“No-show” is a common term in healthcare that is used to describe patients who fail to attend their scheduled appointment. From a system perspective, these ‘no-shows’ are of particular interest as they incur financial losses to hospitals and clinics. Schedulers are therefore encouraged to overbook patients in order to account for potentially lost revenue. While this solves a problem for the hospital, it does not address the issue for the patient: that they are not receiving the care that they need in a timely manner.

An average of 30% of patients of the elective surgery outpatient clinic at ZSFG miss their scheduled appointment. We are using the Super Learner machine to generate predictive models that can anticipate which patients at the ZSFG Surgical Outpatient clinic have a higher likelihood of missing their scheduled appointment. In tandem, we are also using a Human-Centered Design (HCD) approach to conduct semi-structured interviews in order to gather insights regarding patients’ unmet needs accessing their scheduled appointment. The goal of this combined quantitative and qualitative approach is to create a rich data set that can inform interventions that meet both system and user needs.

In this session, we will begin by discussing the benefits and challenges of using predictive analytics and HCD methods to meet both system and user expectations to improve the no-show rate and overall efficiency and experience in clinic.

Slides: https://ucsf.box.com/s/0rhd4027qnu77jjfwm8rarv3ersxaohl (MyAccess login required)