Opportunistic Decision Making Information Needs and Workflow in Emergency Care

In high-risk healthcare settings such as the Emergency Department (ED), managing information needs and supporting clinical decision making is critical for patient safety and healthcare quality. Rather than focusing on a single task at a time, ED clinicians are forced to switch between multiple tasks and multiple patients. Many of these task transition decisions are based on unplanned, unorganized, and unpredictable environmental factors. Rather than acting on a global understanding of the department, clinicians make opportunistic decisions, or those that are informed by the local level, such as selecting the next patient based on proximity. This high level of complexity in the ED is a major contributing factor to potentially preventable adverse events.

This project studied information management and decision making among ED clinicians and focused on task transition decisions. The goal was to develop a Work Domain Ontology (WDO) to capture the essence of work in the ED, including the clinical goals, information needs, and the clinical operations of an ED that are required for patient care. The WDO was used to inform evaluation of dashboard displays.

The specific aims of this project were as follows:

• Develop and validate the WDO of the ED. 
• Identify the information needs and understand the mechanisms and impacts of opportunistic decision making. 
• Develop visualizations for increasing situational awareness and supporting decision making. 
• Evaluate the impact of the visualizations as cognitive interventions. 

Interviews, surveys, and ethnographic observations were used to capture participant workflows and information needs for completing tasks within the ED at three hospitals. Five displays providing a range of information and visual supports were developed using the WDO. The dashboards included: (1) the number of current patients, severity score for each patient, current location of patients, and available beds; (2) status of laboratory and imaging orders; (3) contributing factors to the unit’s current overcrowding score; (4) throughput measures including the duration of each patient’s current stage of care; and (5) workload information for each clinician.

The displays were implemented at 11 hospitals with the aim of providing actionable information for real-time decision making. Formative assessments indicated the dashboards were readily accepted; however, some degree of fine-tuning was needed, such as extending the patient history tracking window from 12 hours to 15 hours and refining timeline graphics. Eye-tracking results indicated users quickly made use of dashboard features and tools. Evaluation of long-term use of the dashboard on quality measures is a potential next step for this project.