer or distributor.
Route the applicability review
This table is a practical review map, not a statement of applicability or exemption.
| Jurisdiction and product | Reference route | Decision to document |
|---|---|---|
| US medical devices | Current Part 830 text; FDA explanatory resources separately | Applicable scope, labeler identity, timing, exceptions and alternatives |
| Products with uncertain device classification | Separate product-classification assessment | Classification basis and unresolved questions before a GUDID decision |
| Mixed drug, supplement, food or chemical portfolios | Product-by-product and market-by-market assessment | Which products enter the US device review; separate routes for the remainder |
| Devices marketed outside the United States | Destination-market assessment, such as the UK medical-device framework | Local identification assessment, kept separate from any US determination |
For internal review, record the rationale, responsible reviewer and unresolved applicability questions. Treat timing, exceptions and alternatives as current-review items rather than assumptions carried forward from another product or market.
03GUDID, UDI, DI, PI and AccessGUDID compared
For US medical devices, FDA’s UDI format explanation distinguishes the complete identifier from its components. The table separates that vocabulary from the database and public lookup service; it summarises FDA explanatory material rather than setting out binding-rule applicability.
| Term | What it represents | Operational use |
|---|---|---|
| GUDID | The reference catalog defined in the opening section. | Keep catalog maintenance distinct from label identifier construction and public lookup. |
| UDI | For US medical devices, FDA describes a UDI as a unique numeric or alphanumeric code. | Use UDI when discussing the complete identifier, rather than treating it as a synonym for DI. |
| DI | For US medical devices, FDA describes the DI as the mandatory, fixed UDI portion identifying the labeler and specific device version or model. | Use the DI–PI distinction to separate model identity from production context. |
| PI | For US medical devices, FDA describes the PI as the conditional, variable UDI portion identifying production attributes when included on the label—not every possible attribute in every UDI. | Identify the applicable production information before interpreting a label. |
| AccessGUDID | For the US device-identification system, FDA describes AccessGUDID as a separate, publicly searchable device-information database. | Use for public lookup; keep that activity separate from the labeler’s submission workflow. |
In practice, name the object involved in each task: the label identifier, its components, the catalog record or the public search result. Field-level nomenclature and the distinction between PI flags and PI values belong in the data-model discussion below.
04DI record content, PI flags and source-to-field mapping
For US medical devices, FDA explains that GUDID contains only the device identifier (DI) portion of the UDI, used as the key to retrieve device information. For US medical devices under 21 CFR Part 830, the GUDID device identifier record excludes production identifier (PI) values but includes flags indicating which PI attributes appear in the UDI. A presence flag describes an attribute used on the label; it is not the actual lot number, serial number or date.
For US medical devices, FDA’s UDI explanation identifies lot or batch number, serial number, expiration date and manufacturing date as label-side production attributes. Keep this DI-versus-PI distinction explicit in internal mappings: map the presence of an attribute separately from the production value printed for a particular device.
Field groups and their limits
For US medical-device information, FDA describes AccessGUDID lookup fields including device name, company name, MR safety status, Global Medical Device Nomenclature (GMDN) Term, Code and Status, and premarket submission numbers. These examples help organize the mapping discussion; they are not a complete submission schema or a universally mandatory field list.
For US medical-device DI records, FDA’s explanation of DI-record business rules describes rules intended to support data quality and directs readers to its Data Elements Reference Table for individual attributes. Use those attribute-level rules rather than assuming that an internal field name defines the destination field.
Example internal mapping register
The following is a practical governance example, not an FDA-prescribed form. Each row represents a mapping decision, not a production unit. The columns are internal controls, not FDA submission fields; source documents and owner titles will vary by organization.
| Field meaning | Internal source | Source revision | Transformation decision | Owner | Review evidence |
|---|---|---|---|---|---|
| Device name | Approved labeling master | Label revision reference | Direct copy or documented conversion | Labeling | Source-to-export comparison |
| Lot-number presence | Approved UDI specification | Specification revision reference | Derive presence, not lot value | Labeling | Flag-to-specification check |
| Premarket submission number | Controlled regulatory register | Register revision reference | Preserve approved identifier | Regulatory affairs | Reference-to-export comparison |
Common master-data practice keeps the approved source value and the export value separately traceable. If someone manually reformats or selects a value for export, record the transformation, its reason and the reviewer’s comparison with the approved source. Do not silently replace the source value with the transformed version. A useful mapping record lets a reviewer reconstruct which revision was used, what changed during preparation and why the exported value still represents the approved information.
