Predicting Medical Appointment Attendance

Abstract

The purpose of this study is to reduce the percentage of patients who do not attend their scheduled doctor’s appointments. The first phase of this project involves determining which characteristics of an appointment are highly correlated with no-shows. The second phase is to develop a model that can predict whether a patient is going to attend their scheduled doctor’s appointment with at least a 75% accuracy. This model will be created using a random forest algorithm on a data set containing information on 300,000 anonymized medical appointments. In this data set, approximately 30% of the scheduled appointments are no-shows. The overall goal is to create a model that can determine in advance which patients are likely to not show for their appointment, thereby providing a better allocation of doctor resources.