Senior Advisor for Patient Safety
Food and Drug Administration
Establishing a UDI System
Combination of 4 distinct steps – 4 panels:
- Develop standardized unique device identifiers (UDI)
- Place the UDI in human readable and/or AutoID on a device, its label, or both
- Create and maintain the UDI Database
- Adoption and Implementations
Panel 1 – Developing the UDI
- Develop UDI code according to GS1, HIBCC, NDC
- Created and maintained by the manufacturer
- Concatenating Device and Production Identifier
- Device Identifier: [static] Manufacturer, make, model [i.e., each catalogue number]
- Production Identifier: [dynamic] if currently serialized – serial number; if currently identified at the lot, the lot number, and expiration date
Panel 1 – Questions
- Created according to what standard(s)? How do we identify these standards? Should it be limited in any way?
- The device identifier must unambiguously and uniquely identify a device – how do we convey this concept?
- For changes to a device, when is a separate (new) device identifier necessary/desired/required?
- Should anything else be included in the UDI (e.g., quantity, GMDN code)?
- Are there devices types that are controlled in some way other than lot or serial number – and if so, what are they – and should we allow that?
- If the UDI is on the label – what information, if any, could be removed from the label?
- Which devices or device types, if any, should be excepted? From which parts (all or some) of the regulation?
- What should be the process for granting exceptions?
- Are there devices that are distributed in kits or are part of a combination product that otherwise would NOT be distributed – and therefore would
- NOT have a UDI? Should we require a “virtual” UDI?
- How do we handle the reuse of single use devices?
- How should components be identified?
- What about devices sold in retail (which requires a UPC code)? What if the same device is distributed both in retail and health care facilities?
Panel 2 – UDI Application
- Applied at all levels of packaging, down to the lowest level (the patient use level or unit of use)
- Human readable and/or encoded in a form of automatic identification technology
- Direct Part Marking (DPM) for some devices
- No specific technology would be identified (technology neutral)
- Identify a series of standards (linear barcode, 2-dimensional barcode, RFID)
Panel 2 – Questions
- The default location for the UDI is on the “label” of the device – is this model sufficient to identify a device at the “unit of use”? Are there devices where this model doesn’t work?
- Are there device types where it is not necessary to identify the individual, labeled device with the full UDI? If not – where does the UDI go – how do we describe its location? Should part of the UDI be on the individual device?
- Is there anything about combination products that should be addressed?
- How should UDI be applied to kits? Convenience packs?
- How should UDI be applied to devices that have no physical “label” – such as software?
- Should we identify specific automatic-identification technologies that should be used? If so – which one(s)? If not, how should auto-id be addressed?
- If we go the standards route – which standards should we identify (linear barcode, 2-dimensional barcode, RFID)?
- Are there devices for which human readable is not necessary (auto-id only)? Are there devices where auto-id is not necessary (human readable only)?
- What if the device or its label carries other forms of auto-id (for purposes other than UDI) – how do we distinguish the UDI?
- Which devices require the UDI to go on the device itself – direct part marking (e.g., for devices where the label is ordinarily removed when the device is distributed or used)?
Panel 3 – UDI Database
Minimum Data Set for each Device Identifier:
- Manufacturer, make/model (catalogue number)
- GMDN/UNSPCS Category/code
- Control mechanism
- Packaging level/number of items
- Country of origin/manufacture
- Labeled as single use or reusable
- Contains known, labeled allergen (e.g., latex)
- Storage conditions (e.g., needs to be refrigerated)
Panel 3 – Questions
- Are there data elements that are mandatory (for all devices); others that are device specific; any that should be voluntary?
- What other data elements should be included?
- Is there any data in the database that should not be made publically available?
- Can the UDID manage change of ownership? Country of origin?
- What is the role of GMDN in the UDID?
- We envision that this list would evolve over time – how do we manage that process?
- Who is responsible for entering and validating the data?
- How do we obtain international harmonization of the data elements (e.g., sterile, latex)?
Panel 4 – Adopt and Implement
- Improve effectiveness and efficiency of recalls
- Adverse event reporting and postmarket surveillance
- Tracking and tracing, supply chain security; and anti-counterfeiting/diversion
- Disaster/terror preparation and shortages/substitutions
- Systems to reduce medical errors
- Assisting clinicians in identifying appropriate device
- Documenting medical device use in patient’s EHR/PHR, hospital information systems, claims data
- Sentinel Initiative - strengthening FDA’s ability to query data systems for relevant device information
In order to accrue the kinds of benefits we envision, the UDI needs to be adopted and implemented by all stakeholders:
- What other benefits can be accrued from UDI?
- How do we facilitate adoption and implementation? For distributors? For hospitals? For other health care environments? For home care?
- What are the obstacles to implementation?
- How can UDI be used to perform device recalls more efficiently and effectively?
- How can UDI provide visibility into the supply chain?
- How can UDI help with traceability? Anti-counterfeiting? Import Safety? Reimbursement? Clinical effectiveness? Cost effectiveness?
- What is FDA’s role?