A novel artificial intelligence (AI)-based virtual patient navigator, MyEleanor, improved colonoscopy uptake among U.S. patients who were previously nonadherent to prior colonoscopy appointments. These findings will be presented by Moadel et al at the 2024 ASCO Annual Meeting (Abstract 100).
Background
Racial and ethnic minority patients in the United States often face barriers to colorectal cancer screening that can delay diagnoses until the disease has advanced to later stages, which can exacerbate disparities in survival. In spite of active outreach attempts made by skilled patient navigators, a large percentage of these patients do not adhere to colonoscopy recommendations. For example, at one New York City cancer center that serves minority and economically disadvantaged patients, nearly 60% of patients either cancelled or did not attend their scheduled colonoscopies, despite outreach from skilled patient navigators; when patient navigators re-engaged, about 20% did complete their colonoscopy—however, 80% never did.
The recent introduction of conversational AI-based applications has allowed health-care institutions to enhance their cancer care offerings by reducing the burden on their human patient navigator workforce, as well as potentially reaching a wider population of potentially at-risk patients.
Study Design
As part of a quality improvement initiative, the researchers examined whether an AI-based virtual patient navigator known as MyEleanor was capable of increasing re-engagement response by facilitating colonoscopy rescheduling in a group of 2,400 patients nonadherent to colonoscopy appointments in 2022 and 2023. Between April and December 2023, they used the MyEleanor tool to contact patients to discuss rescheduling their colonoscopy appointments, provide live transfers to clinical staff, evaluate the patients’ screening barriers, and call patients to remind them of their upcoming appointments. The researchers then assessed several outcomes following interaction with the MyEleanor tool—including patient engagement via identify confirmation, accepted live transfers, predictors of actionable engagement, proportion of patients with barriers to care, and colonoscopy completion rates. The MyEleanor tool was also capable of rendering these services in Spanish.
Major Findings
After a period of 8 months, 57% (n = 1,368) of the patients involved in the study engaged with the MyEleanor tool via outreach by phone, and 58% of the patients who accepted these calls (n = 789)—or 33% of the overall population of the study—accepted live transfers. Among those who engaged with MyEleanor, the mean age was 56.7 years; 66% were female; 41% and 33% identified as Hispanic and Black, respectively; 73% and 25% were English- and Spanish-speaking, respectively; and 37% of them were in relationships.
Further, the rate of colonoscopy uptake among patients who did not adhere to their initial appointments rose from 10% in 2022 to 19% in 2023 following use of MyEleanor, thereby increasing patient volume by 36%.
The researchers indicated that almost 33% of the patients reported at least two barriers to colonoscopy—the most common of which were a lack of a perceived need to undergo screening (19%), limited time (18%), no health-care professional encouragement (16%), medical mistrust (14%), concerns about findings (13%), and concerns related to costs (12%). Additionally, a higher number of barriers to colonoscopy was correlated with actionable engagement. Patients who predominantly identified as Spanish-speaking and those who declined to disclose their racial identify reported nearly twofold the number of barriers.
Conclusions
The new findings revealed how AI-based virtual patient navigators could help mitigate patient nonadherence to colonoscopy, improve patient volume, and ultimately reduce disparities. The researchers plan to further examine the effects of MyEleanor on patient preparation adherence, staff burden, and revenue.
Disclosure: The research in this study was funded by internal resources dedicated to quality improvement at the Montefiore Einstein Comprehensive Cancer Center. For full disclosures of the study authors, visit coi.asco.org.
