Autodok

Almost all acute hospital settings use observation protocols ranked by patient severity to ensure that potential deterioration is identified in time. To achieve this, the departments utilise advanced patient monitors for tracking the state of admitted patients. These monitors are typically capable of continuously tracking and reporting vital values at a much higher rate than documented in clinical knowledge management systems.

Previous research has shown that most patients are not monitored closely enough, while a subset of patients are monitored more than required by clinical protocols. The latter situation, which also applies to patients who are continuously monitored, are poorly represented by spot-based registrations. The former situation complicates clinical assessment and prediction.

AIM

In the AutoDok project, the Patient Deterioration Warning System was extended with functionality for handling clinical documentation for patients who were extensively monitored and patients who lacked monitoring during periods of their admission. Furthermore, the AutoDok project aimed to explore integration with an EHR system (EPJ Syd) through the Medical Device Information Collection (MDIC) platform.

RESULTS

Functionality was developed to calculate summary values ​​for hospitalised patients, and new algorithms for inserting unknown vital values ​​were tested. Unfortunately, these were not integrated into clinical operations due to covid-19, among other things.

PARTNERS

The project was led by Thomas Schmidt and was carried out in collaboration with the Centre for Clinical Artificial Intelligence (CAI-X).

Partners on the project are the Emergency Department at Odense University Hospital and the Maersk Mc-Kinney Moller Institute at the University of Southern Denmark.

EXTERNAL FUNDING

The project was funded by the Innovation Fund of the Region of Southern Denmark.