277 The Impact of Decision Fatigue on Ancillary Test Ordering in the Emergency Department

Abstract

Decisionmaking capability may decrease with repetitive cognitive demand, with a resulting increased reliance on the least cognitively demanding option. We hypothesize that increasing decision fatigue due to increased cognitive demand, resulting from an increasing number of hours on shift, number of clinical encounters, hourly active patient burden, and hourly new patient burden will demonstrate a positive correlation with an increased number of ancillary laboratory and radiologic studies ordered and relied upon in the management of emergency department patients. We also hypothesize that absolute hour of the day will demonstrate a bi-modal increase in rate of ordering during nontraditional work hours. This study utilizes de-identified patient information, collected from the EPIC-based Clinical Research Database, at a single academic institution, related to emergency department visits from January through June 2016. Physician-level data collected includes physician identifying number, shift clock-in and clock-out time, absolute number of ancillary studies ordered by hour of shift and per patient encounter, new patients by physician, and patients transferred to physician from prior shift. Encounter (patient)-level data collected includes encounter number, time of patient arrival, time of patient room assignment, patient triage acuity, physician assignment and time, time stamps for each ancillary study ordered, ordering physician for each ancillary study ordered, and type and time of disposition. The rate ratios and rate of ordering laboratory and radiologic studies were evaluated by shift hour, total patients seen, hourly active patient burden, hourly new patient burden, and absolute hour of the day. The data demonstrates an increased rate (2.57% [95% confidence interval 1.05%-4.12%], p<.001) of ordering laboratory studies but decreased rate (5.73% [0.34%-10.83%], p=.04) of ordering computed tomography (CT) imaging studies as shift duration increased. The data also reflects an increased rate of ordering laboratory and imaging studies associated with an increase in total patient encounters over the course of the shift (laboratory: 5.62% [1.05%-6.42%], p<.001; CT: 3.18% [1.01%-5.93%], p=.02; plain radiographs: 4.15% [1.63%-6.73%], p=.001; advanced imaging such as magnetic resonance imaging, ultrasound, or nuclear: 6.12% [3.40%-8.91%], p<.001), as well as with an increase in hourly active patient burden (laboratory: 6.10% [5.10%-7.10%], p<.001; CT: 5.00% [1.80%-8.20%], p=.002; plain radiographs: 6.00% [3.10%-9.00%], p<.001; advanced imaging: 5.80% [3.00%-8.80%], p<.001). The data did not show significance for correlation between increasing hour of the day and rate of CT or plain radiograph test ordering, but did show a decreased rate of laboratory test ordering (0.66% [0.19%-1.13%], p=.006) and decreased rate of advanced imaging (2.21% [0.24%-4.14%], p=.03) starting from midnight and on through the day. When cognitively burdened and mentally fatigued, emergency physicians may more heavily rely upon ancillary laboratory and radiologic studies in the diagnostic process.