Abstract

This paper providesevidence that in an emergency department (ED), the arrival of an additionallow-acuity patient substantially increases the wait time to start of treatmentfor high-acuity patients. This contradicts a long-standing prior conclusion inthe medical literature that this effect is "negligible". The priormethodology underestimates the effect by neglecting how delays are propagatedin queueing systems. In contrast, this paper develops and validates a newestimation method based on queueing theory, machine learning, and causalinference. Wait time information displayed to low-acuity patients provide aquasi-randomized instrumental variable, and is used to correct for omittedvariable bias. Through a combination of empirical and queueing theoreticanalyses, this paper identifies three mechanisms by which a low-acuity patientincreases the wait time for high-acuity patients: pre-triage delay, transitiondelay when an ED interrupts treatment of a low-acuity patient in order to treata high-acuity patient, and the delay in waiting for test results forhigh-acuity patients. Thus the paper identifies ways to reduce high-acuitypatients' wait time, including: reducing the standard deviation or mean of thetransition delay, preventing transition delays by providing vertical or ``fasttrack" treatment to more low-acuity patients; and designing wait timeinformation systems to divert (especially when the ED is highly congested)low-acuity patients that do not need ED treatment.

Time

12月31日（星期四）9:30-11:00

Speaker

Danqi Luo is a Ph.D. candidate at StanfordUniversity Graduate School of business. Her research applies causal inferenceand machine learning guided by queueing theory and behavioral operationsresearch insights to operations problems. Her primary research application isin health care, but she is also interested in other service settings, includingon-line market places. 

Venue

Zoom会议室ID:99139784980

密码：556789