Patient no-show continues to contribute to the rising healthcare cost, leading to negative impacts on the day-to-day operations of the healthcare system, restricting healthcare delivery efficacy, besides limiting quality healthcare access for all patients. This study addresses the prevalence of patient no-shows, the missed by the patients. Demographic factors particularly age, gender, the time span of appointments and socio-economic status of patients are the most influencing factors on patient medical appointments attendance. Past attendance history, financial information, appointment information are among other factors that are also vital for patient attendance. Five machine learning predictive models namely Logistic Regression, Random Forest, Support Vector Machine, AdaBoost Classifier and Gradient Boosting Classifier were built using the ‘Medical Appointment No-show’ dataset after being treated for all possible types of noises. The Gradient Boosting Classifier was selected as the best performing model with 79.6% accuracy and 0.89 Receiver Operating Characteristics score as Gradient Boosting tends to perform better when it is properly tuned. Future research may include other key factors affecting patient attendance to improve model performance.

Keywords : Appointment Scheduling, Healthcare, Missed Appointments, Non-Attendance, Patient No-Shows, No-Shows Prediction, Predictive Models