Unpacking Averages: Likelihood of FDA Medical Device Inspections
It is common for FDA and others to show a map of the United States with the states color-coded by intensity to showcase the total number of inspections done in that state. Indeed, FDA includes such a map in its newly released dashboard for FDA inspections. In reviewing that map with the U.S. map color-coded to reflect where medical device establishments are located, do you notice anything? Not to destroy the suspense for you, but it turns out that FDA tends to inspect where medical device inspection facilities are located. Really.
We wanted to get beneath those numbers in two ways. First, it’s much more informative to look at the data at a county level because there’s actually quite a bit of variation county by county. Second, and more importantly, we wanted to normalize the inspection data by the number of facilities. In other words, by looking at inspections per facility, we can get a better sense of the inspection frequency in each county.
This is an interactive chart where you can move your cursor over the counties to see the particular percentage. Please do not read the percentages in the legend too literally. For example, 100% means that the number of inspections was equal to or greater than the number of registered firms in that location. But as we all know, that likely means that some firms never got inspected, and other firms got inspected more than once over the course of the 14 years. Thus, it is just a comparative metric that allows you to see the differences across the counties.
This takes a little bit more work than you might guess and the product here should be viewed as an estimate. There are several places where we needed to make some assumptions and gloss over some minor discrepancies. Here’s how we did it.
The FDA inspection database, at the time we did this study, included the dates from October 1, 2008 until September 15, 2021. We chose to leave that entire time period in as it gives a smoother assessment over that roughly 13 years. Some years FDA may conduct many inspections; other years not so many. We didn’t care so much about year-to-year variation as we did establishing a typical inspection level.
There were just over a quarter million FDA inspections during that time. We filtered the data to focus on domestic inspections, and done under the auspices of the Center for Devices and Radiological Health. That left just about 33,000. FDA conducts inspections for many different reasons, and we chose to focus on those that involve device compliance. That took us down to just over 18,000. That is the pool of inspections we chose to use.
Next we went to the registration database. Obviously, registrations change from year to year, and so the fact that we are using 13 years of inspection data adds some complexity. We chose to use the number of registered facilities in fiscal year 2021. We believe that geographic dispersion of registered facilities hasn’t changed so much over the last 13 years as to render the results too inaccurate. In addition, the inspection database that FDA released only included the name of the firm, not the establishment type. But in looking at the range of establishments, it’s possible that any of the different types of registered establishments might be inspected for device compliance, from medical device manufacturers to re-processors of single use devices, to exporters. So we chose to leave all of the establishment types in the data. Obviously it would’ve been better if we could filter both the inspection database and the registration database by establishment type, but because that wasn’t available on the public inspection database, we couldn’t refine the analysis in that way.
We merged the two lists using ZIP codes as the common geographic denominator. In the end, because of the imperfections of the databases and because of the changes over time, there were some counties where there were inspections, but there were no registered establishments in 2021. Fortunately, as you might guess, in these cases there was only typically one inspection in the whole county, probably because there was a facility there that is no longer. But the numbers were very small. We treated those counties as zeros.
Producing the visualization at the county level is a bit tricky. As you may know, ZIP codes do not correspond to counties. Many ZIP codes cover multiple counties. So it took some work to create a crosswalk from ZIP codes to county data (using a data set from the Department of Housing and Urban Development that includes the estimated county location of businesses) and we had to make some assumptions along the way. But, again, we believe this produced a reasonable estimate.
As we suspected, there’s a lot more nuance to the data than just concluding that inspections are most prevalent where registered firms are most prevalent. There are some counties where the inspection rate is very high, and others where it is dramatically lower. Where you operate has a big impact on the likelihood of inspection.
What’s interesting to us is that it doesn’t seem to necessarily correspond with FDA district offices or regions. In other words, we don’t notice any particular region of the United States having a higher prevalence of FDA medical device inspections than other regions of the United States. Indeed, at the state level we don’t really discern any trends. There are certainly some states like North Dakota that don’t have any hotspots. But quite honestly that’s because North Dakota has few medical device establishments to begin with.
Bottom line, it seems as though there are county level hotspots but not state or region hotspots.
What does this map reveal to you?