Analytics

The pSCANNER network has developed and successfully implemented distributed analytic algorithms that do not require sharing of patient-level data. This technology employs privacy-preserving analytics that allow sites to extract de-identified data sets that can be collaboratively analyzed without transferring patient-level identifiers. The figure below depicts the analytics framework and tools of pSCANNER. 

 

 

 

OCEANS

Observational Cohort Event Analysis And Notification System

 

OCEANS, a Java-based analytics tool, provides a standardized request for multiple analytic workflows. The request protocol can be natively consumed via Java applications or via a web service consuming and returning JSON. The intent is to provide a common framework for performing analyses natively or to leverage additionals engine, such as R. In the latter case, OCEANS provides workflow in Java for managing complex data manipulation and leaves the analysis to the supported external engine. The current implementation provides support for Logistic Regression, Risk-Adjusted Sequential Probability Ratio Test, Propensity Score Analysis, and Survival Analysis. The source code link will be updated upon finalization of the current iteration.