Distributed data analytics vendor Databricks Inc. today debuted its third vertical-industry product in as many months with a platform for healthcare and life sciences.

The Databricks Lakehouse for Healthcare and Life Sciences provides datasets and libraries for healthcare-specific applications of data management, analytics and artificial intelligence in areas such as disease prediction, medical image classification and biomarker discovery. The platform is already in use at Providence Health & Services and General Electric Co.’s GE Healthcare, as well as offered by several partners.

The introduction follows last month’s launch of similar offerings for financial services and January’s introduction of a lakehouse for retail. Lakehouse is a Databricks-coined term for a data repository that combines elements of both a structured data warehouse and an unstructured data lake.

“Having a platform where different teams and skillsets can come together with different languages and tools has been the value Databricks has provided,” said Michael Sanky, global industry lead for healthcare and life sciences at Databricks.

The platform comes with a variety of open-source data and AI libraries that address healthcare use cases. One of them is Glow, a toolkit for genomics machine learning and data analytics that’s natively built on Apache Spark. Another compares disease progression against variants at scale.

Databricks calls these toolkits Solution Accelerators and makes them available as notebooks on its website at no charge. “We build the code and release it in a way that makes it easy for a healthcare provider to adopt it,” Sanky said.

Other purpose-built libraries use machine learning to assess patient risk based on encounter history and demographics information and rapidly analyze thousands of slide images to detect signs of metastasis. Yet others map data types to healthcare-specific analytic data models, analyze unstructured medical text using natural language processing and improve biomarker discovery for precision medicine through whole-genome processing.

Sanky said Databricks is actively engaging in the healthcare industry by supporting emerging standards such as the Observational Medical Outcomes Partnership created by the Observational Health Data Sciences and Informatics consortium. “We’ve made sure we’re compatible with those libraries and have released guides to integrating with those datasets,” he said. Databricks is also building mappings between the FHIR interoperability standard and OMOP for translating healthcare messages into patient-level analytics.

Photo: Shutterstock