Published on
Modeling clinical workflow is an important pre-cursor to understanding how clinicians interact with an EMR and its features. Common methods of assessing clinical workflow include video recording or direct observations, or through directed user experience testing in a controlled environment. Assessing clinical
workflow within an EMR can also be accomplished by analyzing EMR log data, which can provide an unbiased view of EMR use. Using log data to build workflow patterns requires appropriate pre-processing of the log data, and the use of several types of data mining techniques: frequent pattern mining, sequence mining, and sub-graph mining. This talk will discuss the challenges of log data preprocessing and workflow mining, and touch on sequence mining and sub-graph mining techniques.