Page 25 - AIH-1-3

Artificial Intelligence in Health Optimizing EHRs to support AI

Figure 1. OpenEHR multi-level modeling . Copyright © 2007 Author(s)
87

Figure 2. Ontology-based framework for an overall system architecture (adapted from a multitude of similar images developed to represent the openEHR
architecture and its relationships to external health data sources, exchange standards, and outcomes). Copyright © 2017 Author(s)
80
systems are applied to these medical data, there is an as part of routine clinical practice. Post-ad-hoc data
imminent and real danger of feeding AI algorithms with collection has been shown to be very expensive and error-
non-optimal EHR data which are highly likely to end up in prone, and at times, it is impossible to capture the clinical
the “garbage in garbage out” problem. 90 context in which the data were collected. Data sources
91
There is an opportunity to capture high-quality and can be very diverse and range from operational EMR
complete structured and computable health-care data systems to well‐structured longitudinal disease registries

Volume 1 Issue 3 (2024) 19 doi: 10.36922/aih.3056