Related Experiment Video
Updated: Jun 3, 2026
E-Patient Counseling Trial (E-PACO): Computer Based Education versus Nurse Counseling for Patients to Prepare for Colonoscopy
Published on: August 1, 2019
Feasibility and Preliminary Effectiveness of the ChulaCancer Mobile Chatbot for Supportive Care of Patients With
Narawitch Sompornpailin1, Thiti Susiriwatananont1, Virote Sriuranpong1
1Division of Medical Oncology, Department of Medicine, Faculty of Medicine, Chulalongkorn University & The King Chulalongkorn Memorial Hospital, Bangkok, Thailand.
Background:
Chemotherapy-related toxicities often lead to unscheduled health care use and diminished quality of life. Digital health interventions, such as chatbots, offer a scalable solution for supportive care; however, evidence regarding their effectiveness in resource-limited, low- and middle-income settings remains limited.
Objective:
This study aimed to evaluate the feasibility, use, and preliminary effectiveness of a closed-loop chatbot (ChulaCancer Chatbot) in reducing unscheduled hospital visits and stabilizing quality of life among patients receiving chemotherapy for breast or colorectal cancer.
Methods:
This pilot randomized controlled trial enrolled 40 patients at a single academic center in Thailand, randomized 1:1 to either ChulaCancer chatbot plus usual care or usual care alone. The primary end point was the proportion of unscheduled hospital visits due to chemotherapy-related toxicities within 12 weeks of treatment initiation. Secondary end points included longitudinal quality of life changes (30-item EORTC Quality of Life Questionnaire) measured at baseline, following chemotherapy cycle 2, and following cycle 4. Use metrics were extracted from the chatbot platform. Data were analyzed using the Fisher exact test and linear mixed-effects models.
Results:
The platform recorded 2393 total messages with a 70.5% (503/713) successful response rate for user-initiated queries. Unscheduled hospital visits occurred in 15% (3/20) of the chatbot group compared to 35% (7/20) of the usual care group (P=.24). While infection-related visits were similar between groups, the usual care group recorded multiple visits for low-acuity symptoms (eg, anxiety, headache, and edema) that were absent in the chatbot group. Regarding quality of life, the chatbot group demonstrated a significant mitigation of cancer-related fatigue following cycle 4 compared with the usual care group (P=.02 between groups). Additionally, the chatbot group significantly improved in global health status (P=.04) and avoided the decline in physical functioning observed in the control arm (P=.04).
Conclusions:
The integration of a closed-loop chatbot into oncology care is feasible and provides a potential secure triage mechanism that may reduce acute care use for low-acuity concerns. Future large-scale trials incorporating agentic artificial intelligence are warranted to further validate clinical and economic benefits.
Trial Registration:
Thai Clinical Trials Registry TCTR20251220014; https://tinyurl.com/5b6k3e63.
