JMIR Publications

ABSTRACT

Background:

Electronic health record information systems' continuous application has accumulated enormous medical data with potential value. Semantic interoperability is the premise of mining and realizing these values by sharing and reusing vast scattered, heterogeneous, and multi-source clinical data. Several initiatives have been developing information models to realize and improve semantic interoperability, among which the openEHR model is one of the outstanding information models. Reusing archetypes from Clinical Knowledge Manager (CKM) instances is the backbone of achieving semantic interoperability based on the openEHR approach. Archetype reuse among different CKMs is of great significance to achieve a broader range of semantic interoperability. However, the reuse of archetypes among different existing CKMs is still unclear.

Objective:

This study aims to investigate the reuse of openEHR archetypes across the CKMs approved by openEHR International for achieving and improving semantic interoperability.

Methods:

This study analyzes and compares archetype reuse across the given five CKMs (Fi-CKMs) and across the four CKMs (Fo-CKMs), respectively. Firstly, the analysis and comparison of reused archetypes are executed in terms of: archetype reuse number and archetype reuse ratio. Then, data elements of the given 25 reused archetypes were analyzed and compared across the Fi-CKMs and across the Fo-CKMs, including the type, number and ratio of the reused data elements.

Results:

The archetype reuse numbers across the Fi-CKMs and the Fo-CKMs are 25 and 58, respectively. Comparison of archetype reuse ratio showed that reducing one CKM increased the archetype reuse ratio from 5.56% - 19.23% to 12.89% - 44.62%. The analysis and comparison of data elements across the Fi-CKMs and the Fo-CKMs show that reducing one CKM increased the reusable data elements from 231 to 355, increased the data element reuse ratio from 54.20% - 70.64% to 90.33% - 100%, and increased the data element direct ratio from 89% to 100%.

Conclusions:

A large number of archetypes have been accumulated in the existing CKM instances, which provides an important foundation for the reuse of archetypes needed for semantic interoperability. With the decrease of the involving person, archetype reuse is improved, including the number of the reused archetype, the number of data element reuse, and the ratio of data element direct reused. A more effective coordination mechanism between multiple CKMs needs to be established to promote archetypes reuse across CKMs.