Feasibility of institution‐agnostic, EHR‐integrated regional clinical trial matching

• Published in: Cancer 2024-01, Vol.130 (1), p.60-67
• Main Authors: Shriver, Sharon P., Arafat, Waddah, Potteiger, Caroline, Butler, Dorothy L., Beg, Muhammad S., Hullings, Melanie, Semy, Salim, Lister, Zach, Khosama, Leticia, Armstrong, Susan, Hadley, David, Pappa, John, Fleury, Mark E.
• Format: Article
• Language: English
• Subjects: Automation; Cancer; cancer clinical trials; Clinical trials; Clinical Trials as Topic; clinical trials screening; Data search; Drug development; Electronic Health Records; electronic health records (EHR); Electronic medical records; Eligibility Determination; Feasibility; Feasibility Studies; Humans; Matching; Medical research; Neoplasms - diagnosis; Neoplasms - therapy; Patient Selection; Patients; R&D; Regional analysis; Regional development; Regional planning; Research & development; Software; Urban areas
• ISSN: 0008-543X; 1097-0142
• DOI: 10.1002/cncr.35022

Background A lack of onsite clinical trials is the largest barrier to participation of cancer patients in trials. Development of an automated process for regional trial eligibility screening first requires identification of patient electronic health record data that allows effective trial screening, and evidence that searching for trials regionally has a positive impact compared with site‐specific searching. Methods To assess a screening framework that would support an automated regional search tool, a set of patient clinical variables was analyzed for prescreening clinical trials. The variables were used to assess regional compared with site‐specific screening throughout the United States. Results Eight core variables from patient electronic health records were identified that yielded likely matches in a prescreen process. Assessment of the screening framework was performed using these variables to search for trials locally and regionally for an 84‐patient cohort. The likelihood that a trial returned in this prescreen was a provisional trial match was 45.7%. Expanding the search radius to 20 miles led to a net 91% increase in matches across cancers within the tested cohort. In a U.S. regional analysis, for sparsely populated areas, searching a 100‐mile radius using the prescreening framework was needed, whereas for urban areas a 20‐mile radius was sufficient. Conclusion A clinical trial screening framework was assessed that uses limited patient data to efficiently and effectively identify prescreen matches for clinical trials. This framework improves trial matching rates when searching regionally compared with locally, although the applicability of this framework may vary geographically depending on oncology practice density. Plain Language Summary Clinical trials provide cancer patients the opportunity to participate in research and development of new drugs and treatment approaches. It can be difficult to find available clinical trials for which a patient is eligible. This article describes an approach to clinical trial matching using limited patient data to search for trials regionally, beyond just the patient’s local care site. Feasibility testing shows that this process can lead to a net 91% increase in the number of potential clinical trial matches available within 20 miles of a patient. Based on these findings, a software tool based on this model is being developed that will automatically send limited, deidentified information from patient medical r