A Machine Learning Approach to Predicting Users’ Intention to Adopt Telemedicine in Healthcare

Although there has been a noticeable ascent in the deployment of telemedicine services, yet their uptake and utilization in healthcare settings have been sluggish, mostly in developing countries. Leveraging machine learning (ML) to precisely predict the adopting factors could significantly contribute to effectively utilizing telemedicine in healthcare. Thus, the study aimed to have a more comprehensive understanding of the factors influencing behavioral intention towards adopting telemedicine applications in healthcare by integrating the well-established Unified Theory of Acceptance and Use of Technology (UTAUT), Task Technology Fit (TTF) models with an additional factor, e-health literacy. Furthermore, the study leveraged the power of ML algorithms to construct a robust model that predicts behavioral intention toward telemedicine adoption. To achieve these objectives, the study collects data from 373 Bangladeshi telemedicine users, and seven widespread supervised machine learning algorithms, including BayesNet classifier, Naive Bayes classifier, Logistic-Regression classifier, Multilayer Perceptron classifier, Classification via Regression, Rule-learner (PART), and decision-tree classifier (J48) were employed to construct the predictive model. A confusion matrix has been calculated for each classification model. Performance metrics, including accuracy, precision, recall, and F1-score, have been computed. The findings indicated that with an accuracy of 74.3% the BayesNet classifier outperformed with other classifiers in predicting the users' intent to adopt telemedicine. The results of this study have the potential to provide a framework for government agencies, and others engaged in telemedicine service implementation and deployment, thereby increasing the uptake and utilization of telemedicine in healthcare endeavors across the globe.