FDA MAUDE Medical Device Reporting Database Analytics

G

gudushi.kutushi

#1
Dear Experts,
I am trying to do some research around effectiveness of running analytics on FDA Maude database for my organization. I have a statistical and analytical background, however my experience around Medical Devices is limited.
I stumbled upon this forum when I was researching online. I was honestly, amazed to find such intellectuals in one single place! and thought I will raise the question to this group.

* What type of analytics can an organization run on FDA Maude database ?
* Does FDA run analytics or mine the MAUDE data to help them capture signals, that flags that they need to go after an organization for Audit?

It would be great if someone can give me some direction, and I shall gladly appreciate it.
Thank you very much in advance!
Regards,
GK
 
#2
We have about a dozen device types which I have matched to the FDA/MAUDE three letter designations. Every month (latest month) I check if any (competitor) devices in these classifications have had alerts as part of my post market surveillance procedure/risk analysis. I also make a note of the total number alerts in that particular month (have to keep it just to ‘deaths’ to get an actual number). I keep a track on an Excel spreadsheet/workbook along with alerts from other regulatory bodies (MHRA, HC, bfarm, Afssaps etc). I also keep hyperlinks to the actual alerts. Certain simple patterns/problems can show up e.g. electrical components failure, misuse of a device, material failure, fire risk etc. In this way I will update my Risk Management File (ISO 14791) – new possible failure mode for a particular device (which you can show an auditor to demonstrate your surveillance skills!:)).:2cents:

Steve
 

robert.beck

Involved In Discussions
#3
I have downloaded the latest Maude data files from FDA website and put all of the data into a SQL database that is full-text searchable. I've used this for several searches; here is my opinion:

1) the data is not very clean. There are a lot of records that aren't very useful.
2) the reports are not consistent, so when I'm looking for one specific topic I have to use a variety of search terms, and winnow records over several iterations.
3) Because of the amount of data, and because this is the only publicly available data of its type, it could be useful, but would be much more useful if the data were coded and categorized beyond what is done by FDA.
4) Unfortunately, this database is not a complete dataset on the topic of adverse events - see FDA's disclaimers, which are:

"MAUDE may not include reports made according to exemptions, variances, or alternative reporting requirements granted under 21 CFR 803.19."
"MAUDE data is not intended to be used either to evaluate rates of adverse events or to compare adverse event occurrence rates across devices"
"Please be aware that reports regarding device trade names may have been submitted under different manufacturer names"
Two exemptions of the FOIA are applied to MAUDE: "Exemption 4, trade secrets and confidential business information; and exemption 6, personnel and medical files"

I don't know how FDA uses this data, but I think your points are valid - FDA probably monitors it somehow and could definitely use it to identify flags and signals.
 

yana prus

Involved In Discussions
#4
* What type of analytics can an organization run on FDA Maude database ?
As SteveK mentioned, it is a very useful tool for the post marketing survelliance, and risk management update. I use it also in the design stage when I would like to identify the potential risks and failures of the device to be taken into account for risk analysis, user requirements and technical specification.

Once I thought to use the database to explore the reliability (incident rate analysis) of the competitors products. The problem is as Robert mentioned that adverse events from the MAUDE database cannot be used to determine the rate of adverse events for a particular product or to compare adverse event rates between products. Manufacturers submit reports according to their interpretation of regulatory requirements and event reportability. For example, global or USA companies might submit more MDRs than small sized/out of USA companies as a result of internal compliance policy. What one company would consider as reportable, might not be reportable in interpretation of other company for the same type of device. Moreover, submission of the report does not mean the device caused the adverse event.
Nice resource on the problems with MAUDE system can be found here:
http ://www. fda gov/downloads/MedicalDevices/NewsEvents/WorkshopsConferences/UCM283555.pdf - DEAD 404 LINK UNLINKED
it mentions other challenges for the search in the MAUDE system:
- Duplicate and triplicate entries (follow-up reports),
? Blank narrative entries,
? Duplicate MDR ID numbers,
? Reports filed through the Alternative Summary Report (ASR) System not consistently updated to MAUDE
* Does FDA run analytics or mine the MAUDE data to help them capture signals, that flags that they need to go after an organization for Audit?
FDA has created the TPLC - total product life cycle database that combines data from various CDRH databases (data sources like MAUDE) to present an integrated record of premarket and postmarket activity for medical devices. The TPLC database integrates data by procode, a three-letter code associated with a generic type of medical device.

here is the link to the TPLC dabase - you can perform the search according to the device name or product code: TPLC - Total Product Life Cycle

For specific product code you could see there the types of device problems and number of the problems reported according to each type.
Another important source of the information that might be useful for your research is FDA's report for one of specific types of product - you could see there the method for mining the MAUDE data and the Public Health Notification that was issued as results of large number of adverse event received, with recommendations how to mitigate the risks. http :// www .fda. gov/downloads/AdvisoryCommittees/CommitteesMeetingMaterials/MedicalDevices/MedicalDevicesAdvisoryCommittee/ObstetricsandGynecologyDevices/UCM270402.pdf - DEAD 404 LINK UNLINKED

Regarding the signals for the audit - in my experience, I'm aware of 1 FDA inspection where the reason for inspection was lack of correlation in number of the user facility reports with the number of the manufacturer reports for the same cases. For that specific company there were about 20 user facility reports with 0 manufacturers reports for these cases as MDRs to FDA and the only reason for the FDA inspection was to inspect how that company responses, investigates and decides on the reportability for MDRs.
 
Last edited by a moderator:
V

vaneswh

#5
hi,yana! Thanks for your response and I learned a lot. I have some questions:

How do you find duplicate records in MAUDE database?
Can you estimate the numbers of reports that FDA received each year?

I did some analytics recently and used the variable "event_key" to find duplicates, here's an overview of report numbers by each year:

year frequency
1900 1
1991 15
1992 3,098
1993 4,407
1994 11,200
1995 9,724
1996 31,361
1997 73,275
1998 58,248
1999 49,329
2000 49,956
2001 56,188
2002 64,981
2003 70,288
2004 75,672
2005 94,262
2006 115,137
2007 149,273
2008 156,651
2009 204,213
2010 252,193
2011 369,321
2012 415,393
2013 488,359

The result is a little weird that the report number is dramatically increasing,especially from 2008. So I doubt whether I did enough "pre-processing" of the database.

I'd like to hear your opinion or advice.Thanks again!
 
D

drewsky1

#6
I am currently doing a Post Market Surveillance and am searching the FDA MAUDE database. It appears there may be duplicate and maybe triplicate entries for the same event. Can anyone advise if there is a way to identify within the individual reports if it is connected to another event without having to read sometimes 100 reports that come up? It doesn't appear that the "key" numbers assigned or date of event would distinguish them.

Thanks:thanks:
 
#7
For those looking for
Nice resource on the problems with MAUDE system can be found here:
http ://www. fda gov/downloads/MedicalDevices/NewsEvents/WorkshopsConferences/UCM283555.pdf - DEAD 404 LINK UNLINKED

It can be found here:
At the wayback machine web site just search for the above URL. I would paste it here but for some crazy reason, it will not let me :-( sorry
  1. Search google for Wayback machine,
  2. On the site search for the above link,
    1. be sure to fix the link remove spaces and add a . between fda and gov: fda.gov
  3. Click on 2015
  4. Click on June 12 to see the pdf
 

Top Bottom