No-show appointments are a significant financial and operational burden for the entire healthcare system. In primary care, the rate of no-show appointments ranges from 19%  to 42% . The data set contains just over 988,000 unique encounters spanning 7 years’ worth of appointment, clinical, demographic, and financial data from a large rural Federally Qualified Health Center. Prediction of the probability of a patient missing a scheduled appointment in a primary health care center can be a significant advantage to operational and financial success for a primary care practice. Using the predictive data generated, in combination with targeted interventions, can benefit FQHC practices which typically operate on very small margins. Nine machine learning algorithms were tested against each other to determine the most predictive model generator. Compared to the results achieved using the most commonly used algorithm previously, Logistic Regression, Adaptive Boosting showed a statistically significant improvement in accuracy and recall.
Denney, Joseph; Coyne, Samuel; and Rafiqi, Sohail
"Machine Learning Predictions of No-Show Appointments in a Primary Care Setting,"
SMU Data Science Review: Vol. 2
, Article 2.
Available at: https://scholar.smu.edu/datasciencereview/vol2/iss1/2
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial 4.0 License