A system for the semi-automatic evaluation of clinical practice guideline indicators

This paper presents EXEMED v2, a system that allows the evaluation of clinical practice guideline indicators. EXEMED v2 includes a knowledge base that supports the definition of executable rules applied over Electronic Health Records (EHR) in order to measure its compliance with a specific clinical guideline. Taking into account that an EHR may include structured attributes and narrative text attributes, EXEMED v2 analyzes both types. The process of evaluation in EXEMED v2 is to define the rules; once the rules are defined EXEMED v2 extracts from the EHR the facts that allow evaluating whether each one of them was accomplished or not. This evaluation includes different levels of certainty, allowing in some cases the interaction of a human evaluator to confirm (or not) automatic evaluation decisions. The functionality of EXEMED v2 was validated applying it in a case study of Acute Myocardial Infarction.