Back to search

IKTPLUSS-IKT og digital innovasjon

Virtual House Doctor

Awarded: NOK 0.50 mill.

Can we provide timely diagnostic judgment for outpatients with chronic neurologic disorders? An envisioned system collects passive and elicited bio-behavioral data and links these data to a patient's cumulative bio-behavioral profile and medical records to detect significant events and help guide the management of the patient's care. Developing such a diagnostic monitor will require ICT advances. First, automated in-situ collection of data needs to be so simple, unobtrusive and helpful that it is widely accepted by outpatients; while it has to be accurate enough to reliably deliver key information that drives an effective diagnostic process. Second, an automated diagnostic process needs to learn how to combine an individual's continuous incoming data with that individual's medical records so that the process can approach the accuracy of a skilled doctor. These requirements suggest four research challenges (A) Find the most acceptable and reliable ongoing patient information streams; (B) Identify the most specific and useful patient data for early diagnosis and monitoring; (C) Build an accurate diagnostic inference engine that merges ongoing data streams with individual medical histories; and (D) Find user-interfaces that promote patient acceptance and compliance. To address these challenges, VHD team will build a research system to collect data, and then develop and evaluate diagnostic algorithms. The piloting system will track outpatients' location, movement, and activities and analyze them in relation to that person's typical daily patterns. It will also implement several interaction modalities to elicit verbal and other task behavior to probe for symptoms and measure signs. The core project focus will be the development of an accurate inference engine that uses deep learning to select among diagnostic responses. If the project has success with neurological conditions, it will expand to other chronic conditions and be made available to other research groups.

Funding scheme:

IKTPLUSS-IKT og digital innovasjon