The medical technology sector has the dubious distinction of having enormous amounts of data being created in relation to their use, but few tools to usefully analyze it. This has been highlighted by a difficulty to quickly identify and understand the causation of problems, such as with metal-on-metal hip replacements and breast implants with defective materials.

This is now changing, both through action from regulators and support by device manufacturers and analytical companies introducing tools to better understand the outcomes in the use of devices.

FDA’s introduction of Unique Device Identification (UDI) in 2014 has led the way in creating better standards in master-data sets in their Global UDI Database (GUDID), to unlock previously hidden ‘signals’ in the data, which is collected from patient records and device adverse-incident reports.

The volume of data collected by healthcare providers is enormous, with more than 2 million different medical devices recorded so far in the U.S. market alone, and the FDA MAUDE database recording more than 1 million reportable adverse events with medical devices in 2019.

The problem with medical device data is not the collection of data but having the necessary tools to analyze the results in a meaningful way. That is why the FDA chose to specify a nomenclature to enable the accurate naming and grouping of devices. The decision to use the GMDN as their nomenclature was not done in isolation, because the FDA followed recommendations from the International Medical Device Regulators Forum, which brings together countries interested in regulatory harmonization. Other countries including Canada, Australia and New Zealand, the Russian Federation, and the UK are observing the progress being made in the United States in its use of UDI and nomenclature to collect better data and make better healthcare decisions.

New data analysis techniques using a nomenclature is providing regulators, healthcare providers and manufacturers with new ways of understanding different types of problems associated with devices and also more importantly an indication of the route cause. For example, by grouping adverse event data based on a group of similar devices and by manufacturer, we can quickly identify if the problem is manufacturer specific or maybe more systematic and therefore there may be an inherent design problem to be investigated. Alternatively, if we can isolate a higher rate of incidents with a group of devices to a particular location, maybe the problem is related to how the device has been used, and thus we may need to focus on better training for clinicians at a local level.

Using a nomenclature to group devices has already been recognized as adding value by the Learning UDI Community (LUC). The LUC was created by the FDA and Advancing Health Care through Supply Chain Excellence (AHRMM), the U.S. membership group for health care supply chain professionals. The LUC’s 2019 clinician lead report on ‘High-Risk Implants’ concluded that “Use of the GMDN term assigned to UDI-DIs and their associated implantable collective codes supports the most accurate programmable approach to identifying implantable devices.” Another review on a study on implants records at the NHS Breast and Cosmetic Implant Registry concluded, “It appears that the GMDN data would be extremely useful in building further submission validation for the registry, which we could explore once we are in a position to develop the registry further”.

Inconsistency in collecting and identifying devices held in implant registries is currently a major challenge. Using a national device database (UDI Database) that has both the UDI and a nomenclature included will help to eliminate this problem. When a product that has a UDI on its label is scanned, systems can quickly retrieve all the information about the product, just like when we are at the check-out doing our weekly shop, removing many of the ‘keying in errors’ and reducing the administrative burden on our clinical staff when entering data about the product into the implant registry.

Consistency also improves if the national database is a single source of device data shared by all the possible users. Data accuracy can then be monitored at a single point rather that each hospital trying to keep thousands of product records up to date.

As data becomes more available, analysis tools are picking up on the utility of the data to refine searches and group devices in a more consistent way, such as the Reed Tech Navigator for Medical Devices and the Symmetric Medical Device Database applications. Both these products take data from the U.S. GUDID to identify and group devices more consistently.

Analysis of data should not be limited to identifying safety-related matters, but it should also support other needs of healthcare providers. The nomenclature should be able group devices in different hierarchies depending on the user’s requirement. For example, not only grouping devices by name or function, but also by principle material or power source. This is very useful when rationalizing hospital inventory, organizing maintenance schedules and improving utility.

In times of national emergency, it is important to be able to quickly and accurately identify shortages of specific equipment and excess stock of supplies, so they can be reallocated to the greatest need. Long-term planning of healthcare resources and preparation for unexpected demand is an important activity that depends on using a nomenclature to accurately sort devices into clinical groups. International aid charities, like Médecins Sans Frontières, now use this more accurate identification to better prepare for the clinical emergencies they support worldwide.

How close are other countries to getting a national medical device database?

In the UK, the Independent Medicines and Medical Devices Safety Review headed by Baroness Cumberlege, has sought evidence to improve medical device safety and has strongly indicated that improvements in data collection will be an important recommendation when their report is expected to be published in early 2020. Written evidence to the Review, submitted by NHS Digital stated “In short the speed and affordability of set up of registries will be greatly helped by standardisation most notably the use of common data standards, which could enable complex data sets needed for a comprehensive and granular registry to be built. In addition when things go wrong there needs to be a way of tracking the supply chain using GMDN…”

In predicting the outcome of Cumberlege, the newly published UK ‘Medicines and Medical Devices Bill’ has already introduced the possible provisions about “requiring information in relation to a medical device to be entered in a register”. As part of the legislation supporting Brexit negotiations, the UK seems prepared to move quickly on this important initiative in time for the end of the transition period, if agreement cannot be made with the EU by January 1, 2021.

The introduction to better data collection in the EU region seems hampered by the delays in providing the EUDAMED database for data entry until mid-2022 at the earliest and in the creation of a brand-new nomenclature based on a variant of the Italian national classification system, CND. Not only will this mean meaningful product data may not be available from EUDAMED until the publication modules are completed, but also use of a specially created nomenclature will limit opportunities to exchange vigilance data with other countries and regions, most of which are now using the GMDN.

Optimistically the introduction of an UDI requirement within the EU will be a welcome improvement, although global trends in medical device production already predict that manufacturers are seeking to create globally harmonized packaging to include the UDI. Technical standards are also seeking to reduce the impact of diverse national requirements on device labels, symbols and instruction for use, such as e-IFU initiatives.

One important development on the horizon is the growing need to understand more about device performance using real world evidence. Regulators and healthcare providers are seeking more assurance of device effectiveness as healthcare technology assessment becomes an emerging requirement. Finding sources of reliable data to continuously support claims of compliance by manufacturers is expected to be in big demand in the future. Publishers like Elsevier, specializing in a bibliographic database of published literature, has recently announced its use of nomenclature data to provide important clinical and vigilance information on medical devices.

It seems that the use of better data to make better decisions is finally being realized.