Representing and reasoning about contextually changing organizational behavior using simulation models of medical work

Medical practice is a team-based activity requiring coordination among physicians, nurses, and support staff. The decisions each clinician makes can affect the work of the others unpredictably. Systematic differences in healthcare activities based on non-clinical factors have been documented. These differences arise because of the work process context of clinical decisions. Healthcare administrators need tools to understand how an organization's structure, resources, and work processes interact to affect care delivery.; I have developed a system, the Context Aware Virtual Health Administration Team (CAVHAT). CAVHAT uses agent-based simulation to links changes in individual behavior to their effects on organizational performance. CAVHAT represents the actors in an organization: their roles, skills, experience levels, and work tasks. Non-clinical process contexts arise dynamically during simulation, causing predictable deviations in actor behavior. CAVHAT extends prior research of the Virtual Design Team research group by adding context-driven variations in actor behavior. This work focuses on four contexts that are systematic, ubiquitous, and documented in the medical literature. The goals of this research are two-fold: to represent the heuristics used by actors in response to contexts, and to predict the effects on organization-level outcomes.; CAVHAT has been validated against published literature and with an empirical study of a medical organization. Results show that contextual behaviors have a significant effect on organizational performance, both positive and negative. The organization studied relies upon contextual responses in order to keep abreast of the planned workload. This can cost the organization in terms of poorer communication and coordination quality. Sensitivity analyses show that contextual behaviors differ in their degree of impact on organizational performance. Some of the benefits of locally optimizing contextual behaviors work in opposition, leading to no net benefit in organizational performance.; Contextual changes are an understudied phenomenon with potentially large effects on work process execution. CAVHAT's representation and reasoning framework can be used to study service organizations involved in non-routine, diagnose-and-repair tasks. The results can be used to design work protocols that insure uniformity of care.