Data-driven transformation in healthcare: Vida Health's approach using dbt Cloud, Fivetran, Looker, and more from Coalesce 2023

"We're focusing on building member and provider apps, which is our rocket ship, and we want to have a data platform that is deeply integrated into that application experience."

Trenton Huey, senior director of data and frontend engineering at Vida Health, discusses how to leverage a data-driven platform to transform healthcare delivery.

Building a data-driven platform for optimizing healthcare delivery

Vida Health uses a data-driven strategy to optimize healthcare delivery. They employ a five-step playbook that includes integrating into value delivery systems, encoding complex decision triggers, building a production-grade data platform, establishing a bidirectional hub and spoke, and demonstrating data value chain creation.

Trenton emphasizes the importance of integrating into value delivery systems: "We want to take our data platform and integrate it as closely and as deeply into these downstream systems as possible because that's where we've found that more value is created from the data platform that we've built."

He explains the process of encoding complex decision triggers. "We want to actually encode those decisions as much as possible directly into the data platform... we're building to the decisions that were made in using those dashboards,” he says.

The importance of establishing a bidirectional hub-and-spoke model

Vida Health's approach also emphasizes the importance of establishing a bidirectional hub-and-spoke model. This involves getting everyone comfortable with using dbt Cloud and encouraging cross-functional teams to contribute directly to dbt Cloud.

Trenton explains their approach, stating, "What we found is so much valuable analytics and data were being built in the spokes, but they weren't getting pushed back into the central platform to be reused in other places... we've really focused on getting everybody comfortable that's using data to build in dbt."

He also emphasizes the benefits of bringing in domain expertise from different parts of the business. "Not only are they giving extra sets of hands to build in dbt, they're bringing so much domain knowledge from their areas of specialty which is helping us design better data models," he says.

Demonstrating and celebrating data value chain creation

The final step in Vida Health's playbook involves demonstrating and celebrating the data value chain creation. Trenton emphasizes the importance of communication in this process, particularly in helping other business leaders understand the value of data platforms.

He explains: "We've created this association of when something gets built in dbt, something good happens for our members or for our customers or for our business." He adds, "The idea is to list out all the stages because it helps the data teams better understand why this work is valuable."

Trenton also stresses the importance of raising the technical rigor of data platforms, embracing complexity in decision-making, and over-indexing on communication. "The idea is, if you have a strong foundation for your data platform strategy, it should enable future adaptability in a much more streamlined scalable way," he says.

Trenton’s key insights

• Integrating data platforms as deeply as possible into downstream systems creates more value
• It's important to encode complex decisions directly into the data platform
• Building a production-grade data platform requires a technical systems maturity checklist
• Establishing a bidirectional hub-and-spoke model allows for more domain knowledge to be incorporated into the data models