If we could predict a patient’s future risk of developing illnesses such as depression or lung cancer in the next three years, then we could potentially intervene and improve the patient’s future health. The PatientLevelPrediction R package provides a standardized analytic framework for developing diagnostic and prognostic models using observational healthcare data (e.g., electronic healthcare data and insurance claims data). It utilizes the OMOP common data model, a standard data structure, to enable rapid but reliable model development. The package contains a library of binary classifiers and survival models (with R, Python, C++ and Java backends) for users to select but also enables the flexibility of writing custom supervised learning methods. In addition, the package contains a suite of recommended performance metrics and visualizations. In this talk we will demonstrate how to use the package to develop and internally validate data-driven models and then show how the standardized approach makes large-scale external validation possible. We will also illustrate the built-in shiny app that enables interactive visualization of multiple models.