Abdominal Pain Triage (APT) Score: Machine-Learning Identification of ED Patients Requiring Urgent Surgery

Departments Included in Project: Medicine, Surgery,

Project Classification: Timeliness, Efficiency

Does the work incorporate equity?

Current ED triage assessment of the acuity of a patient's condition is largely subjective - it relies on patients to quickly portray the severity of their clinical symptoms to nursing staff, and for nursing staff to interpret that information in context with only the help of a single set of vital signs. As one might guess, a significant amount of bias can be introduced into this system. In communities with poor access to healthcare and low health literacy, a patient may not feel comfortable in the triage setting nor have health care experience to portray the issue at hand, leaving much for interpretation by the triage staff. Moreover, patients in socially vulnerable neighborhoods more often have a number of co-morbidities (e.g. diabetes) which can cause abnormal presenting symptoms, complicating triage precision. These things in combination can result in inequitable ED wait times for patients requiring time-sensitive emergency general surgery. Our APT Score attempts to avert bias and create equitable care by using a diverse set of our own patients to create an algorithm that removes the subjectivity involved in current triage practices to ensure expeditious care for all.

Does this work address any of the UChicago Medicine quality priorities as listed on the Clinical Excellence Scorecard?

Mortality Reduction