Artificial intelligence (AI) technology is now being utilized to predict patient attendance behavior for hospital medical appointments. Researchers from the Banjarmasin State Polytechnic, Fitria and Nitami Lestari Putri, developed a model based on the Decision Tree algorithm to analyze medical appointment data. Published in 2026, this study aims to assist hospitals in optimizing scheduling systems and reducing operational inefficiencies caused by unplanned patient no-shows.
Patient no-shows without prior notification represent a major challenge in hospital management information systems. This phenomenon not only leads to wasted physician time but also impacts operational losses and decreases the effectiveness of healthcare services. Consequently, the ability to predict patient attendance patterns is crucial for hospitals to manage medical resources more effectively.
In this study, the team used a dataset of 200 hospital appointment records, which included doctor information, reasons for visits, days, and appointment time sessions. The development process involved several stages, ranging from data cleaning, feature engineering to create new variables, to encoding categorical data for machine learning compatibility. The Decision Tree model was selected due to its ability to classify categorical data and provide results that are easily interpretable by hospital management.
The analysis revealed several significant findings regarding patient no-show patterns:
- The highest number of no-shows occurred on Wednesdays.
- Morning and afternoon examination sessions experienced higher no-show rates compared to evening sessions.
- There were variations in no-show rates depending on the specific doctor assigned.
- The Decision Tree model achieved an accuracy of 60% in predicting patient attendance status.
Implementing this prediction model offers tangible benefits for the healthcare sector, particularly in supporting operational decision-making. By detecting potential no-shows earlier, hospitals can adjust schedules or perform calculated overbooking to minimize idle physician time. Although the current accuracy stands at 60%, the model is considered to have great potential for further development by incorporating additional variables such as patient demographics, medical history, and payment methods to enhance predictive performance in the future.
Author Profiles: This research was conducted by Fitria and Nitami Lestari Putri from Politeknik Negeri Banjarmasin (Banjarmasin State Polytechnic). Fitria specializes in Accounting Information Systems, while Nitami Lestari Putri possesses expertise in Smart City Information Systems.
Research Source:
Fitria, F., & Putri, N. L. (2026). Machine Learning-Based Patient No-Show Prediction for Hospital Appointment Systems. International Journal of Education and Life Sciences (IJELS), 4(6), 761-772. DOI:
0 Komentar