Background Chatbots are artificial intelligence–driven programs that interact with people. The applications of this technology include the collection and delivery of information, generation of and responding to inquiries, collection of end user feedback, and the delivery of personalized health and medical information to patients through cellphone- and web-based platforms. However, no chatbots have been developed for patients with lung cancer and their caregivers. Objective This study aimed to develop and evaluate the early feasibility of a chatbot designed to improve the knowledge of symptom management among patients with lung cancer in Japan and their caregivers. Methods We conducted a sequential mixed methods study that included a web-based anonymized questionnaire survey administered to physicians and paramedics from June to July 2019 (phase 1). Two physicians conducted a content analysis of the questionnaire to curate frequently asked questions (FAQs; phase 2). Based on these FAQs, we developed and integrated a chatbot into a social network service (phase 3). The physicians and paramedics involved in phase I then tested this chatbot (α test; phase 4). Thereafter, patients with lung cancer and their caregivers tested this chatbot (β test; phase 5). Results We obtained 246 questions from 15 health care providers in phase 1. We curated 91 FAQs and their corresponding responses in phase 2. In total, 11 patients and 1 caregiver participated in the β test in phase 5. The participants were asked 60 questions, 8 (13%) of which did not match the appropriate categories. After the β test, 7 (64%) participants responded to the postexperimental questionnaire. The mean satisfaction score was 2.7 (SD 0.5) points out of 5. Conclusions Medical staff providing care to patients with lung cancer can use the categories specified in this chatbot to educate patients on how they can manage their symptoms. Further studies are required to improve chatbots in terms of interaction with patients.

BACKGROUND Chatbots are artificial intelligence-driven programs that interact with people. The application of this technology is wide—including collecting and delivering information, creating and answering inquiries, collecting customer feedback, and even delivering personalized health and medical information to patients through mobile and web-based platforms. However, there are no chatbots for lung cancer patients and their caregivers. OBJECTIVE Our research objective was to develop and evaluate the early feasibility of a chatbot designed to improve symptom management knowledge of patients with lung cancer and their caregivers. METHODS We conducted a sequential mixed method approach that included conducting a web-based anonymized questionnaire survey of physicians and paramedics from June to July 2019. Two physicians conducted a content analysis of the questionnaire to organize it into frequently asked questions (FAQs). Based on the FAQs, we developed and integrated a chatbot into a social network service. The chatbot was then tested by the physicians and paramedics (alpha test), thereafter, by the patients or their caregivers (beta test). RESULTS We obtained 246 questions from 15 healthcare providers. After the test, we identified 91 FAQs and their corresponding answers. In the beta test, 11 patients and one caregiver participated. The participants were asked 60 questions. Eight questions (13%) did not match the appropriate categories. After the beta test, seven participants (64%) answered the post-experimental questionnaire. The mean satisfaction score was 2.7 points out of 5 (standard deviation: 0.5). CONCLUSIONS Medical staff caring for lung cancer patients would be able to use the categories specified in this study to educate patients. Further research is warranted to improve chatbots in terms of interaction. CLINICALTRIAL Chatbot; lung cancer; symptom management education; mixed method approach