The Spotlight Analytics Platform
Any Data. All Data. New Insights.
Taking any and all data, Spotlight uses proven machine learning to optimize risk insights and infuse them into any clinical workflow and any technology platform.
Population Risk Management
Acute Episode Risk Management
Quality and Performance Measures
What Makes Us Different
Many healthcare analytics companies offer advanced algorithms that analyze healthcare data. But the data science at the heart of Spotlight was developed from the ground up specifically for healthcare applications. Our models are the most comprehensive on the market, and offer predictions that are highly accurate and actionable.
Any and all data
While many of our competitors rely only on medical claims data that’s often 30 to 90 days old, we start with a comprehensive, up-to-date data set. We use real-time clinical, billing and claims data, as well as proteomic and metobolomic data to derive insights from cellular processes and metabolic pathways.
Machine learning as a service
Unlike other vendors that provide a single model tested on a specific data set, HBI takes the data you have—any and all data—to calibrate and optimize our proven models to the data you make available. Then we recalibrate each year or whenever your data set changes to assure you’ve always got the best results for your populations.
While others may provide only a simple risk score or stratification, we assign an individual probability and expose risk features and weights to help clinicians understand what’s driving that risk. We also connect modifiable risk features to intervention recommendations and clinical profiles. This gives users confidence as they use our predictions to make better treatment decisions.
More risk models
Many competitors require multiple static algorithms (one for congestive heart failure patients, one for Medicare populations, etc.). But HBI’s algorithms work with all available data to deliver risk results for all ages, all diseases and all payer types. We offer over 30 dynamic condition, cost and event-based prediction models across acute, population and community care settings.
We’ve spent years proving our models using actual client clinical data, with practicing health providers, in real patient care situations. We’ve validated our models through peer-reviewed research and continuously improve them with feature engineering and machine learning processes.