A Model for the Development and Deployment of Machine Learning Solutions in Healthcare

Abstract

The development and deployment of Machine Learning models in healthcare is increasing. However, skepticism remains regarding their reliability, accuracy, and relevance in a clinical setting. As a result, there is limited adoption of these machine learning solutions in healthcare despite the availability of many prototypes from numerous studies over the years. The adoption challenge can be attributed in part to the lack of an established and acceptable process for developing these solutions. In particular, the lack of clinical trials for these solutions remains the most significant challenge. This study, therefore, developed a model integrating research, scientific and ethics review, the machine learning process, Software as a medical device development, software testing, and a clinical trial. Critical considerations for the model's success include meticulous documentation at every step and a multidisciplinary team drawn from Information Technology and healthcare for the entire scope of development. Future work should look into the development of a clinical trial process for Software in general as well as for machine learning solutions in particular.