LAUSR

Study Objective To evaluate the accuracy of the postmarket surveillance process in the MAUDE database for devices approved via the Food and Drug Administration (FDA) premarket (PMA) and 510(k) approval processes and used in Obstetrics and Gynecology, comparing the accuracy of death and injury reports. Design A retrospective observational study. Setting Medical devices used in Obstetrics and Gynecology. Patients or Participants Two hundred and nine product codes encompassing over 12,000 FDA approved devices. Interventions n/a Measurements and Main Results Death and injury reports were collected from November 1, 2002 to April 25, 2018 in the MAUDE database. This raw data was adjusted to improve accuracy. There was a total of 1,732 raw death reports of which 350 were attributed to the PMA process and 1,382 attributed to the 510(k) process. The adjusted death reports generated a total of 218 death reports of which 26 were attributed to PMA and 192 to 510(k). There was a total of 161,376 raw injury reports of which 36,033 were approved via PMA and 125,343 via 510(k). The adjusted injury reports generated a total of 2,254 reports, of which 440 were approved via PMA and 1,814 via 510(k). There was an 87% drop in number of adjusted death reports and a 98% drop in the number of adjusted injury reports as compared to raw data (p=.0015). Conclusion Death reports in the MAUDE database are more accurate than injury reports, whether approved by the PMA or 510(k) process, and more likely to contain an actual death, as compared to actual injuries in injury reports. Raw death reports overestimated deaths 7.9 times as compared to adjusted death reports; injury reports overestimated total injuries by 71.9 times as compared to adjusted injury reports. The warrants a call for a more accurate a national device registry with concurrent robust statistical analysis so trends of potentially harmful devices can be identified.