This project aims at sharing clinical decision support (CDS) knowledge that has been proven in practice. CDS knowledge interchange standards have existed for as long as clinical data interchange standards (about 2 decades), yet sharing of computable, actionable knowledge in practice is extremely limited. Diverging forces abound in the highly innovative CDS research field and developed methods and content are deeply incompatible even though each piece may have been validated in practice. This project will impact the present CDS knowledge sharing impasse at its roots by creating, validating and refining a methodology for conceptual integration of CDS knowledge through systematic application on a large body of high value, practically validated, but incompatible resources. Initially a web-based Clinical Knowledge Library is established that will grow in contents and in its degree of integration. The contents include (i) the Regenstrief Institute's data dictionaries and complete CDS knowledge base;(ii) the published Columbia University Medical Center'Arden Syntax Medical Logic Modules (MLM), (iii) MLMs of the Child Health Improvement through Computer Automation (CHICA) system;(iv) FDA/NLM published Structured Product Labels (SPL) enhanced with the VA's NDF-RT medication knowledge base, (v) Clinical data sets and templates published by various continuity of care projects in the U.S., Netherlands, U.K. NHS, Australia, (vi) The CDC Data Elements for Emergency Departments, clinical LOINC panels, and other detailed clinical models, (vii) Cancer data protocols (e.g., College of American Pathology Cancer Protocols). Applications and tools for editing and adapting the content, as well as for implementation on a diverse set of systems will also be disseminated. The integration of content will reach all levels: (1) lexical integration, i.e., mapping of all idiosyncratic codes to standard terminologies;(2) syntactic integration across different representation languages including Arden Syntax Medical Logic Modules (MLM) and Regenstrief CARE and G-CARE languages;(3) semantic integration, including the transformation procedural knowledge (e.g., CARE and MLMs) into declarative ones (e.g., G- CARE);and finally (4) pragmatic integration across different knowledge modalities including (a) CDS reminder rules, (b) declarative CDS knowledge representations (e.g., drug-interaction tables), (c) definitions of calculated observation rules (e.g., glomerular filtration rate estimate), (d) clinical data sets, templates and assessment instruments, (e) healthcare quality measures, and (f) clinical research queries for both individual and population data. Thus widely useful CDS knowledge will be unlocked from its concrete representations, conceptualizations and applications, and adaptation and adoption will be simplified by the software developed and released in open source collaborations.
The project will create new methods to translate and integrate computer information to make medical decision support information shareable. This information will be made easily accessible to all by establishing a public library of harmonized, high value computer information to support better clinical decisions, and thus help improve the quality and reduce the cost of health care.
