Creating seamless healthcare experiences from start to finish

Accolade, a healthcare company that provides care delivery and advocacy services, is on a mission to help every member lead their healthiest life. The company’s technology-enabled solutions combine virtual primary care and mental health support, expert medical opinions, and care navigation. Accolade’s data group sought to connect the fragmented world of healthcare data in the name of better health. Spread across multiple platforms, Accolade’s data was siloed and lacked real-time access, hindering accurate member stratification (i.e., the process of categorizing members into different groups based on health status, risk levels or healthcare needs) and timely care delivery.
 
Managing multiple data systems led to inefficiencies and increased operational costs for Accolade. According to Accolade’s Vice President of Enterprise Data and Clinical Platform, Kapil Ashar, “Fragmented data across multiple different platforms hampered our efficiency and ability to provide timely, personalized care. Our internal teams struggled to quickly access the necessary information, leading to slower response times and a diminished ability to assist our members effectively.” Accolade recognized the vital need to fix their data, remedy their challenges and enhance their service delivery, which could now be done with the help of generative AI. However, the complexity of managing and coordinating multiple models across various systems, such as AWS Redshift and Snowflake, proved to be a significant obstacle.
 
Since the company’s complex data management landscape also complicated the development of AI-driven initiatives, numerous teams wanted to collaborate more efficiently and innovate at the pace needed in the competitive healthtech market. Therefore, Accolade’s first step was to transition from a traditional BI infrastructure to an AI-driven framework.

Accolade leveraged several Databricks Data Intelligence Platform components to solve their data fragmentation, ML model development and customer service challenges. Central to this implementation was a lakehouse architecture that enabled the combination of data storage and management to facilitate easier access and analysis. This unified environment allowed Accolade to break down the silos between different types of data and systems, making it possible to harness data for their desired artificial intelligence use cases. Ashar elaborated upon the transition: “Databricks served as the foundational step for us. It’s not just about having data; it’s about making it actionable and efficient, which Databricks enabled us to do effectively at a scale we previously couldn’t manage.”
 
At the core of these real-time analytics, the Databricks Platform tapped into Apache Spark™ for streaming data capabilities, enabling continuous data ingestion from various sources. As streaming data flowed into the system, Databricks Unity Catalog was critical for management and governance, supporting HIPAA compliance requirements with stringent access controls and detailed data lineage. Now, Accolade could seamlessly govern their structured and unstructured data, machine learning models, notebooks, dashboards, and files across any cloud or platform. This unified approach to governance accelerated data and AI initiatives, simplified regulatory compliance and allowed data teams to access data and collaborate securely.

Databricks Mosaic AI tools also paved the way for Accolade’s use of retrieval augmented generation (RAG), a generative AI workflow that uses custom data and documents to provide context for LLMs. Accolade used the Mosaic AI Agent Framework to develop a RAG solution specifically tailored to improve the efficiency and effectiveness of their internal teams. The Databricks Platform allowed the company’s power users to access diverse data sources — from PDF files to online protocols and customer information — providing them with the guidance to answer questions confidently and correctly.
 
Accolade’s RAG solution relied on DBRX, an open source LLM developed by Databricks, to help them use proprietary enterprise data for more accurate and context-aware outputs. For example, Accolade leveraged DBRX to enhance internal search functions and enable internal users to retrieve accurate information quickly. With the help of GenAI to analyze customer contracts and other relevant data, the internal teams could better understand and fulfill customer commitments. Once the DBRX model generated insights, Databricks Model Serving came into play. It deployed the model as a RESTful API, enabling real-time predictions that could be integrated directly into Accolade’s decision systems. With Model Serving, the healthcare technology company had Databricks’ support for version control, automatic scaling and integration for comprehensive lifecycle management. As a result of the Databricks integration, Accolade experienced greater transparency and control over resources and costs to simplify the development and maintenance of ML models, while ensuring compliance with data governance standards.