ABSTRACT
Integrating biomedical knowledge with patients’ health record systems is a common approach to helping clinicians make more informed decisions. We aim to explore and compare different integration mechanisms and identify integration paradigms to enlighten future improvement. We searched PubMed (n = 471) and set up the initial inclusion/exclusion criteria, and the initial paradigms via a pilot classification (n = 61). We then conducted brainstorming sessions to achieve consensus. We then compared, discussed, refined the results, and formed the integration paradigms. We then reviewed, selected, and classified the remainder of the publications (n = 410). There exist at least five knowledge integration paradigms: knowledge inscription, knowledge catalog, knowledge agent, expert systems, and knowledge modeled as an ontology among 73 publications. Integration mechanisms, 238scalability, maintenance, and reusability of the knowledge sources were reviewed and compared among paradigms. The knowledge modeled as an ontology paradigm has advantages in scalability, reusability, and maintainability. However, computational efficiency can be a challenge. Our review provides an initial summary of approaches to knowledge integration with patients’ records. The work demonstrates the quick transitions among different paradigms when integrating knowledge with patient records and highlights the roles of ontology applications in healthcare computing.