Device Reference
Searchable device identity.
- Device identifier (DI)
- Device name
- Company name
Safety and References
Additional lookup attributes.
- MR safety status
- Premarket submission numbers
Nomenclature
Term and status for lookup.
- GMDN Term
- GMDN Code
- GMDN Status
PI Presence
Which production attributes appear in the UDI.
- PI flags
Label-Side Production Data
Label values, not GUDID PI values.
- Lot or batch number
- Serial number
- Expiration date
- Date a specific device was manufactured
05Ownership, human approval and the boundary of software support
For a US medical-device GUDID workflow, use a documented responsibility model that separates preparing data, approving content, transmitting it and resolving discrepancies. The following is a practical governance model, not an FDA-prescribed staffing or segregation-of-duties structure.
- Regulatory affairs assesses applicability, documents the labeler-status assessment and resolves regulatory questions.
- Product-data owners maintain controlled source attributes and investigate conflicts between source records.
- Labeling personnel reconcile proposed submission values with approved artwork and flag differences for resolution.
- A named approver authorizes the submission content after reviewing the proposed values and disposition of open questions.
- Submission operators transmit the reviewed content and record the result; they refer content discrepancies to the responsible owner rather than silently editing approved values.
A service-provider handoff can allocate these tasks, but should not substitute for the organization’s labeler-status determination. Specify who prepares, who approves, who operates the channel and who closes discrepancies. Where an additional check is useful, define its scope using a documented double-check approach, rather than treating another signature as proof of accuracy.
Make the handoff reviewable
Identify the source revision, intended DI record, unresolved questions, named approver and person responsible for closing the submission activity. Distinguish questions that block approval from follow-up work. If content changes after approval, return the changed content for review before transmission. Closure should document the outcome and disposition of outstanding issues, not merely that an operator sent data.
Software support is an internal design choice: controlled staging, value comparison, task routing and issue tracking can support this workflow. Keep submission-content approval and any separate product-release decision with designated humans. Neither a completed software task nor a successful transmission should replace those decisions.
06Submission routes, integrations and recovery from uncertain results
FDA-described submission and retrieval routes
For US medical-device labelers, FDA’s UDI overview describes requesting a GUDID account before submitting information. This is FDA explanatory guidance; the integration controls below are practical recommendations, not additional regulatory requirements.
For US medical-device labelers, FDA’s submission options are manual entry through the GUDID web application for individual devices, or HL7 SPL submission through the FDA Electronic Submissions Gateway for bulk uploads.
For public US device-reference data, FDA’s AccessGUDID overview describes downloading GUDID data and exporting search results. Treat that retrieval activity separately from labeler submission: a public-data export is an input to reconciliation, not a submission route.
An illustrative controlled integration
- Stage the intended content from controlled product and labeling sources. Retain source versions so the submission can be reconstructed without relying on today’s master data.
- Map source values to submission fields and inspect transformations. Keep the mapped content linked to its source; these are practical data-integrity controls, independent of the chosen submission route.
- Have a designated reviewer approve the staged content before transmission. Record who approved which version using attributable recording.
- Transmit through the selected channel, retaining the attempted content and available response information. Reconcile the outcome against that approved version and, separately, compare public-record values when available.
