Collaborative Patient Portals: Computer-based Agents and Patients’ Understanding and Use of Numeric Health Information

Dan Morrow, Mark Hasegawa-Johnson, Thomas Huang, and William Schuh; AHRQ R21HS022948


This project emulates, in a Patient Portal environment, best practices in face-to-face patient/clinician consultations. Self-managing health often hinges on patients’ comprehension of and reasoning about numeric information such as lipid panel results. This especially challenges patients with limited numeracy. Such patients are more likely to misunderstand test results, which undermines health decisions, behaviors and outcomes. Technologies such as Patient Portals, though intended to help patients, may exacerbate this problem by expanding delivery of numeric information. Clinicians are traditionally a key to helping patients create gist representations by providing a context for interpreting numeric information. They do so through verbal (evaluative commentary) and nonverbal strategies (e.g., facial expressions and prosody) that help patients understand and act. Such strategies are associated with better health behaviors and outcomes. Our team has developed a CA with stress/intonation, facial expressions and other nonverbal cues to convey affective information; by using tools developed in that prior research together with new data, we expect to develop a CA that can interact with patients in a Collaborative Patient Portal, providing succinct interpretations of test results that help patients to understand the gist of their test results.


Funded by this grant

  • Renato F. L. Azevedo, Daniel Morrow, Mark Hasegawa-Johnson, Kuangxiao Gu, Dan Soberal, Thomas Huang, William Schuh , Rocio Garcia-Retamero, "Improving Patient Comprehension of Numeric Health Information," Human Factors Conference, 2015
  • Yang Zhang, Nasser Nasrabadi and Mark Hasegawa-Johnson, ``Multichannel Transient Acoustic Signal Classification Using Task-Driven Dictionary with Joint Sparsity and Beamforming,'' Proc. ICASSP 2015, 2591:1-5

Relevant prior work by this team


Useful External Links