Researchers at the Stanford Graduate School of Business (GSB) have developed a more accurate method to predict hospital wait times.
The researchers’ new model surpasses standard practices by taking a larger range of factors into account when projecting wait times. It aims to improve predictions in the hospital setting and increase patients’ satisfaction by helping them to be treated on time.
According to the researchers, hospital wait time models tend to be inaccurate — sometimes predicting times up to several hours more than than actual wait — and can leave many patients unhappy.
The researchers identified flaws in current hospital wait time models and then used data from the emergency departments of four different hospitals to evaluate the accuracy of their new prediction method, called “Q-Lasso.”
Associate Professor of Operations, Information and Technology Mohsen Bayati told Insights by Stanford Business that most wait time estimates use models that oversimplify reality in an emergency wait room. They assume equal wait times for all patients, a steady flow of patients in and out of the waiting room and equal times for treatment of each patient. But this does not always capture reality: Some patients require hours-long surgery while others wait for less threatening ailments like an arm sprain. Additionally, the flow of patients in and out of the emergency department fluctuates throughout the day.
The Q-Lasso method combats these issues with a mathematical model that combines statistics with fluid model estimators. When the researchers applied this technique to waits for low-acuity patients at four hospitals they examined, the model reduced the margin of error for hospital wait time predictions by 33 percent.
The Q-Lasso method, however, is not completely accurate; when tested at hospitals, Q-Lasso was still off by around 17 minutes to an hour. Q-Lasso errs toward overestimating wait times in an effort to increase patient satisfaction when people wait for less time than they initially expected.
Contact Kelly Yookyeong Kim at ykkelly ‘at’ gmail.com.