The ENGAGE study: evaluation of a conversational virtual agent that provides tailored pre-test genetic education to cancer patients

Kala Visvanathan; Dana Petry; Michelle S McCullough; Betty May; Ramkrishnan Tenkasi; Nitin Sharma; Catherine A Klein; Angelisa Johnson; Gisselle Killian; Melissa Camp; Channing J Paller; Rima Couzi; Mary Wilkinson; Lisa Jacobs; Julie Lange; Danijela Jelovac; Michael A Carducci; Mehran Habibi; Gauri Naik; Ashwin Kotwaliwale

doi:10.1007/s11764-023-01495-x

The ENGAGE study: evaluation of a conversational virtual agent that provides tailored pre-test genetic education to cancer patients

J Cancer Surviv. 2023 Dec 8. doi: 10.1007/s11764-023-01495-x. Online ahead of print.

Authors

Kala Visvanathan¹, Dana Petry², Michelle S McCullough³, Betty May³, Ramkrishnan Tenkasi⁴, Nitin Sharma⁵, Catherine A Klein², Angelisa Johnson², Gisselle Killian³, Melissa Camp², Channing J Paller², Rima Couzi², Mary Wilkinson², Lisa Jacobs², Julie Lange², Danijela Jelovac², Michael A Carducci², Mehran Habibi², Gauri Naik⁵, Ashwin Kotwaliwale⁵

Affiliations

¹ Johns Hopkins School of Medicine and Bloomberg School of Public Health, #E6142 615 N. Wolfe Street, Baltimore, MD, 21231, USA. kvisvan1@jhu.edu.
² Johns Hopkins School of Medicine, Baltimore, MD, USA.
³ Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA.
⁴ Goodwin College of Business, Benedictine University, Lisle, IL, USA.
⁵ Optra Health, San Jose, CA, USA.

PMID: 38064163
DOI: 10.1007/s11764-023-01495-x

Abstract

Purpose: Novel approaches are needed to ensure all patients with cancer have access to quality genetic education before genetic testing to enable informed treatment decisions. The purpose of this study was to test the use of an artificial intelligence (AI) intervention for the delivery of genetic education by non-genetic providers to patients with cancer undergoing active treatment.

Methods: A conversational AI-based application was developed on the HealthFAX platform to provide tailored genetic education to patients with cancer and tested at Johns Hopkins Hospital between April 2021 and Feb 2022. Patients' responses around the adoption, use, and experience of the AI application were assessed.

Results: Out of 64 individuals who consented to the study, 51 accessed the tool. The responding participants had a mean age of 61 years (ranging from 30-90 years) with 39 individuals undergoing active treatment for breast cancer and 12 for advanced prostate cancer. All patients chose to complete the tool at home. The median time between study enrollment and AI application initiation was 1 day, and the median time to complete the application was 24 min. All participants in their survey responses felt that the tool was secure, easy to use, liked the convenience of viewing it at home, and felt it provided valuable information. Eighteen percent of participants viewed the application with a family member. Ninety-eight percent of the participants completed their genetic education prior to receiving their test results. In 16%, a pathogenic variant was identified.

Conclusions: The 51 patients who adopted the AI application were highly satisfied with its usability and convenience. Our results support the continued evaluation of this cost-effective AI application in a large-scale study.

Implications for cancer survivors: Tailored pre-test genetic education can be successfully delivered to patients with cancer undergoing active treatment via an AI application at their convenience.

Keywords: AI for cancer survivorship; AI in cancers; AI in healthcare; AI-assisted patient education; Conversational AI; Scaling clinical processes with conversational AI.