Navigation auf uzh.ch
Navigation auf uzh.ch
Digital agents (DA) proliferate in healthcare. On the one hand, they are used for treatment, monitoring, and education of patients or as persuasive devices to induce behavioral change. On the other hand, they support healthcare professionals during provision of health services or diagnosis. DA as a digital system that offers a virtual character capable of various communication methods to help the user navigate their work or personal settings. A DA simulates intelligence and social skills. A DA for health (DA4H) is a digital agent designed to improve a patient’s health. The positive health effects occur through psychological or structural empowerment of relevant stakeholders including patients, healthcare professionals, or informal caregivers.
The existing research attends primarily to actual and envisioned benefits related to the use of DA4Hs in specific medical contexts. However, a common and overarching understanding of DA4Hs with relation to human agents in healthcare domain remains missing. In this stream of research we explore mutual relationships between the DAs and various stakeholders. Thereby we engage in multiple smaller design projects and scientific conceptualization with various partners from industry and academia.
Patients spend only a small part of their time in face-to-face sessions with their therapists, while important processes take place outside of therapy, i.e., between sessions (referred to as intersession processes and activities). Since smartphones are widespread, and their use is independent from location and time, DA4H in an app in a promising tool to complement psychotherapy in everyday life and to support intersession activities. The first prototype of the app including a chat-based DA4H has been finalized in Summer 2021. Subsequent steps are to evaluate the usability and acceptability together with therapists and patients.
Partner:
Department of Psychology: Clinical Psychology with Focus on Psychotherapy Research (Gvantsa Jinashvili, Dr. Markus Wolf, Prof. Dr. Birgit Watzke)
Temporomandibular disorders and chronic orofacial pain can have many different roots. Those disorders often occur as a result of psychomotor problems and their treatment require adequate patient information rather than interventional procedures. Information therapy includes the provision of information on symptoms such as pain, discussing symptom-related biological and psychological aspects, as well as the consideration of psychosocial context factors. Before admitting patients to an information-therapy specialist it is important to learn as much as possible about patient's psychosocial and physical condition. However, many patients dislike the long questionnaires used so far and feel exhausted when filling them out. In this project we develop a chat-based DA4H who helps the patient feel out the survey, cheers them up, and identifies questions to be asked based on free-text chat exchange with the patient.
Partners:
Department of Psychology: Clinical Psychology with Focus on Psychotherapy Research (Dr. Markus Wolf)
juli provides a platform to help people with chronic conditions better manage their health. The platform has a mobile app which gathers data from the user's wearables (e.g., fitness tracker or digital scale), the user's surrounding (weather, airpolution, pollen) and data entered by the user (daily mood and health questions, intake, or medication adherence). juli measures a user's health state, provides insights about it and personalised recommendations on how to improve it. A DA4H based on multimodal AI can help the user make sense of the data and to guide them. It make the interaction with juli more engaging and interactive, thus possibly improving adherence to the recommendation, and improve patient's decisions.
Partner:
juli (Joris Straatman, Bettina Duehrkoop)
A significant amount of research on the use of artificial intelligence (AI) in psychiatry, psychology, and psychotherapy has been conducted. At the same time, there is a need for a qualitative analysis of this conducted research, including a systematic assessment of new technologies' benefits and potential harm. The development of AI technologies and their application in mental health practice are important, they can make significant changes in this field in modern times. This study aims to examine the quality and accuracy of AI-generated responses compared to human-generated responses in mental health diagnosis and treatment. The study involves certified psychologists or psychotherapists creating diagnoses and providing answers to specific cases, both independently and in collaboration with a chatbot based on OpenAI's GPT language models. And involves evaluating the generated answers from the perspective of potential patients to assess their quality, empathy, and trustworthiness. Besides examining the potential and limits of AI-assisted health coaching in the psychological and psychotherapeutic domain, the study also explores the collaborative potential between certified mental health professionals and AI language models. By evaluating the generated answers from the perspectives of both specialists and potential patients, valuable insights can be gained regarding the integration of AI in guided medical and psychological coaching and its impact on patient care.
