Between ICH E6(R3) and AI, clinical data management is moving upstream. At SCDM this year, that shift was impossible to miss.
I’ve just come back from the SCDM Annual Conference in Raleigh, and one theme has stayed with me.
I’ve been part of the SCDM community for years, on stage, in sessions and in the hallway conversations that often tell you as much as the agenda does. The conference is always a useful barometer for where clinical data management is heading.
This year, the conversations felt different.
The questions were not only about faster edit checks or cleaner listings. They were bigger: What is our role becoming? Where should data management sit in the study lifecycle? And what becomes more important as technology takes on more of the mechanical work?
My read is that clinical data management is increasingly becoming a strategy function. Two forces are accelerating that shift: ICH E6(R3) and AI.
E6(R3): Quality is something you design, not inspect
Clinical data management has traditionally been positioned relatively late in the process. Data comes in, it is reviewed and cleaned, queries go out, and ultimately a database is locked.
That work remains essential. But E6(R3) pushes us to think further upstream.
Quality is not something that can simply be inspected into a study at the end. The factors that are critical to the reliability of the data need to be understood early, and the study needs to be designed with the risks to those factors in mind.
That changes the conversation.
The question becomes less, “How many checks did we run?” and more, “Did we design this study so that data integrity was built in from the start?”
That work does not begin at database lock. It begins much closer to protocol and study design.
For data leaders, that means data management being involved earlier. The more context we have around the protocol, endpoints, operational realities and potential risks, the better positioned we are to help design a data strategy that supports the study rather than simply reacting to it later.
AI: The mechanical work is automating. The human work is becoming more important.
AI was, unsurprisingly, everywhere at SCDM this year.
Automation can already support activities such as reconciliation, anomaly detection and query drafting, and those capabilities will continue to develop.
It is understandable that this creates some anxiety within the profession. But I think focusing only on the uncertainty of the change that AI will bring misses the more important shift.
When the mechanical work becomes easier to automate, human value moves toward judgment, context and interpretation.
A system can flag a value as an outlier. An experienced clinical data professional still needs to understand whether that value is an error, an expected result, or potentially one of the most important signals in the study.
That distinction matters.
It is also the thinking behind how we approach platforms such as BioGRID at Bioforum. The objective is not to remove expert reviewers from the process. It is to centralize and visualize the data so they spend less time assembling the picture and more time understanding what it means.
Technology will make the ability to apply good judgment faster, not make judgment unnecessary.
As automation grows, I believe that is where the role of the data professional becomes even more valuable: understanding risk, asking the right questions and deciding where human attention matters most.
What this means for sponsors
Put E6(R3) and AI together, and the direction is clear. Data management has an opportunity to shape how studies are designed and managed, rather than primarily responding to issues once data starts arriving.
There are three things I would encourage sponsors to think about:
- Bring the data management team into study design discussion earlier. The best time to prevent many data issues is before the study starts. Data expertise should be part of those early conversations.
- Define risks before planning the execution. Be clear about which data and processes are most critical to the study, where the risks sit, and how the design will protect the output. Automation can then support a strategy that has already been thoughtfully defined.
- Invest in judgment as well as systems. Technology matters, but systems alone do not determine what is important for a particular protocol, endpoint or patient population. Critical thinking, clinical context and risk awareness are the most valuable skill set.
The goal is not a cleaner database. It is earlier identification of risks and triggers, better-informed decisions throughout the study and ultimately trusted data that supports the outcome of the trial.
Our moment
What struck me most in Raleigh was that these conversations were happening everywhere.
Clinical data management is not becoming less important as technology advances. It is being asked to contribute at a higher level.
That is a significant opportunity for our profession.
Clinical data management is having its moment. The opportunity is no longer just to clean up after the science. It is to help shape how reliable evidence is generated from the start.
For those who were in Raleigh, I would be interested to hear whether you felt the same shift.
Tanya du Plessis is Chief Data Strategist & Solutions Officer at Bioforum the Data Masters and served as Conference MC at SCDM 2026. Bioforum is a biometric CRO delivering data management, biostatistics, statistical programming and FSP services, and the developer of BioGRID, a study data visualization and oversight platform.
Share with friends: