Capturing Real-Time Emergency Department Sentiment: A Feasibility Study Using Touch-Button Terminals

• Published in: Annals of emergency medicine 2020-06, Vol.75 (6), p.727-732
• Main Authors: Agarwal, Anish K., MD, MPH, MSHP, Hahn, Lauren, BFA, Pelullo, Arthur, MS, Rosin, Roy, MBA, Merchant, Raina M., MD, MSHP
• Format: Article
• Language: English
• Subjects: Emergency
• ISSN: 0196-0644; 1097-6760
• DOI: 10.1016/j.annemergmed.2019.07.004

Study objectiveProviding care in emergency departments (EDs) affects patients and providers. Providers experience high rates of work-related stress. Little is known about the feasibility of measuring real-time sentiment within busy clinical environments. We test the feasibility of measuring sentiment with touch-button terminals in an academic, urban ED. MethodsTerminals offered a choice of 4 sentiment buttons (very positive, positive, negative, and very negative). They were placed central to physician workstations, nursing workstations, and the patient exit. Pearson correlation coefficients ( r) were calculated to estimate correlation between sentiment and volume metrics (arrivals, length of stay, waiting patients, and number of boarding patients) over time. ResultsA total of 13,849 sentiments were recorded (June 2018 to October 2018); 9,472 came from providers (52.6% nursing) and 4,377 from patients. The majority of provider sentiments were negative (58.7%). Negative provider sentiment was associated with increasing number of patients waiting to be seen ( r=0.45) and boarding ( r=0.68). Positive provider sentiment was associated with increasing numbers of patients who left without being seen ( r=0.48). Increased boarding was associated with more recorded sentiments ( r=0.73). Negative patient sentiment was associated with increasing number waiting ( r=0.55), boarding ( r=0.67), and leaving without being seen ( r=0.46). ConclusionThis study demonstrates the feasibility of a novel approach to measuring “on-shift” sentiment in real time and provides a sample comparison to traditional volume metrics.