openFDA Unveils Cache of Medical Device Data

Taha A. Kass-Hout, MD, MSOpenFDA is releasing information on medical devices that could spur innovation and advance scientific research.

OpenFDA’s Application Programming Interface (API) expands the previous openFDA resources about medical device-related adverse events and recalls by incorporating information from the total medical device product life cycle. This includes current data on device classification (6000 records), 24,000 registrations of device companies and establishments, and the companies’ listings of over 100,000 devices. Data since 1976 on 30,000 device approvals and approval supplements, and 141,000 device clearance decisions (510(k) and de novo types) are now on openFDA. In addition, more details about device recalls (9,500 records going back to 2002) and adverse event reports (4.2 million records since 1991) were added. Although this information has been available on our public databases for many years, now developers can more easily access and use the data.

The flexible openFDA interface functions well even when greater demands are made on it and is designed on a common platform so developers can harmonize and integrate data from various sources and build their own applications. For example, they could develop a smartphone app to determine all the recalls associated with a particular class of devices or find all companies that manufacture certain types of devices.

Roselie A. Bright, Sc.D., M.S., P.M.P.There are some important safeguards. For example, the information doesn’t contain anything that potentially could be used to identify individuals or reveal confidential commercial information.

Everything available in these datasets should be used in the appropriate context. FDA has harmonized the data, but there may be instances when a query does not return a full and complete result. For example, if the name of a manufacturer is listed with different spellings, some variations may not be captured in the result.

Some datasets are snapshots in time. For example, the 510(k) data shows who submitted the 510(k), the device name and other information at the time of clearance. In addition, the types of information that FDA has collected has changed over the years, which can make it difficult to look at data over time. Also, the data may not have enough information to establish cause and effect, incidence, nor prevalence.

Part of openFDA’s mission is to engage the public to advance the understanding and use of the data. This API is the latest in a series of openFDA releases that have made publicly available data more easily accessed and queried. We believe these tools can be used by developers and researchers to make insights that fuel new, innovative products that help protect and promote public health.

Ann FerriterBy design, openFDA is a research and development project that draws on community involvement. We are actively involved in the openFDA communities on GitHub and StackExchange, and we encourage everyone to participate in those communities.

Over the last year, there have been dozens of tools created using openFDA resources. We hope this enhanced devices data will be put to similar use. Together, we can make openFDA an even more useful, more powerful resource for all.