April 12, 2021
AI is Restarting Clinical Trials Put on Hold During the Pandemic – and Redesigning the Trials of the Future
BLOG AI ConcertAI Perspectives
By Jeff Elton, PhD, CEO
COVID-19 has negatively impacted oncology care disproportionately and in a unique way. As a study published in JAMA Oncology demonstrated earlier this year, rates of screening for cancer went down considerably as patients avoided health facilities and as the facilities implemented access restrictions. That, in turn, meant that patients’ cancers progressed without detection and that initiation of treatment occurred at a more advanced stage, often with less positive outcomes. While this is just a single illustration, we see other issues negatively impacting patients – high vulnerability of patients with hematological malignancies; reduced accruals to potentially beneficial clinical trials; and fewer new trials opening. Together these trends have serious implications for patients and for advancing the biomedical innovations they need. It puts at risk the enormous progress we’ve seen over the past several years – a time when cancer deaths were finally declining.
Fewer diagnosed patients mean fewer prospective patients who may consider clinical trials. Limited clinic access means more arduous processes for sponsors to interact with clinical sites and less available staff time for clinical teams to screen patients for prospective study eligibility. This limited access and staff time limitations further means that the burden for patients to enter trials will be much higher as there will a strong preference for standard of care treatment.
But with the help of AI, cancer centers are restarting and initiating new trials. Foremost in this process is reconsidering the operating models for executing and redesigning studies to be less burdensome for sites and patients. Reconsideration has involved using AI-driven solutions to identify eligible patients; remote access tools to upload clinical documents that will support electric case report form completion (eCRF, and other study artifacts that previously might have required direct, on-site interactions.
Now more than ever, sites want to broaden eligibility criteria to allow more patients to receive consideration. AI-driven software and similar tools can help, assuring the study intent and design will reach the targeted number of participants in as short a timeframe as possible. We are already seeing tremendous progress in AI enhancing the ability to identify patents for clinical study eligibility. AI solutions can be tuned to assuring a limited number of false negatives – meaning, if a patient is eligible for a clinical trial, they’re identified. It also ensures that there aren’t too many false positives, which would only create extra work with results that won’t be used by providers or research personnel. Perhaps most critically, this AI processes structured and unstructured clinical data sources.
The new uses of AI and other technologies to get trials back on track relies on evidence-based approaches as opposed to depending exclusively on protocols and study designs from past randomized controlled trial studies. This is a subtle, but important, change. Standard of care data combined with advanced AI can optimize trials for this broader inclusion as well as provide higher assurance that the study can demonstrate benefits relative to the current standard of care.
AI-driven approaches to trial design, site optimization, patient identification and escreening – even the processing of data for digital trials – will become a new standard once the pandemic recedes. It took an exogenous shock of unprecedented proportions to accelerate solutions that had been slowly maturing on the edge of the industry for years. These approaches are now serving as the basis of all new trials in a number of biopharma companies. While these solutions will complement traditional ways of working in the beginning, supporting new ways of working for about 25% of all trials, they will improve rapidly and supplant more and more of our inefficient legacy models. There will be no going back.
In the end, it will be patients who benefit. These alternative trial models will make clinical studies more accessible as more studies of greater sophistication are moved into advanced community oncology and regional health systems – important as this is where 80 percent of U.S. patients receive their care. Having access to standard of care therapeutics and clinical trials means a broader set of treatment options for the most devastating cancers. It also means that patients of different racial and economic groups now have access to more trials – a real issue in the ‘research disparities’ of the past. Lastly, designing and conducting trials closer to standard of care clinical workflows holds the potential of advancing outcomes and benefits more rapidly than in years past.
AI-driven solutions have protected the interests of providers, patients, and biopharma sponsors while accommodating the challenges of COVID-19. It’s been really remarkable what we have seen in terms of new models of collaboration, cooperation, technologies, and innovative approaches, and those changes are going to drive us towards an entirely new trial model – one that ConcertAI is calling “Integrated Digital Trials.”
(Read more here in Healthcare IT News’ Q&A with Jeff on how enterprise AI is driving decentralized clinical trials)