Classify exceptions before taking action
| Internal issue category | Recommended response |
|---|---|
| Content problem | Resolve the incorrect source value or mapping, then review the corrected content before another attempt. |
| Transmission uncertainty | Preserve the attempted content and available responses. Establish what is known about delivery and processing before authorizing a retry. |
| Public-record difference | Compare the retrieved value with the approved submission and source version. Assign an owner to investigate rather than immediately overwriting either record. |
These categories are internal triage labels, not FDA response states. Do not build recovery around assumed acknowledgment names, error codes, publication latency, automatic retries or rollback behavior. Where the outcome remains uncertain, document the unresolved question and route it to the submission owner. Control any retry against the preserved attempt so an investigation does not become an untracked sequence of submissions. Local file creation alone is not proof of completed submission. This workflow presumes no native connector, write-back API or particular system architecture.
- Scope ReviewAssess product classification and identify the labeler.
- OwnershipName the data owner and human approver.
- ReconciliationCompare DI, label and device attributes.
- ApprovalApprove the content and retain the decision.
- TransmissionUse web entry or HL7 SPL through FDA ESG.
- RecoveryResolve discrepancies and control retries.
- MonitoringCompare the public record and assess changes.
07Lifecycle management: assess changes before editing a record
GUDID maintenance commonly uses documented change-impact assessment and linked revision history. The stages below are organizational workflow stages, not official GUDID status names. They separate a proposed data edit from the regulatory assessment of its identifier implications.
- Proposed change: Open a change record when source information changes, labeling is revised, terminology needs maintenance, or a downstream user reports a discrepancy. Describe the issue before selecting a replacement value; a reported mismatch may originate outside the submitted record.
- Impact assessment: Identify the affected controlled source and revision, existing DI record, relevant label revision, terminology reference, and downstream consumers. Compare the previous approved value with the proposed value. Assess the data correction separately from whether the underlying device or labeling change has identifier implications.
- Approved revision: Have the designated reviewers document their decision and rationale. For an edit, link the approved replacement value to its supporting source. For no change, explain why the existing value remains appropriate and how the reported concern was resolved.
- Submission activity: Link the approved revision to the submission attempt and its available outcome information. Keep planned edits distinguishable from edits actually submitted; an internal approval alone is not evidence of an external update.
- Reconciliation: Compare the intended revision with the resulting record information. Assign unresolved differences to an owner rather than treating submission activity as closure. Include affected downstream consumers in the reconciliation where they hold copies or mappings.
- Closure: Record the reviewer’s conclusion, outstanding follow-up if any, and the relationship between the previous state, proposed state, decision and submission outcome. Preserve abandoned proposals as distinguishable from approved revisions.
For US medical devices, check the current text of 21 CFR Part 830 for scope when resolving a regulatory change question. This workflow does not establish new-DI criteria, update deadlines or retention periods.
Maintain the linked internal history through an audit trail and proportionate history review. Keep that history separate from whatever public information is currently displayed: the internal record explains the decision path, including proposals that never became published changes.
08Worked cases: production changes, obsolete terminology and uncertain delivery
These qualitative cases illustrate internal review practice, not legal determinations. Each separates the affected data from the decision and the evidence supporting it.
Production change: a different labeled lot number
Assumptions: the labeler and stated device version or model are unchanged, while the labeled lot number changes. Unit of analysis: the manufactured lot versus the DI reference record.
Apply the attribute classification above: assess the changed lot value against production evidence rather than treating it as a replacement DI-reference value. Review the lot assignment, executed batch record, label-generation inputs and label inspection results. Separately document whether any associated change has regulatory identifier implications; unchanged assumptions alone are not a categorical new-DI exemption.
Decision: route the lot-level evidence through production review and obtain a separate identifier-impact decision before changing reference data. Supporting evidence: approved lot assignment, label comparison and the documented scope of the change.
Terminology: an obsolete GMDN code
Assumption: a public lookup displays an obsolete GMDN status for the device’s recorded code. Unit of analysis: the nomenclature assignment associated with a DI record.
For the US FDA device-reference system, FDA describes AccessGUDID GMDN Code status values as Active or Obsolete. For US device labelers investigating terminology, FDA points to GMDN resources for identifying appropriate codes, including alternatives to obsolete codes.
Compare the selected term with the controlled device description, investigate alternatives and explain their fit. Do not guess a replacement or treat the status as proof of a completed terminology transition.
