In a recent video interview with Applied Clinical Trials, Sunny Kumar, MD, partner, Informed Ventures, discussed the slow adoption of DCT models despite their benefits in retention and diversity. Key barriers include high upfront costs (millions per trial) and the current down cycle in pharma investment. To address the digital divide, Kumar emphasized providing devices to patients rather than relying on their own, especially in under-resourced areas. Operational and cultural challenges, such as regulatory compliance and risk-aversion, contribute to a six-year lag in adopting clinical trial innovations. Generative AI is seen as a promising tool to reduce costs. Kumar also highlighted the need for tech platforms that fit seamlessly into the pharma ecosystem.
ACT: What operational or cultural barriers do you think contribute most to the six-year lag in adopting clinical trial innovations, according to the Tufts CSDD report?
Kumar: Yeah, it’s a great point. We talked a little bit on the operational side, and that these are very large organizations, especially on the pharma side, where in order to adopt these solutions, you have to get approval, not just from the clinical trial leads, but from IT, from security, from legal, and making sure that you do all of that properly in compliance with the regulatory framework is critical. Those are all good things, but we have to recognize that that takes time, that takes effort, that takes money, and that’s part of the reason why it takes so long for any technology, including things like decentralized trials, to get widely adopted across pharma. I would say that’s all in the operational domain, and that’s still occurring. We are looking at ways to speed that up. One way that seems to be very promising is that new generative AI tools are decreasing the cost barrier, the service barrier, in order to integrate some of these solutions. One example of that is, if you look at the costs associated with spinning up a trial protocol, you might have to, for example, take a trial protocol and translate it into 50 different languages, all of which typically involves a significant amount of manual work. Now, with generative AI, you can use AI to assist that translation, significantly reducing the amount of labor required to translate those protocols into those different languages. Of course, given that this is involving healthcare, we want to have a human in the loop reviewing all of that material, but even still, the amount of human labor required is significantly reduced.
There’s still a cultural component, as you mentioned, pharma and healthcare in general is very risk averse, and there is a regulatory component to all of that, and that we have not found a way to get around this today. That may be for the best, but the reality of it is that pharma generally lags behind other industries outside of healthcare when it comes to adopting new technology. It’s not the slowest, but it does mean that we are not the first. We’re usually not the forefront when it comes to adopting new technologies, and I don’t expect that to change anytime soon.