This repository is an assembly of tools to automatically pull FDA 510(k), De Novo, and PMA summaries from the FDA's website for easier parsing and manipulating the data therein.
In this repository, there are classes for easy scraping FDA 510ks, De Novo, and PMA devices based only on a Submission Number (eg "K000000", "DEN000000", or "P000000"). In addition, there are helful functions for downloading the full list of FDA 510(k) devices as databased by the FDA and updated on a monthly basis. Combined, these two options allow a user to conduct rudimentary FDA predicate device searches in a near-automatic fashion. Below, I walk through a small tutorial on how to employ this repo to conduct pointed predicate-device searches.
In writing this repository, it is apparent that one could use this code to querey the FDA database to identify all 510(k)'s and De Novo's and build a large database of all ~100,000 cleared and granted devices. While this was not the intent of the code, it would be irresponsible if I did not acknowledge this possibility and council against it. The FDA CDRH provides this information at no cost, and it is a useful tool for medical device developers and companies. It would be a loss for the greater community if such information was put behind a firewall, paywall, or other rate-limiting measure. If you are interested in routinely and aggressively querying the FDA's servers, please be mindful to not overly burden the FDA's resources and website.
- tbiscof's Github "FDA" repository - Really cool predicate tracing methodology.
- McClain98's Github "FDAexplorer" repository - FDA database download and manipulation.
- JustIceQAQ's Github "Downloads_510K_Data" repository - FDA database parsing.