How Memorial Sloan Kettering Cancer Center Scaled EHR-to-EDC Technology
Memorial Sloan Kettering (MSK) Cancer Center’s Senior Director of Clinical Research Informatics and Technology, Joe Lengfellner, tells us how MSK validated and scaled EHR-to-EDC technology across MSK’s vast portfolio of trials.
Why was scaling EHR-to-EDC technology at MSK such a priority?
At MSK, we have about 1800-1900 open research studies and a huge part of running a clinical trial is the data management that goes along with that. Traditionally, that has required manual data transcription. Our research staff must go through the EHR, looking for the right clinical data and then moving it over to the EDC system of the sponsor. We have over 400 research coordinators who are moving that data, but it’s not necessarily the job they signed up for.
When people get into research, they want to be experiencing the science of research, working with doctors and patients. They don’t want to copy-and-paste data from the medical record into another database. Beyond that, it introduces inefficiencies and quality gaps into the process. Therefore, addressing the transfer of data has always been a priority for MSK. Over the past decade, we’ve done several pilots and proofs of concept to show the ROI and value that a more streamlined process can bring.
"We see that sponsors are getting their data more quickly, in days rather than weeks, and at a higher quality. It doesn’t require back-and-forth queries, and it allows research coordinators to focus more heavily on the unstructured data or spend more time with patients and investigators."
Why is now the right time to move from pilots to reality?
EHR-to-EDC is now past the novelty stage. We are no longer proving efficacy, we are scaling it. Up until the past couple of years, it wasn’t truly scalable. One reason the time is right is because, at least in the US, we have the 21st Century Cures Act. This regulation required EHR systems to have a certain level of interoperability with FHIR, and allows vendors to connect to the clinical record in order to extract data. Any EHR, Epic, Cerner, etc, uses FHIR for interoperability.
What was the process for transitioning and scaling an EHR-to-EDC technology?
We identified a technology partner, IgniteData, that could help us scale EHR-to-EDC technology across our entire portfolio of studies. Initially, we established our partnership with IgniteData through our iHub program at MSK. Since the formation of the initial partnership, we have continued to expand our collaboration and have transferred north of 40,000 data points and hundreds of patients.
IgniteData is a mapping engine that sits between the EHR systems and the sponsor’s EDC system using FHIR. For the research coordinator, the transition has been easy. Instead of copy-and-pasting from a medical record, they launch the IgniteData platform, they see their patient clinical data appear in the interface, confirm it is correct, and then it moves into the EDC system. It’s critical to have a research coordinator still in the loop, confirming that the data being pulled is correct, but they don’t have to do it manually.
We’re seeing efficiency gains, data quality improvements and increases in job satisfaction from our research coordinators. We’re particularly excited about that last one because we are making the job of research more appealing to new people entering the profession. We hope to show that this has a positive impact on retention as well.
What were the challenges in adopting an EHR-to-EDC system?
The biggest challenge is the slow-moving nature of our industry. The technical challenges are quite minimal for sites implementing this software. There are some change management activities, but on the whole, sites are eager to take on this technology because they don’t want to do this work. Manual data transfer is an example of research coordinators not operating at the “top of their licenses”.
We have spent many years proving the efficacy and the ROI, and I would like to see us move faster as an industry. The slow rollout is something we’re constantly pushing against.
Have you witnessed any misalignment between sponsors and MSK when it comes to EHR-to-EDC?
Just the opposite, EHR-to-EDC technology has been a project that has brought sites and sponsors together, because the wins are so easily quantifiable on both sides. It requires collaboration between sites and sponsors to set up that initial study, but it’s also opened conversations around future additional projects between sites and sponsors.
What data is already fit-for-purpose for EHR-to-EDC transfer? What data is not yet ready?
Today, we are able to do this transfer very well with highly structured data: labs, vitals, demographics, adverse events, etc. However, there is still 30-40% of data that is unstructured and cannot, as of today, be transferred directly into the EDC. We’re working on strategies to address this and are making progress. That said, there are still some that will say, “Call me when you can do 100% data transfer.” I push back on this, my argument is that getting to 60% data transfer frees up the capacity to work on that harder 40%.
