Quote from Guest on June 6, 2023, 7:01 pm

In my experience devices/apps assigned in studies are mapped to the randomized patient. There isn't really a way for sites to override that. 

I'm curious to hear more about what could make ePRO data "obviously" generated by machine? Are you saying that patients had the epro collection tool (let's say an app) and their responses were being overridden by some other input source? Or that you suspected that the end results were potentially deleted and replaced with something manufactured? It's a pretty bold claim either way.

I mean, if a study is using a fitness tracker or something, there isn't really a way (that i'm aware of) to know the patient didn't put in on their kid, or strap it to the roomba or something - - but the premise of this question feels rooted in sponsor paranoia along the lines of 'every site is consenting people just to get paid for screen fails!'

 

Quote from Guest on June 6, 2023, 7:02 pm

This is core to why the industry needs to continue the push towards Risk Based Quality Management and centralized monitoring and data analysis.  If the data was obviously being generated, it should be easy to identify the trend in central data analysis.

Quote from Guest on June 6, 2023, 7:30 pm

You should look into Verified Clinical Trials. 

To answer your original question,"So how do we ensure the legitimacy of data in these digital times? With compliances being bad enough, to discourage the use of CAPTCHAs, but having the obvious need for such measures" it comes down to being able to qualify the best patients at the best sites.

 

 

 

Quote from brad on June 6, 2023, 8:24 pm

Quote from Guest on June 6, 2023, 6:31 pm

The premise for discussion is quite simple, in times of patients remotely generating data, where do we get the confidence, that the patient is generating the data, and not the staff of a misconducting site.

Recently i encountered an unusual case, of ePRO data obviously being generated by bots/scripts, and the sponsor being all cryptic and silent about it. 

So how do we ensure the legitimacy of data in these digital times? With compliances being bad enough, to discourage the use of CAPTCHAs, but having the obvious need for such measures. 

There is a definite need to adress this question, but i never see it mentioned in the industry. I'm eager to hear some hot takes on this.

I admit that this topic is a bit over my head but I do find it hilarious that I have to use 2FA and have insane security requirements for the systems that I log in to that aren't even containing source data but we wouldn't even dream of doing that to participants. Tough spot.

Quote from Guest on June 7, 2023, 5:59 am

Quote from Guest on June 6, 2023, 7:01 pm

 what could make ePRO data "obviously" generated by machine? Are you saying that patients had the epro collection tool (let's say an app) and their responses were being overridden by some other input source? 

If the subject is in control of when ePRO data gets sent, what are the odds of it happening at exactly the same time every day? 

What are the odds that hundreds of subjects do the reporting in alphabetical order?

Why would Aaron take a measurement at 8:00, Bethany at 8:02, Chris at 8:04, Dave at 8:06 etc., and never break this alphabetical order?

Yes, there is a certain sense of paranoia, but once you see random data not being random at all, and only one site in the trial displaying such behaviour, you can't help but wonder

Quote from Guest on June 7, 2023, 10:48 am

That sounds pretty suspect.  I would do two things:

• Report it to the ePRO company, with a subtext of “our next stop for reporting this is the FDA.”  If they weren’t inclined to investigate seriously, they will if they know the FDA may audit studies where their tool is in use.  They should have the tools to know how this is happening (is an API endpoint insecure?  If they capture location when the app is open, do all the devices show that they are in the same spot? Etc.)  Give them some time to respond - they should be very motivated to fix it and report to the FDA themselves.
• If you don’t hear back from the ePRO vendor, next stop is the FDA.