Analyzing interactions and processes in an adaptive and evidence-based consumer health information system.

Published on: 2024

Bedek, M., Schramm, N., Lengauer S., Krenn, C., Albert, D., & Schreck, T.

The A+CHIS project develops an interactive and adaptive Consumer Health Information System (CHIS) that incorporates information on Type II Diabetes Mellitus (T2DM). More and more people are relying on the internet, social media or hallucinating LLMs as their source of health information, with varying degree of credibility. Many existing CHIS are of poor quality and not suitable as a basis for informed decisions by patients. Since patients have different pre-knowledge and information needs, a static one-fits-all approach seems inadequate. To counter these critical issues, A+CHIS aims to provide strictly evidence-based information (curated by medical experts) in an interactive and adaptive way. To individually tailor user recommendations, tracking and interpreting the behaviour is necessary. As part of an evaluation study, twelve participants with different levels of pre-knowledge on T2DM and computer literacy, applied a cognitive walkthrough consisting of eleven tasks that required different A+CHIS components. During the walkthroughs, the participants were asked to think aloud. The audio and on-screen recordings enabled to identify processes, ranging from ‘basal’ (e.g., mouse clicks) to more cognitive ones (e.g., reading), the order in which they were applied as well as their durations. While ‘basal’ processes can be tracked automatically and quite easy, some of the more cognitive processes can only be assessed indirectly. Different approaches like the lag-sequential analysis or metrics inspired by behavioural mapping and graph theory, are suitable to analyse patterns in the behaviour. These analyses can be used to assess efficiency and make recommendations for system improvement or user recommendations.