Decision: route a proposed correction for human review. Supporting evidence: the lookup capture, current assignment, device description, candidate-term comparison and reviewer rationale.
Uncertain delivery: transmission result unknown
Assumption: the team cannot determine the result of an attempted transmission. Unit of analysis: the transmitted DI-record content and its attempt history.
Compare the intended content with the retained transmission copy, available response information and observed record state. Preserve the audit trail linking those observations to the attempt. Neither silence nor an inconclusive lookup establishes an FDA processing result.
Decision: investigate unresolved discrepancies before choosing correction or retry; record why the selected action is justified. Supporting evidence: content comparison, attempt history, response information, record-state observations and the reviewer’s recovery decision.
09Common data failures, their causes and targeted controls
For a labeler’s GUDID workflow, use this table as a practical diagnostic aid, not a list of mandatory controls or measured FDA findings. The weaknesses are possibilities to investigate, not established causes. Apply the integration and change-control processes covered above to implement corrections.
| Failure symptom | Likely process weakness | Targeted control | Evidence of closure |
|---|---|---|---|
| Submission preparation uses a stale source export. | Export selection may rely on filenames rather than controlled revisions. | Reconcile export revisions against approved source revisions before reuse. | Dated reconciliation identifying the replacement export and resolved differences. |
| Spreadsheet transformations alter intended field values. | Formulas, lookups or manual substitutions may bypass review. | Review transformation logic and compare affected outputs with source values. | Reviewed logic version and documented input-to-output checks. |
| A record stalls between teams. | The receiving owner may never have accepted the handoff. | Record explicit acceptance by a named owner with an agreed next action. | Accepted handoff linked to the completed action. |
| An issue is closed but corrected content remains wrong. | Closure may depend on task completion rather than content verification. | Compare corrected content with the approved intended values. | Reviewer-recorded comparison identifying the content checked and result. |
| Downstream users retain an outdated internal extract. | Replacement notices may not reach all consuming teams or jobs. | Identify consumers, communicate replacement and confirm adoption. | Replacement notice plus consumer confirmation or job logs showing the revised extract. |
Connect recurring handoff and revision problems to master-data governance. Use a data-integrity risk assessment to prioritize follow-up where incorrect values could propagate. Keep closure artifacts linked to the affected record and issue so reviewers can trace the correction without reconstructing email chains.
10Implementation checklist and evidence for an internal readiness review
For a US medical-device GUDID workflow, use this practical checklist to assemble a reviewable decision record. It supports internal readiness, not FDA certification or assured inspection acceptance. Refer to the earlier scope, ownership, mapping, submission and lifecycle sections for how each control operates.
| Review question | Internal evidence | Unresolved action |
|---|---|---|
| Is scope and responsibility documented? | Product scope, labeler assessment and assigned ownership | Assign unresolved applicability questions. |
| Are the intended sources approved? | Approved label revision and approved data snapshot | Resolve conflicting or unsuitable sources. |
| Are field selections traceable? | Field mapping decisions and unresolved questions | Assign ambiguous mappings for review. |
| Was the submission content approved? | Submission content linked to human approval evidence | Reconcile differences from the approved snapshot. |
| Is the submission outcome understood? | Available responses and discrepancy disposition | Assign uncertain outcomes for investigation. |
| Was public-record follow-up completed? | Public-record comparison or documented pending check | Assign discrepancies or pending follow-up. |
| Can changes be reconstructed? | Change rationale linked to record history | Restore missing decision links. |
Keep two judgments separate: whether the record matches its intended source, and whether that source is appropriate for the device and field. An exact match to an outdated label revision answers only the first question. A technically successful transmission is not the internal review conclusion.
Record a disposition for the review, with the approver, rationale and an accountable owner for each unresolved item. Link subsequent resolutions through change control and the relevant audit trail. Keep pending actions visible rather than treating submission completion as closure.
Assess retention and authorized access through the organization’s applicable record-governance process; this checklist prescribes neither a retention period nor an electronic-signature rule.
