hc1 Uses AWS to Turn Billions of Quest Lab Tests into Healthcare Insights

Reducing low-value testing in the healthcare system requires developing a high-performance, scalable analytics platform for the data of billions of lab test results to derive insights.

hc1 designed its cloud-based hc1 High-Value Care Platform to lower costs and improve patient care by reducing low-value practices in healthcare. The company partnered with Quest Diagnostics, a diagnostic information provider with one of the largest lab datasets in the world, to tackle the issue of low-value care—things such as unnecessary services, inefficient delivery, and missed prevention—which accounts for 30 percent of healthcare dollars spent and amounts to $765 billion. To reduce low-value testing, hc1 had to quickly process data from Quest’s 48 billion lab events at scale and generate real-time insights in the form of dashboards and visualizations for healthcare providers. “Whether it’s a lab, a hospital, a health system, or a health plan, or even looking at this nationally, we have the capability to organize and normalize billions of lab results to identify where low-value testing is occurring,” says Zach Berg, senior product director for hc1.

A third of Americans have had a test performed by Quest, which generates billions of test results. hc1 uses several HIPAA-compliant Amazon Web Services (AWS) tools to power its platform to store and process the data from Quest. To ingest the billions of records, hc1 uses big data platform Amazon EMR to streamline the data pipeline before loading the data into Amazon Aurora, a cloud relational database. Amazon Elastic Container Service (Amazon ECS) scales capacity so that as the test volume fluctuates, the platform can efficiently tackle the application of medical lab guidelines or rules to identify low-value testing on every transaction. So far, the platform has processed more than 10 billion results, including 290 million lab orders organized and associated with 153 million patient profiles, with some tests being flagged more than 30 percent of the time.