FDA develops software to select medical device inspection tasks
2019-01-11 10:06:42
Recently, the Biostatistics Division of the FDA’s Center for Devices and Radiological Health (CDRH) developed a new inspection software for selecting medical device manufacturers for inspection tasks. The inspection software is currently in the pilot phase.
The software tool is part of the FDA’s efforts to strengthen product life cycle supervision and will change the way medical device manufacturers are selected for on-site inspections and their frequency from the Biological Investigation Monitoring (BIMO) program. Adoption of this tool will improve FDA’s resource allocation—an area emphasized by recent legislative actions (such as the 21st Century Act) and can also save a lot of time and money for the medical device industry and FDA staff.
According to BIMO statistics, CDRH conducted a total of 287 domestic inspections in 2017, up from 270 in 2016 and down from 301 in 2015. The main challenge facing the current process is the lack of sufficient resources to conduct on-site inspections of all companies found to have large deviations in the study. According to CDRH officials, the current method of selecting on-site inspection sites is mainly based on manual verification or visual inspection of descriptive factory statistics, which is very difficult for a large number of manufacturers and variables, and lacks rationality and rigor.
The CDRH Biostatistics Department has proposed a technical solution to statistically test each valuable variable in the on-site inspection enterprise selection process, including the size of the enterprise and the number of related serious adverse events. Then two different methods, Fisher combination method and LRT method, are used to translate the results into a matrix, and finally a concise and clear enterprise ranking is presented. This software tool is called BIMO LRT Inspection Statistical Software (BLISS).
The director of the CDRH Biostatistics Department said that the FDA is currently using simulated data sets to verify this tool and use it in some actual medical device clinical research cases. It is expected that this work will have a direct impact on pre-market and post-market review work. Next, the FDA will further develop indicators that can measure the performance of BLISS, hoping to eventually obtain companies that need to be inspected on site more quickly through automated means.