Can another authorized reviewer reconstruct what was submitted, why it was selected and how the outcome was resolved?
Frequently asked questions
Q.How can a company decide whether it is acting as the labeler rather than only a distributor?+
For US medical devices, FDA’s labeler explanation considers who causes a label to be applied or modified. Under the US FDA UDI framework for medical devices, adding only a distributor’s name and contact information does not constitute label modification for this test. A practical assessment examines the intended final labeling arrangement, artwork authorization and instructions to packaging contractors rather than relying on commercial titles. Document the assessment within UDI governance, and use change control to revisit it when labeling responsibilities change.
Q.How should a low-volume labeler choose between web entry and bulk submission?+
For US medical-device labelers, FDA describes manual entry through the GUDID web application and bulk HL7 SPL submission through its Electronic Submissions Gateway. A practical choice considers record complexity, update frequency, available technical support and review effort—not volume alone. Web entry may suit occasional, straightforward changes; bulk submission may suit reusable structured data. Compare preparation, correction and maintenance effort, then document the chosen workflow and reviewer responsibilities within master-data governance.
Q.What should a hospital check before using a public device record in its item master?+
For US medical-device information, FDA describes AccessGUDID as searchable by identification details including the label DI, device name and company name. A practical hospital intake check compares the physical package identifier, model description and purchasing unit with the candidate record before updating material master data. Investigate packaging-level mismatches rather than matching on brand alone. Keep DI and PI distinctions explicit in local mappings, and retain the lookup date and reviewer’s decision.
Q.How should discrepancies between a supplier file and an approved label be resolved?+
A practical reconciliation workflow pauses the affected data change and identifies the exact disagreement: identifier, model, packaging description or another attribute. Compare the supplier file and approved label with controlled product specifications and revision history; neither file age nor apparent completeness settles the issue. Assign an owner to investigate whether the label, supplier mapping or submission needs correction. Record the decision through a deviation workflow, preserve data integrity, and obtain human approval before publishing the corrected mapping.
Q.What information should be handed over when a submission service provider changes?+
A practical handover package includes the DI inventory, field mappings, source-document references, submission files, response messages, unresolved errors and pending changes. Include authorized contacts, account roles, transfer instructions and the distinction between submitted and accepted records. Transfer access through an authorized process rather than sharing personal credentials. Preserve attributable recording and the relevant audit trail. The incoming provider and labeler should reconcile the inventory together, with named human reviewers approving the handover and subsequent data changes.
Q.Can a missing public search result establish whether a device is exempt from UDI?+
No. Treat a missing result as an unresolved lookup, not an exemption determination. A practical investigation checks the identifier transcription, search method, packaging level and supplier’s explanation before drawing conclusions. Keep the search finding separate from the regulatory assessment: absence from a result list does not identify the legal basis for an exemption. For US medical devices, check current UDI provisions for scope and document the applicability assessment alongside the product’s medical-device classification.
Q.How should a mixed device, drug and supplement portfolio separate its data-governance work?+
A practical governance model uses shared controls for ownership, versioning and approval while maintaining separate product-specific data dictionaries and applicability decisions. Assign each marketed item to its relevant regulatory workflow before mapping identifiers or publication fields. Keep device DI work distinct from drug and supplement product records and from transactional batch records. Where systems share a product master, use explicit product-type fields, responsible reviewers and controlled interfaces so a change in one portfolio does not silently overwrite another.
Q.What should trigger a review of an older GUDID operating procedure?+
Practical review triggers include changed submission channels, service providers, account responsibilities, product families, field mappings or recurring rejection patterns. For US medical-device GUDID records, FDA explains that attribute-level business rules appear in the GUDID Data Elements Reference Table. Compare the procedure with current operating references after relevant system notices. Test whether staff can still execute its instructions, update responsibilities and examples, and document any resulting instruction through the training record rather than merely changing the procedure’s revision date.
Primary sources
Want to see how GUDID could fit into your own records and workflows? Explore the related V5 pages or talk to our team about what applies to your operation.
