Apply by 3/31 for career development funding to advance Learning Health System science. Learn more.

Happy or Not

Happy or Not

Capturing patient and provider sentiment in real time

Project status

Pilot/study with results

Collaborators

PPMC ED staff 

Innovation leads

Opportunity 

Health systems have historically relied on traditional surveying tools to understand patient satisfaction. Unfortunately, these surveys can be long, tedious, and burdensome to submit. Additionally, they are often distributed and completed long after an experience occurs, compromising the quality of feedback. 

Similarly, traditional approaches to measuring satisfaction among clinical providers are reactive and sporadic. 

Having a real-time pulse on the experience of patients and providers is essential to driving quality improvement in health care and supporting the long-term well-being of the health care workforce. 

Intervention  

In partnership with Penn Presbyterian Medical Center (PPMC) experts, we set out to test the feasibility of measuring patient and provider experience in the emergency department (ED) in real time. 

We used Happy or Not feedback terminals to collect the sentiments of doctors, nurses, and patients. Terminals were strategically placed in three locations to optimize participation from each group. Each stand presented users with four sentiment options: very positive, positive, negative, or very negative. Sentiment data was collected longitudinally and compared to ED metrics like arrivals per hour, length of stay, waiting patients, and boarding patients. 

Impact  

The five-month pilot we conducted proved that collecting large amounts of sentiment data in real time in the ED without placing an additional burden on patients and providers is feasible. Doctors, patients, and nurses all used the terminal at high rates. Nearly 14,000 sentiments were recorded, with roughly 68 percent coming from provider-facing terminals. 

Timestamps on the submissions enabled our team to identify trends and patterns over time and offered insight into specific drivers of dissatisfaction. For example, we saw that negative sentiments submitted by doctors and nurses were moderately associated with an increased number of patients waiting to be seen. There was also a strong correlation between negative sentiments and an increased number of patients being boarded in the ED. 

These low-effort methods are being translated into more personalized ways of connecting with clinicians in real time. Our future pilots will explore the use of text messaging and wearable devices to provide a more personalized approach to understanding the complex situations that arise in clinical settings.