MedCon: Experts discuss approaches to large-scale data collection and management

COLUMBUS, OH – As companies are required to collect more and more clinical data to stay in compliance with global regulatory requirements, the role of becoming a good steward of that data has become more challenging, a panel of experts said at the 2025 MedCon conference, sponsored by the AFDO/RAPS Healthcare Products Collaborative.

“No matter where you are in the world, no matter what regulations you’re trying to satisfy, you have a lot more data that’s at your fingertips,” Mindy McCann, vice president of US operations and principal consultant at Qserve Group, told attendees during a panel discussion on the global shift in clinical expectations for medical devices.

McCann said the data are being collected primarily to achieve regulatory compliance, but the purpose is for improving safety and effectiveness for patients, and potentially drive innovations that would improve products, make better decisions, and reduce the time to market for devices. However, she acknowledged the collection of fit-for-purpose data and aligning it to added value for patients, users, and the manufacturer can be challenging.

“One of the big things that I see in management of data and good practices is this whole concept of managing communication in silo,” she said. For instance, departments within a company can collect different data, or even the same data, and use it for different purposes. There can also be a lack of communication in companies surrounding data collection.

“One of the big tools that I've seen that I think a lot of companies are now getting much better at doing is improving the cross functional communication in order to use the data more efficiently, have one data collection source that is then used by multiple teams, defining the objectives of why you’re collecting that data, to be able to allow more effective use of analysis of that data,” she said.

Having a central source of data can increase the number of stakeholders involved, potentially leading to “more innovation, more perspective and different ideas that are coming to the table than there used to be,” she said.

AI tools for data management

Gert Bos, executive director and partner at Qserve Group, said he has noticed the mixing and matching of data, including data from different sources, data with different levels of quality, and even the mixing of pre-market and post-market data from different countries. There is the potential for artificial intelligence (AI) tools to aid in management of this data, but the panel said their experience with the technology has been mixed.

Matt Albert, senior fellow, corporate clinical quality at Boston Scientific, said his company has begun looking at AI tools to aid in the clinical evaluation process. His company has specifically looked at how AI tools perform in creating completed or updated clinical evaluation reports when compared them with manually created reports, but has run into issues with hallucinations, mismatched data, incorrect data tabulations and summaries.

“I think the maturity of the tools haven’t been adequate, required a lot of re-work, and quality checks and didn’t end up in the time savings we were hoping for,” he said. “We’re eager for that tool that does increase efficiency to be available to us.”

Andrew Webster, senior manager in post market surveillance team at Zimmer Biomet, said AI is still a new technology. “No one’s really sure what it’s able to do for us and what we can accomplish with it,” he said.

Albert said part of the problem may have to do with the scale of the data the AI tools need to learn on. A resource model is also likely needed to onboard an AI solution, and training the model would require a full-time resource. “Obviously, as AI continues to advance, it’s going to continue to get better. It's just not there where we would need it to implement it in any sort of way,” he said.

McCann said there are some areas where AI works better than in others, such as in the state-of-the-art section of a clinical evaluation report. “I've seen some tools that seem to give pretty decent data, but in the actual data analysis piece, it may not be as good,” she said.

One good use case for AI seems to be is keeping information aligned in the technical documents of a European Medical Device Regulation submission. “It's very easy for things to get—as the documents are updated—for information to not be aligned across the tech doc,” Albert said. “I think that is something that AI could really help us with, is ensuring [we’re] doing those quality checks and consistency checks across the technical documentation.”

MedCon: Experts discuss approaches to large-scale data collection and management

Related topics

Recommended content

Welcome to the new RAPS Digital Experience