Putting the SWORD to the Test: Finding Workarounds with Process Mining

Abstract

Workarounds, or deviations from standardized operating procedures, can indicate discrepancies between theory and practice in work processes. Traditionally, observations and interviews have been used to identify workarounds, but these methods can be time-consuming and may not capture all workarounds. The paper presents the Semi-automated WORkaround Detection (SWORD) framework, which leverages event log traces to help process analysts identify workarounds. The framework is evaluated in a multiple-case study of two hospital departments. The results of the study indicate that with SWORD we were able to identify 11 unique workaround types, with limited knowledge about the actual processes. The framework thus supports the discovery of workarounds while minimizing the dependence on domain knowledge, which limits the time investment required by domain experts. The findings highlight the importance of leveraging technology to improve the detection of workarounds and to support process improvement efforts in organizations.