Roosevelt Innovations, LLC is the first technology solution to deliver a simple, seamless, and smart platform for health insurance companies. With industry-leading claims processing capabilities, Roosevelt can transform operations, enabling insurance carriers to focus exclusively on their customers and growing their business.

With advanced machine learning techniques, Roosevelt Innovations performs analysis that identifies potentially anomalous attributes of healthcare provider claims for review. A key focus of this analysis is the identification of providers who may be engaging in fraud, waste, and abuse (FWA) activities. When found, providers potentially engaging in FWA are added to a manual watch list that places their future claims under scrutiny.

Our Provider Anomaly Analytics Toolkit streamlines this process by compiling the data sources and visualization tools necessary for FWA analysis into an interactive web application.

Upon opening the app, users are presented with a table of providers flagged as potentially problematic by machine learning algorithms. From this table, the user selects providers of interest and navigates to either a summary or comparison view for further investigation.

The summary view automatically displays data for a single provider, while the comparison view visualizes user-specified fields for multiple providers. Users leverage this information to confirm or deny potentially anomalous behavior.

By optimizing the identification process of problematic providers, our toolkit effectively reduces carrier vulnerability to FWA while improving the expected quality of care and cost of insurance for members using the Roosevelt platform.

Our toolkit is written in Python with Streamlit as the front-end framework. Data is stored in a Snowflake database and the outlier identification models are developed using scikit-learn and PyTorch.