As organizational processes have become more interconnected and interdependent, contextual factors have become central to both information systems and process management. Despite the importance of context, few studies investigate the influence of contextual factors on the structure of business processes. Thus, in this dissertation, I examine the role of contextual specifics in the structure of the clinical documentation process using data from electronic health records in outpatient clinics. The dissertation includes three essays. In the first essay, I address the influence of internal contextual factors on enacted complexity. The findings of the first essay provide a unique opportunity to theorize on the specialization in enacted complexity of process by examining the effects of: 1) the number of roles and 2) specialization index. Contrary to expectations, I find that complexity decreases when a greater number of roles are involved in the clinical process and the roles are highly specialized. In the second essay, I turn my attention to the effects of exogenous shocks: When routines are disrupted, are some patterns of action more likely to be affected than others? I show that cohesion (defined as the consistency of context between pairs of actions) has a particularly strong influence on the persistence of action patterns. In essay three, I suggest a path prediction model in a process based on action sequence and its contextual specifics. The model uses a recurrent neural network that models both the observed sequence of actions and the contextual factors in the process. As expected, the results show that context can improve predictive models. In the case of outpatient medical clinics, the strongest improvement in accuracy comes from two attributes: 1) the workstation (location) where work is performed and 2) whether or not the system has been upgraded. Together, these essays represent a rigorous framework for analyzing the role of context in organizational processes and routines.
Read
