Interactive Process Mining in Surgery with Real Time Location Systems: Interactive Trace Correction

Abstract

The use of Real Time Location Systems in combination with Process Mining solutions can produce precise and unobtrusive methods for discovering and analyzing models in hospitals. However, noisy environments and building structures can interfere with RTLS signals and can affect their accuracy. This noise produces an undesirable effect on Process Mining algorithms called Spaghetti Effect. This effect dramatically decreases the human understandability of Workflows produced by the Process Mining discovery algorithms. The use of interactive paradigm for supporting the semi-automated correction of this data can suppose a solution to this problem. In this work, we present an Interactive Trace Correction method based on Error-Correcting (EC) paradigm and using an Edit Distance framework, to correct the data and obtain an insight of the executed process. This method was tested in a surgical area of a hospital in spain using the existing RTLS infrastructure deployed in that environment with 3915 patients collected between March 2012 and May 2013.