And with that 60%, we see that sponsors are getting their data more quickly, in days rather than weeks, and at a higher quality. It doesn’t require back-and-forth queries, and it allows research coordinators to focus more heavily on the unstructured data or spend more time with patients and investigators.
"The wins are so easily quantifiable on both sides. It requires collaboration between sites and sponsors to set up that initial study, but it’s also opened conversations around future additional projects between sites and sponsors."
Do you have a priority towards getting that unstructured data into the EHR-to-EDC transfer?
The goal is 100% data transfer. We have a few ongoing pilots where we're using large language models to try to extract unstructured data. The first phase will be extracting the data and confirming its validity with the coordinator to help surface the right data. The goal would be to slowly train the models to be better and better at surfacing and transferring the right data.
What activity would you want to see sponsors take that would make EHR-to-EDC transfer easier?
One would be standardization, for example the case report form design in the EDC. Every sponsor, even different studies within a sponsor, will design their case report forms differently. That makes integration take just a little bit longer because we must tweak it for each sponsor, or for each study.
Another would be sharing results. Sponsors are often hesitant to share ROI metrics from these initiatives. It slows the adoption of technology when people don’t share success or failure, so that the industry can move more quickly in the right direction.
What would be your message for sponsors to make trials less burdensome for sites?
As trials become more complex and require longer follow-up, and the work required by sites increases, we need to be more aggressive about looking for opportunities to reduce burden and rolling them out when there is success, such as EHR-to-EDC or trial-matching.
And in that vein, trust sites when they say that a technology is working for them. Let them use it and support them in their use, because sponsors reap the downstream benefits. I have seen a shift in the industry over the last 3-5 years, in which sponsors have gone from dictating technology to being more open to sites bringing in their own technology, and I hope that trend continues.
I attribute this partially to the fact that a lot of vendors have reached a level of maturity that technology implementation at the site level is much less risky. It also reflects the realization that sponsor-driven technology, no matter how well-funded, falls short without strong site adoption. That’s a sunk cost for sponsors. When sites get to use their own technology, that works for them, sponsors aren’t wasting resources on implementing technology that doesn't work for sites.
What else are you doing to make the running of clinical trials less burdensome/more accessible?
I co-lead our Clinical Research Innovation Consortium, which we stood up at MSK within the last 16 months. The consortium is tasked with looking for industry-wide bottlenecks, challenges, etc. and identifying solutions. EHR-to-EDC would be one example. We will test solutions within high-volume research environments, such as MSK, and then will share what worked, what didn’t, to help companies iterate on their product to improve them. We share with the larger industry as well, so that it can become industry standard.
We’re currently pushing forward our decentralized trial platform, which is a hub-and-spoke model for getting more research studies to more patients across the country and internationally. The hypothesis is that we can use MSK’s highly capable research sites (and/or other large Academic Medical Centers) and their robust infrastructure, personnel and technology to bring trials to sites that might not have the same robust research infrastructure. We’re testing out different aspects of this platform, using a mix of personnel and new technologies, to run the trials centrally and distribute/decentralize them to more sites.
"I have seen a shift in the industry over the last 3-5 years, in which sponsors have gone from dictating technology to being more open to sites bringing in their own technology, and I hope that trend continues."
How are you advancing into more decentralized trials?
We are in discussions with our sponsors about how they would like to work with a network of sites. We understand the questions that sponsors might have, but the real win for them is that they can contract and interact with MSK and get access to a much larger patient population and sites that they don’t have individual working relationships with. We think it cuts down on the complexities on the sponsor side, once they understand and trust in the model.
With that, we are also looking into technologies around clinical trial matching and asking how we can use the data in a patient’s clinical record to find trials that are a good match. We’re looking at the supply chain to understand how we can use technology or centralized processes to deliver drugs directly to patients or closer to them so that they don’t have to come into our AMC sites to receive their study drug.
It's been encouraging to see the recent surge in investment in clinical trial innovation. I’m eager to see how this accelerates the drug development pipeline and, most importantly, improves patients' lives.
