Abstract
Objective Self-anamnesis is a procedure in which a patient answers questions about the personal
medical history without interacting directly with a doctor or medical assistant. If
collected digitally, the anamnesis data can be shared among the health care team.
In this article, we introduce a concept for digital anamnesis collection and assess
the applicability of a conversational user interface (CUI) for realizing a mobile
self-anamnesis application.
Materials and Methods We implemented our concept for self-anamnesis for the concrete field of music therapy.
We collected requirements with respect to the application from music therapists and
by reviewing the literature. A rule-based approach was chosen for realizing the anamnesis
conversation between the system and the user. The Artificial Intelligence Markup Language
was exploited for encapsulating the questions and responses of the system. For studying
the quality of the system and analyzing performance, humanity, effect, and accessibility
of the system, we performed a usability test with 22 persons.
Results The current version of the self-anamnesis application is equipped with 63 questions
on the music biography of a patient that are asked subsequently to the user by means
of a chatbot conversation. The usability study showed that a CUI is a practical way
for collecting anamnesis data. Users felt engaged of answering the questions and liked
the human characteristics of the chatbot. They suggested to extend the conversation
capabilities of the chatbot so that the system can react appropriately, in particular
when the user is not feeling well.
Conclusions We could demonstrate the applicability of a CUI for collecting anamnesis data. In
contrast to digital anamnesis questionnaires, the application of a CUI provides several
benefits: the user can be encouraged to complete all queries and can ask clarifying
questions in case something is unclear.
Keywords
intelligent assistant - patient empowerment - self-anamnesis - patient safety - AIML
- conversational user interface
