i don’t think using the pharmaceutical industry as a comparison is a good idea since drugs like thalidomide were allowed on the market and cigarettes caused “no health issues” for such a long time. I believe that the government should have strict design of experiment guidelines, require companies to submit all raw data results and analysis methodology, as well as hire skilled statisticians,who would look at the methodology to ensure significance of the statistical analysis. With a good enough statistician, poison ivy laced bandages could be proven to be harmless and have no adverse dermal effects. (no joke). With the rise of big data analytics, man power required to do tasks deemed to time consuming or taxing on the workforce 20 years ago can be completely autonomously with proper system designs. The FDA as well as other governing agencies should all go through a modernization transition to better improve their oversight and efficiency without putting a burden on companies and facilities that are fully compliant with nothing to hide. A lot of the data mentioned above are already recorded in the facilities and companies, its just that it is not required for submission. If a secure cloud server or some other collection mechanism is created to facilitate the transmission of such data, the AI big data analytic programs can do the assessment and point out issues that human employees would then address.
Making assessing testing procedures a part of government audits would also help rid certain unethical or inadequate testing method currently used by companies. Also note that not all defects are reported. If gauze rips in surgery, the doctor will simply grab more. If its a systematic issue they may get a new supplier but the defects don’t always make their way to the FDA’s ears. Assessing inventory use may also provide hospitals mechanisms of improving cost while giving a big data analyzer the information needed to find potential issues with products even if the users don’t notice.