EduSource Specialty: Data Analytics

Brett RidouxThe EduSource team can craft all kinds of custom software, but one of our sweet spots is big data cleanup and data analytics.

If you aren’t getting the information you need out of your data, email Brett Ridoux at for a free consultation.

The Problem

A major insurance provider wanted to mitigate their risk by greater use of a government-funded support program for Medicare and Medicaid patients. The state administrator of Medicare and Medicaid did not have a systematic way to identify and help patients who qualify for their support services.

The Goal

Create a risk stratification pool based on demographics, behavior, and past health history to determine the list for referral to additional government assistance. Insurance company will now know who they can refer.

How We Solved it

The EduSource team developed an algorithm to determine risk pools, using weighted criteria based on demographics, past hospitalizations, current medications and ICD10 diagnosis codes. We also built the algorithm to account for duplicate records since it runs once per month and may pull in the same insured multiple times. The database was developed using Hadoop, Apache Spark, SQL queries, and Kylo to automate the process of getting files from insurance providers into the database to determine the risk stratification score.

The Result

There is now systematic criteria based on data analysis to determine qualified patients for government assistance. Anyone over a certain risk stratification score will be eligible for a referral, and duplicates are filtered out of the report.

The EduSource solution has increased referrals from the major insurance company to the government program by approximately 500%.

There are implied cost savings to the major insurance provider and improved outcomes for patients that utilize government services in this program, and those results will become quantifiable by mid-2019.