Use of a machine learning model to predict retention in care in an urban HIV clinic.

Abstract

Identifying barriers to retention in care (RIC) is critical to ending the HIV epidemic in the United States. Therefore, we developed a machine learning model (MLM) to identify predictive factors for RIC in an urban HIV clinic. Our MLM yielded a positive predictive value of 84%, higher than previously reported MLMs. We found that MLM can be used to develop interventional strategies to enhance RIC in HIV care.