Imagine 5.6 million pieces of data collected by the U.S. Food and Drug Administration (FDA) concerning adverse events, the FDA term for complications.
Yet the agency has trouble identifying early warning signs for thousands of troublesome medical devices.
Madris Tomes, worked at the FDA as an IT project manager in its adverse event reporting program. She knew there was a better way of making data useful to consumers, hospitals, doctors and even attorneys.
The fixes being sought by the current programmers to the MAUDE (Manufacturer and User Facility Device Experience) adverse event database used an Oracle-based system. Tomes thought that works for day-to-day data entry, but not for identifying the big picture safety signals.
With 70,000 new adverse events reports coming into the FDA every month, Tomes knew it was impossible for the FDA to keep up with needed recalls. So she quit the agency two years ago to build her own user-friendly system with Marklogic developers.
In other words, what devices are working in real life, and which are not.
“I knew there were lots of death reports and malfunctions and injuries from medical devices. The data is so dirty, you never see the true scope of any device company,“ she told Mesh News Desk.
What do we know about pelvic mesh and its complications?
Well, it depends on how you ask.
And there is no shortage of search terms, which essentially describe the same thing.
To do some practice runs, Mesh News Desk and Tomes used a few terms - "mesh," “surgical,” “polymeric,” “synthetic,” “urogynecologic” “sling,” “pubovaginal,” and “sepsis,” among them.
Running some numbers for MND, Tomes found:
Using the search terms “Mesh” in general, and “Sling” showed 126,349 adverse event reports. And there were 49 recalls. (When J&J pulled its Prolift off the market, that does not qualify as a recall).
"Mesh Surgical" and "Polymeric" yields 66,000 complication reports. How many were deaths?
It's hard to tell because they are sometimes incorporated into terms like "malfunction" or "injury".
Tomes tells MND, “Each company has its own way of saying how a patient died. Use the terms “autopsy,” “failed to live,” “patient expired,” “injury,” or "malfunction." So I created an algorithm to say the way the patient died rather than say 'death'.”
*Various search terms found 5,900 death reports. The word “autopsy” uncovered another 500 deaths.
*Put in mesh, “near death,” yields 1,500 deaths from mesh, but only 400 boxes are checked, which could be misinterpreted, she says.
*The terms can be broken down by company, using “mesh” or “sling” and J&J there were 43,552 adverse events. If you used the term “pelvic” you wouldn’t see that.
* "Mesh/Sepsis" found 287 AE (adverse event) reports, but you can also look at "perforate" and "rupture" which found 378.
*"Mesh/colon" and "colon removed" found 722 reports and 26 deaths.
* "Tension-free" and "polymeric mesh" found 65,062 adverse events and 18 recalls. Deaths in that group - 281 have death box checked.
*"Polymeric sling"- showed 23,392 adverse event reports, 430 deaths. Of those, 7,762 reports came from physicians.
*"Physiomesh", the recalled J&J hernia mesh, shows 689 adverse event reports, 9 deaths and two recalls.
As has been reported in the past, the FDA’s database is not particularly user-friendly or even known to some who work with medical devices. The number of reports coming into the agency represent only about 14 percent of actual real-life adverse events occurring in hospitals, for example, according to a 2012 Office of Inspector General report (here).
Tomes says some reports are redacted by FDA reviews. She doesn’t know why. When looking up a particular super bug, there was a Freedom of Information clause that said it was protected information or HIPPA and the information was redacted. She doesn’t know whether companies tell the FDA to redact information.
“That worries me,” she adds.
Also clouding the true picture of medical devices is that when the FDA adds a new year of data, it removes a year. MAUDE shows the last 10 years, but the raw data goes back 20 years. It isn’t a question of space on the server. The government requires records are kept a certain amount of time, that’s all.
Reports are still not coming in from the Veterans Administration for some unknown reason, she says.
“There is an entire universe out there the FDA doesn’t see.”
But Tomes admits the reports that are coming in are “scary.”
For example, recently Essure activists asked Tomes to run the numbers and she found 303 fetal deaths from the permanent birth control device. The FDA data had shown only five deaths.
This information was used by Rep. Mike Fitzpatrick and a press release pushing for in legislation to get Essure off the market.
Based on a subscription service, hospitals will be able to use the Device Events reports to find out which devices are yielding the most complications so they do not reorder them.
“There was a group purchase of 200,000 Essure devices in June. They weren’t supposed to purchase anymore but the person didn’t know that and put through another order. Once you make a decision, there has to be some way to turn off the mechanism,” says Tomes.
Tomes believes this work is needed now to support true patient advocacy.
“With 70,000 new reports coming in every month, how can a doctor and provider keep up with this? They rely on the FDA, but the agency takes two months to two years to recommend a recall. Rather than wait for the FDA, I’m trying to put the data in front of the public so they can decide!” ###