68aa8042-b884-4b41-a32b-ca92aaeb83a2

Status: Original and derived artifacts are available for this document.

Download: pdf

Kiran Myalur Dharmaputhra
Lakeville, Minnesota
[email protected]
August 9, 2026
Re: Docket No. FDA-2026-N-7129 Adverse Event Program for Medical Devices; Medical Product
Safety Network (MedSun) (OMB Control No. 0910-0471)
Dear Dockets Management Staff:
I am writing in response to FDA's July 17, 2026, notice (91 FR 44863) requesting comment on the
extension of the information collection that supports the Medical Product Safety Network (MedSun).
These comments are my own, submitted in my individual capacity.
Some quick background so the Agency knows where my perspective comes from. I have spent over a
decade in regulated quality engineering; in my current role I serve as the quality subject matter expert
for Urology implantable product lines, providing day-to-day production support and owning the CAPA
and nonconformance programs. My work sits on the manufacturer side of the adverse event data that
MedSun produces complaint investigations, MDR decisions, and root cause analysis. I also write on
organized my comments around the four questions in the notice.
1. Is the collection necessary, and does it have practical utility?
Yes, and I would like to explain why from the receiving end. When an adverse event report lands on a
manufacturer's desk, the problem is never that we have too few reports. The problem is that most
reports do not tell us enough to figure out what happened. Was it the device? A use error? A
maintenance issue at the facility? A report that just says,
MedSun reports are different because the reporters are trained and the program follows up with
additional questions. A smaller network of three hundred facilities producing complete, clinically
detailed reports is worth more for signal detection than a much larger volume of thin ones. I support
extending this collection.
2. Are the burden estimates accurate?
Thirty minutes per report is right for a simple malfunction report. I doubt it holds for the harder cases,
and the notice says FDA is specifically promoting reporting from pediatric intensive care units,
electrophysiology laboratories, and hospital laboratories, which is where the harder cases live. These
events tend to involve implanted or life-sustaining devices, multiple products interacting, and someone
hunting through several hospital systems for serial numbers, lot numbers, and chart details.
Two suggestions:

Check the 30-minute figure against a sample of actual reports, broken out by complexity (simple
malfunction vs. serious injury or death vs. multi-device event), instead of using one average for
everything.
Count the supplemental follow-up questions as their own line item. They happen later, often
involve a different person, and folding them into the same 30 minutes hides real work.
Getting the burden model right also matters for question 4; it is hard to reduce burden that has not
been measured.
3. Can the quality, utility, and clarity of the information be improved?
A few things that would make MedSun data significantly more usable downstream:
Use IMDRF adverse event codes. If the report form captures (or maps to) IMDRF terminology
for device problems and clinical outcomes, MedSun data becomes directly comparable to
MAUDE and international vigilance data. Free-text narratives are valuable, but structured codes
are what make trending possible.
Ask for the UDI. Ambiguous device identification is the single most common gap I see in adverse
event data. A prompt to scan or enter the Unique Device Identifier would fix identification and
let the system auto-fill model and lot information, which also saves the reporter time.
Separate the device problem from the use environment. A field that distinguishes what the
device did from the conditions around it (training, maintenance status, configuration) makes
root cause attribution far more reliable. It also helps FDA spot use-related hazards that cut
across manufacturers, which no single company's complaint file will ever show.
Give the sites feedback. People who see their reports used write better reports. Even a periodic
aggregate summary back to participating facilities would help.
4. Can the burden be reduced, including through automation?
Let facilities pull data from their electronic health record and materials-management systems
instead of retyping it: patient demographics, event date, device identifiers. FHIR-based
exchange makes this practical now in ways it was not a decade ago.
UDI barcode scanning, per above. It serves both data quality and burden reduction.
Use a short form for simple malfunctions and expand into the full question set only when the
event involves death, serious injury, or multiple devices.
Make sure reporters can save a partial report and come back to it. Clinical details arrive in
pieces; the form should tolerate that.

Closing
MedSun is one of the more sensible things in device postmarket surveillance: a small, trained network
producing high-quality data rather than a large, untrained one producing noise. I support the extension
and offer these suggestions to make the data more usable, and the burden estimates more accurate.
Thank you for considering these comments.
Respectfully submitted,
Kiran Myalur Dharmaputhra
Lakeville, Minnesota