UDIs in recalls and beyond

I have been researching and writing about the value of unique device identifiers (UDI) for nearly 15 years, but it really all comes down to this: Call the same thing the same thing regardless of what you are doing with the thing.

For those who have not been with me on this nearly 15-year journey, the U.S. Food and Drug Administration, UDI regulation requires manufacturers to assign and label their products with UDIs, and publish additional information about those products in the FDA’s Global UDI Database (GUDID). The UDI itself is made up of two parts: 1) the device identifier (UDI-DI) representing a specific version or model of a product at a particular unit of measure, and 2) the production identifier (UDI-PI), which includes data related to production of a specific device, e.g., the lot, batch or serial number, expiration date, etc.

The ability for multiple stakeholders (manufacturers, distributors, and hospitals, etc.) to identify products in a consistent manner across the product lifecycle offers myriad benefits from better patient safety to greater supply chain efficiencies and cost savings. When the rule was published in 2013, the primary focus was the ability to identify and manage devices associated with adverse events and subject to recalls more quickly, thereby getting potentially dangerous devices off the market faster and making the lives of patients safer. (See Standard Practices, July 2019).

The good news: The percentage of medical device recall notices containing UDIs has grown nearly 10-fold since 2016, the year the highest risk devices (Class III) were first required to bear UDIs. Deadlines for less risky devices have since followed. The bad news: Still less than 20 percent of those notices (and less than one-quarter of those for implantable devices) contain UDIs today.

This caught the attention of the Association of Healthcare Resource & Materials Learning UDI Community (LUC), which launched a workgroup to study the value of UDIs for recalls, and develop recommendations to achieve those benefits. After two years of work by more than 70 representatives from across the healthcare supply chain, the group has published a comprehensive report at www.ahrmm.org/luc. Here are a few key findings and recommendations.

Highly manual, variable, and error prone

The workgroup found significant variation in not only the information included, but also the process by which recalls are communicated and managed. Most of that information is shared in paper-based formats, which require manual (and in turn, error prone) entry by stakeholders into multiple technology systems. As a result, the ability to find and remove recalled products from the market can be delayed.

Manufacturers have a variety of means to report recalls to the FDA, but most are handled via emails using pdfs and excel documents. Across those submissions, there is variability in the data shared, including UDIs.

Manufacturers notify providers almost exclusively via mail or delivery service, e.g., FedEx, although some mostly fee-based systems handle this for subscribing hospitals and health systems.

UDI data is needed across multiple systems by multiple stakeholders

Currently, when manufacturers issue recall notices, they blanket the market (both distributors and healthcare delivery organizations). When providers receive such notices, they often have to query multiple systems: enterprise resource planning (ERP) and other procurement software, inventory management systems and electronic health records (EHRs), to determine if they bought the product, if they have the product on hand, and/or if they used the product in patient care.

The challenge is: UDI is just one of the codes that the FDA says manufacturers can use to identify the product being recalled, and as a result, manufacturers often choose a proprietary number, such as a catalog number. Providers, on the other hand, sometimes create their own proprietary numbers to document products in their respective technology systems.  As a result, when it comes to looking to see if they bought, stored or used the devices, they may not have the corresponding identifier in their systems. This can critically delay their response, creating the potential to use the devices in patient care and making it hard to track down patients who may have a recalled device implanted in them.

Providers are also only mandated to record UDIs for implantable devices in patient records as part of the U.S. Core Data Set, which means they may or may not have record of all recalled products used in patient care.

Beyond the patient safety aspects, this lack of a consistent identifier makes recall management a time consuming and expensive process.  Based on real world examples, the workgroup found that managing a single recall, depending on the kind of device, can cost a provider from $480 to more than $3,000 per item recalled.

Recommendations

A primary recommendation of the workgroup is to urge Congress and the FDA to create a structured electronic recall process leveraging the data and process used by the GUDID. By using APIs and/or downloading capabilities similar to those used for the GUDID, data about recalled products could be electronically populated in distributor and provider systems, while the FDA could stand up a searchable database as a single source of information about all recalled products.

Currently, the US FDA manages nine different sites where data on medical device recalls is stored, often containing different information. The manual processes involved can also delay posting recalls in these systems from weeks to even months. Leveraging the GUDID, standardizing how recalls are reported (using a standard template and data elements) and enabling automated queries could save time for all involved, not to mention supporting the needs of patients.

The success of such a system requires coordination among a variety of parties in the healthcare ecosystem:

• Manufacturers should use the full UDI (both DI and PI) in their recall notices.
• Distributors and providers should also leverage the UDI in their respective systems to know if they bought, are storing and/or used a product in patient care. Fortunately, many ERP and EHR systems can hold the UDIs, especially with the increasing prevalence of cloud-based systems.

Creating an electronic medical device system and changing processes comes at a cost, but based on the workgroup’s research, there is a sizeable ROI for all involved. To help organizations estimate their own savings, the LUC workgroup has created cost calculators for each of the key parties: manufacturers, distributors and providers.  If you consider that there have been an average of 2,900 recalls per year since 2014, even if a provider only handled one-quarter of those recalls (725) and at the lowest estimated cost ($480), the annual savings would be nearly $350,000. Using a prototype medical device recall system developed by the work group, distributors estimated they could reduce time to handle individual recalls from 50 to 30 hours, while manufacturers attribute their biggest costs to the notification process and provider response time, both of which can be significantly streamlined with use of standard data and automated processes.

But as with so many things in healthcare, the real bottom line cannot be measured in dollars. The most important and compelling reason is our collective ability to protect patient safety through better visibility, and faster response to adverse events and recalls. We will be able to minimize any negative impacts of care already delivered as well as the number of patients who could otherwise be treated with a faulty device. That’s delivering value in healthcare.