FDA Proposed UDI: The GUDID Database

The FDA is proposing the creation of a Global Unique Device Identification Database (GUDID), a publicly accessible database that would hold information about each medical device marketed in the U.S.  The GUDID has some similarities with GS1’s Global Data Synchronization Network (GDSN) but rather than accepting the use of GS1’s standard and service, they appear to have decided to roll their own.  Is that a good idea or will it just result in unsynchronized duplication?  Is there a way to use GDSN to serve GUDID (or vice versa)?  Could the GUDID concept work for drugs too?

This is the third in a multi-part series of essays examining various aspects of the recently proposed Unique Device Identification (UDI) rule by the U.S. Food and Drug Administration (FDA) (see the two earlier essays in this series, “FDA Proposed UDI: A Revolution In Number Assignment” and “FDA Proposed UDI: AIDC Requirements”).


The FDA GUDID would be a publicly accessible database that holds basic information about each device on the market, now or in the past.  It would be keyed by UDI so that supply chain members, healthcare professionals, regulators and patients would be able to look up a device to quickly find certain information about it.  The GUDID, as envisioned by the FDA in their proposal, would be “FDA-administered”.

Here is a summary of the type of data elements that the FDA is proposing.  See the FDA proposal for full details.

  1. The device identifier portion of the UDI
  2. Any previous device identifiers
  3. A permanent device marking statement where required
  4. The proprietary, trade, or brand name of the device as it appears on the label
  5. Any version or model number or similar reference that appears on the label
  6. If the device is labeled as sterile, a statement to that effect
  7. A natural rubber latex statement
  8. Size information if applicable
  9. The type of production identifiers that appear on the label
  10. The FDA premarket submission number of an approved or cleared device, or a statement that FDA has by regulation exempted the device from premarket notification.
  11. The FDA listing number assigned to the device
  12. The Global Medical Devices Nomenclature (GMDN) code for the device. (GMDN is a comprehensive system of generic descriptors with definitions used to generically identify medical devices.)
  13. The number of individual devices contained in each device package


The specification of the GDSN family of standards is contained in a set of documents that are available on GS1’s website.  The purpose of GDSN is to allow companies in any given supply chain to exchange various kinds of data using a publisher/subscriber approach.  The types of data that can be exchanged with GDSN include party and product data.  Each major supply chain seems to have their own type of data that GS1 has documented in an extension specific to that industry and that includes healthcare.  Manufacturers and repackagers must publish their data so that downstream trading partners can subscribe to that data.

GDSN is a fee-based service where both publishers and subscribers must pay for their use of the system.

At a high level GDSN has an elegant design but when you get down to the field level it can seem a little inflexible.  That is, if the data you need to exchange isn’t already included, about all you can do is join the GS1 GDSN groups and ask them to add it.  See the GDSN Trade Item Implementation Guide – Healthcare for the healthcare-specific extension data elements allowed and the GDSN Healthcare Use Cases.

The primary key for publishing and looking up product information in GDSN is the GS1 Global Trade Item Number (GTIN).  That is, all products in GDSN must have a valid GTIN assigned to them first.  GDSN is capable of carrying basic and detailed product information and it can also carry the full packaging hierarchy as expressed in GTINs.  The number of potential data elements is quite large and is too extensive to list here.  It includes general physical, logistics and retail information as well as healthcare-specific information.

Last December GS1 published a position paper called “Global standards pave the way for Unique Device Identification (UDI)” which states “The GDSN can be leveraged by manufacturers as a data feed to the UDI databases”.  I believe they mean that when manufacturers publish information to a GS1-certified GDSN data pool for supply chain purposes, there will eventually be a way for that data to be automatically posted into the FDA GUDID database on their behalf.

At first glance it seems like GS1’s GDSN might be a good candidate for fulfilling the requirements of the proposed GUDID, but even though GDSN shares some similarities with the plan for the GUDID, it would have been a poor choice for FDA to adopt it as the sole repository for UDIs.  Here are the reasons I can see, perhaps you can see more:

  • GS1 will likely only be one of multiple UDI “issuing agencies” so the GUDID database will need to be able to support non-GS1 UDIs.  That’s probably not in the future plans for GS1’s GDSN service;
  • FDA says that they will administer the GUDID, which means they intend to exercise control over its contents.  For example, the FDA says they “…would reject or remove information submitted to the GUDID […to] prevent misuse of the GUDID for purposes other than those that underlie [the] rule and would help ensure the accuracy and reliability of information in the GUDID.”  I don’t think GS1 or their affiliated data pools would be able to provide the FDA with that level of authority over the data published by manufacturers on GDSN;
  • The contents of the FDA GUDID will be publicly available.  GDSN is a private, for-fee network of repositories that aren’t set up to assume partial exposure to the public;

GDSN is intended to assist commerce within all kinds of supply chains.  The GUDID is intended to assist the FDA, medical professionals, members of the supply chain and patients gain access to common information about medical devices.  These purposes have some similarities, but the differences are great.  GS1’s idea of enabling GDSN to feed manufacturer data into the GUDID on their behalf seems like a good approach.


UDI would be a standardized way of identifying medical devices that is similar to the way the NDC has been used for 40 years to identify drugs in the U.S.  The GUDID would be a central database that holds information about medical devices at the product class level.  The Food and Drug Administration Amendments Act of 2007 (FDAAA)–the law that required FDA to setup a medical device information data repository–also required the FDA to set one up for drugs.  With little fanfare the FDA has done just that.

That repository is called the National Drug Code Database and it is available today.  It can be used to search for drugs by NDC, proprietary name, application number, labeler name or active ingredient.  Try it yourself with your favorite NDC.  You can also download the entire NDC database in Zip file form (here it is if you want it).  I assume this is similar to what would be offered by the GUDID.

When you look at the NDC database you realize that the GUDID doesn’t really need to be very complex to be useful, but without that single national (global?) identifier it would be more complex and probably much less usable.  So the magic of the GUDID is in the UDI.

Could these databases be laying the foundation for national/global serial number-based track and trace repositories for devices and drugs?  No.  That kind of repository would be orders of magnitude larger and more complex with performance requirements that would rival the credit card servicing repositories.

See you at the FDA’s UDI Conference.  I’ll be there, but only for the first day.


2 thoughts on “FDA Proposed UDI: The GUDID Database”

  1. Peterson’s Law: “Any government intervention into peaceful, private activity tends to make things worse rather than better.” I tend to doubt that government ideas, projects are better than what the market can produce. The problem here is that the market wasn’t producing anything. Lots of talk but little action. I suspect if the market had developed the database, implemented it and made it available the gov’t would have not made the move. No one to blame for this but ourselves.

    Having said that, anyone that’s been involved in EDI with pharma knows how easy the NDC made identifying drugs. Never a doubt or problem with that. Remember the fiascos with DEA, HIN, etc. for customer identifying? Still no concensus on what to use and still problems!

  2. FDA has the mission to protect public health, not GS-1. I have limited faith in the abilities of either government or non-governmental organizations, GS-1, incuded. Here, I trust government more with the data.
    The questions repeatedly asked about the cost of GS-1 have never been adequately addressed. I hope industry continues to ask those questions.
    An NGO in a dominant role is an invitation to excessive costs and ineptitude. How did Fannie and Freddie work out for the economy?

Comments are closed.