Monitoring adherence to evidence-based practices: a method to utilize HL7 messages from hospital information systems

Clinical pathways are evidence-based recommendations for treating a diagnosis. Although implementations of clinical pathways have reduced medical errors, lowered costs, and improved patient outcomes, monitoring whether a patient is following the intended pathway is problematic. Implementing a variance reporting program is impeded by the lack of a reliable source of electronic data and automatic retrieval methods.

Our objective is to develop an automated method of measuring and reporting patient variance from a clinical pathway.

We identify a viable and ubiquitous data source for establishing the realized patient's path- Health Level Seven (HL7) formatted message exchanges between Hospital Information Systems. This is in contrast to current practices in most hospitals where data for clinical pathway variance reporting is obtained from multiple data sources, often retrospectively. This paper develops a method to use message exchanges to automatically establish and compare a patient's path against a clinical pathway. Our method not only considers pathway activities as is common practice, but also extracts patient outcomes from HL7 messages and reports this in addition to the variance.

Using data from our partner hospital, we illustrate our clinical pathway variance analysis tool using major joint replacement patients. We validate our method by comparing audit results for a random sample of HL7 constructed pathways with data extracted from patient charts. We report several variances such as omitted laboratory tests or additional activities such as blood transfusions. Our method successfully identifies variances and reports them in a quantified way to support decisions related to quality control.

OUR APPROACH DIFFERS FROM PREVIOUS STUDIES IN THAT A QUANTITATIVE MEASURE IS ESTABLISHED OVER THREE DIMENSIONS: (1) omissions from the pathway, (2) additions to the pathway, and (3) patient outcomes. By examining variances providers can evaluate clinical decisions, and support quality feedback and training mechanisms.