Design and Development of a Conversational Agent for Cancer Patients

• Autor(a) principal: Félix, Beatriz Jorge
• Data de Publicação: 2022
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10362/138630
• topic: Chatbots; Co-design; Conversational agents; Cancer patients; eHealth; Dialogflow; Domínio/Área Científica::Engenharia e Tecnologia::Outras Engenharias e Tecnologias

Cancer is a major public health issue and represents the most significant burden of disease in Portugal. With the overall increase in cancer patients, there is growing demand for bet- ter, more patient-centred healthcare. The lack of reliable information about the disease, including treatments and symptoms, is one of the most common unmet needs of a cancer patient. Consequently, user-friendly tools should be created to empower patients by ad- dressing their concerns. Chatbots are efficient tools to bridge the communication between health providers and patients, and have already been used successfully in healthcare. This study explored how people perceive a cancer chatbot and viewed preferences and expectations pertaining to communication between a chatbot and newly diagnosed cancer patients. A user exploratory study was carried out involving observations, interviews with stakeholders and co-creation sessions. The observations and interviews revealed the overall difficulties and concerns of cancer patients. The results allowed the conversational agent to be targeted to newly diagnosed cancer patients. The co-creation sessions allowed the exploration of the expectations and preferences for the tone of voice, content, and human characterization, which were then applied to develop the chatbot. This work culminated in the development of a conversational agent capable of recog- nizing and classifying the intention of the user and providing a perceptive and prompt response. The chatbot was created with the help of a Natural Language Understanding framework – Dialogflow. The chatbot was evaluated through user testing, scoring above the benchmark of the questionnaires employed in the usability test, and its completion rate was 82.9%. Further research and testing is needed, but these results demonstrated the viability of using a chatbot in a cancer hospital. By adding more data, the chatbot’s performance is expected to improve.