Sofia: a Conversational Health Agent

Study to investigate the effectiveness and the challenges of conversational agents to support the self-report of health and well-being

Study description

Pen and paper and online tools such as mobile and web apps are used in current practice to log daily diaries and self-report on health questionnaires for monitoring and assessment of mental health illnesses such as depression and bipolar disorder. While these methods are considered efficient, far too little attention has been paid in the collection of self-reports using voice. Recent advancements in natural language processing and the popularity of virtual assistants such as Amazon Alexa and Google Assistant offer a new way to collect self-reports through conversational agents where the users can verbally converse with the agents to report about their health and well-being. This could engage users in a more meaningful conversation which might foster more honest and insightful forms of self-disclosure.

To investigate the effectiveness of such agents in collecting self-reports, and to identify the challenges in designing them, we intend to conduct a between-subject field study involving 42 participants half of whom will be randomly assigned to use a web app (WA) and half will use a conversational agent (CA) named Sofia to log daily diaries and self-report a health questionnaire.


Participation in this study will consist of entering daily diaries and self-reporting to the WHO-5 questionnaire every two weeks via Sofia or a web app. Participation will also include pre-study questionnaires containing demographic information such as age and gender. Participants assigned to use Sofia will be provided with a smart speaker, other participants will use a web app on their own device. At the end of the study, participants will fill out a questionnaire regarding their experience and will be interviewed. 


Who can participate?

Volunteers over the age of 18 with previous or current diagnoses of affective disorder (bipolar, depression, anxiety or other mood related disorders). If you meet these criteria and are interested in participating in this study, please contact Raju Maharjan at


What data do we collect?

We collect demographic data on paper, platform usage data, and post-study interview recordings. We do not collect any voice data from Sofia.

The demographic data we collect include:

  • Name
  • Email
  • Age
  • Sex
  • Education level
  • Occupation
  • Health conditions
  • Technological ability

The data we collect from the platforms include:

  • Daily diary entries
  • Self-reports to the WHO-5 health questionnaire
  • Data reflecting your usage of the platform for the duration of the study

Informed Consent

Participation is voluntary. As a participant, you are free to withdraw your consent at any time, without the need to provide a reason, by contacting Raju Maharjan at

Who can access your information?

All data gathered during this study will be stored securely in accordance with the General Data Protection Regulations and other requirements. Participants’ anonymity will be maintained in all publications. The Technical University of Denmark, CVR no. 30060946, is the data controller responsible for processing personal data in this research project.
15 APRIL 2024