Empowering Cancer Patient-Provider Communication: A Personal Virtual Assistant for Automated Post-Discharge Engagement

Baylor College Of Medicine, Houston TX

Investigators

Linked publications & trials

Abstract

Project Summary/Abstract This application is submitted in response to the Notice of Special Interest (NOSI) identified as NOT-CA-23-041. Inadequate communication between patients and providers following hospital discharge poses a significant challenge for oncology patients. The complex, multidisciplinary care these patients require often results in insufficient follow-up and misunderstandings of treatment plans, leading to increased anxiety, stress, and reduced compliance. This in turn elevates the risk of hospital readmission, cancer recurrence, heightened suffering, and even mortality. To address these issues, our team has collaborated with the Dan L Duncan Comprehensive Cancer Center (DLDCCC), supported by a P30 grant, to develop an AI-based Personal Virtual Assistant (PVA) designed to enhance the frequency, quality, and timeliness of patient-provider communication. The PVA is a hands-free, voice-activated AI-driven system that utilizes care plan data for both patients and healthcare providers, fostering continuous communication. The PVA comprises: 1) a voice-enabled, user-friendly AI kiosk-based interactive tablet that presents patients with essential plan details, such as reviewing discharge instructions, receiving post-discharge reminders (e.g., medication, care tasks), providing guidance for exercise and stretching therapy, and facilitating bi-directional messaging with clinical staff; and 2) a comprehensive web portal that enables care providers to monitor compliance, modify patient-centered multidisciplinary care plans, engage in shared decision-making, and remotely update the PVA tablet to prevent care overlap. The web portal can be accessed through any computer, tablet, or phone app. Our pilot study has demonstrated the feasibility and acceptability of the PVA among oncology patients and clinical experts. Building on this preliminary data, we propose a supplemental grant to further assess the PVA's feasibility, scalability, and proof-of-concept effectiveness in enhancing patient-provider communication and adherence to care plans. We will conduct a pilot randomized controlled trial (RCT) with 20 oncology patients undergoing surgical procedures to evaluate the improvement of communication, acceptability, and scalability of the PVA, as well as its effectiveness in reducing post-discharge adverse events. This supplemental grant will help establish the necessary power for clinically validating the PVA's acceptance and effectiveness in mitigating adverse events associated with inadequate patient-provider communication.