Rebecca Jackson: EHR-To-EDC Needs A Reality Check
By Dan Schell, Chief Editor, Clinical Leader
“I don’t come to work to be a politician or to make friends.”
That’s not a sentence you hear very often from someone leading technology initiatives inside a major pharmaceutical company. But it perfectly captures Rebecca Jackson‘s approach to innovation.
Jackson, who is associate director, global development systems at Johnson & Johnson Innovative Medicine, continued, “I come to work for the good of the patients. I don’t want to be wasting time or money on some [initiative or technology] that is slowing down the critical path and that offers no benefit to the patient.”
That perspective caught my attention because it wasn’t anti-technology. Quite the opposite. Jackson believes EHR integration, AI, and other innovations all have an important role to play. She simply believes the industry has a habit of treating new technologies as silver bullets before fully appreciating the work required to implement them successfully. During her presentation, she gave a great analogy when comparing EHR systems to Monica’s secret closet from Friends — a place where everything technically exists, but nothing is organized, complete, or easy to find.
After her presentation, I asked her if she’d be willing to continue the conversation in a formal interview. She agreed, but only with the disclaimer that she would be speaking in a personal capacity and not on behalf of J&J. During our interview, she offered a refreshingly candid assessment of where the industry stands today, where she believes it’s headed, and why practical thinking often gets overshadowed by excitement over the next big innovation.
Innovation Should Solve Problems, Not Create Them
One thing became clear early in our conversation: Jackson isn’t interested in challenging industry consensus simply to be provocative. Her skepticism comes from years of implementing technology, not watching demonstrations of it. In fact, she’s quick to note that she welcomes innovation, but she simply believes organizations need to distinguish between ideas that genuinely improve clinical research and ideas that merely shift work from one group to another.
She also believes the pharmaceutical industry often approaches innovation backward. Rather than anticipating technologies years before they’re ready for widespread adoption, companies tend to react once a new trend has already captured everyone’s attention. “In this industry, we can’t do innovation like Apple and Google. Innovation, for us, is a three-year timeline.” That longer timeline is a reality of working in a highly regulated industry that requires the tech to be validated, documented, tested, supported, and accepted across multiple stakeholders before it is deployed and delivers meaningful value.
The Reality Behind EHR-to-EDC Integration
Jackson believes EHR-to-EDC integration is a perfect example of expectations outrunning reality. “The technology and the infrastructure isn’t there for us to really grab this by the horns at the moment. Site models aren’t really set up for it. It is, at this stage, an extra piece of work.”
I asked her to explain, and she started by saying that simply moving data from one system to another is only a small part of the overall process. Sites often need additional agreements with intermediary vendors serving as data brokers. Sponsors must establish support models that define exactly who handles technical issues. Every study requires new mapping between the site’s EHR and the sponsor’s EDC system. Every mapping exercise requires testing and validation before patient data can flow reliably.
“You think you are just literally picking up data from A and putting it in B. It’s not that simple, because there are probably at least six ways to do it, and each has a failure point where the whole system/process could just collapse.”
Those failure points multiply because no two studies are identical. CRFs differ. Visit schedules differ. Data standards differ. Individual health systems often customize their EHRs differently. Even if the same site participates in multiple studies, much of that work has to be repeated.
Jackson also pointed out that today’s integrations primarily capture highly structured information such as laboratory values and vital signs. Those are relatively easy to standardize, but they also represent some of the lowest-value data from an operational perspective. The information sponsors most want — complex oncology data, physician notes, and other less structured clinical information — remains far more difficult to extract consistently.
COVID Offered A Valuable Reality Check
Jackson’s views aren’t based solely on theory. During the COVID pandemic, her team participated in what amounted to a mirror study, collecting the same patient information both through traditional manual data entry and directly from EHRs. The project demonstrated both the promise and the limitations of EHR integration.
One clear success involved hospitalization events. Because hospitals immediately record patient admissions in their EHR systems, researchers were able to identify hospitalizations faster than they could through traditional manual reporting. The broader data, however, told a different story. “We really started to understand the mess that is EHR data,” Jackson said. “That really opened our eyes in terms of how difficult this is to map.” Information wasn’t consistently stored in the same locations. Different sites documented similar concepts differently. Some information appeared in pharmacy records, other data lived in billing systems, and terminology varied widely between organizations.
As a result, while the HER data proved useful for analysis, the study still relied on manually collected data for regulatory submission because it had already passed through established quality-control and cleaning processes. The experience reinforced Jackson’s belief that the industry’s biggest challenge isn’t collecting more data; it’s making existing data usable.
No Drastic Changes, Just Smoother Trials
Jackson is just as measured when discussing AI. Essentially, she expects it to remove many of the repetitive administrative tasks that currently consume enormous amounts of time. But it’s not going to drastically “reinvent” clinical trials like some folks have predicted.
One opportunity she finds particularly promising is data mapping, which is useful, again, when talking about the promise of EHR-EDC integration. Today, mapping EHR information into an EDC system often requires study-by-study and site-by-site configuration. Jackson believes machine learning could eventually recognize recurring patterns, automate much of that work, and continually improve as more studies are completed.
“I don’t think AI is going to change the way clinical trials are run … nor is it going to replace ClinOps professionals,” she said. “But, I do think it’s going to make trials run smoother.”
That outlook also shapes her expectations around return on investment. Organizations shouldn’t expect dramatic cost savings within the first few years of implementing new technologies. Building expertise, training staff, validating systems, and supporting new processes all require significant investment before meaningful efficiencies begin to emerge.
Her perspective may sound contrarian in an industry eager to embrace the latest technological breakthrough. But after spending an hour discussing interoperability, AI, and operational execution with Jackson, it became clear that she’s simply questioning whether we’re asking the right questions before declaring victory about AI or any tech “revolution.”