Sumitovant more than doubles its research output in its quest to save lives

Here at Sumitovant, we’re developing next-generation therapies for treating cancer, life-threatening immune disorders in infants, women’s health issues, and other medical conditions. Thankfully, increasing access to anonymized medical data, pharmaceutical studies, clinical trial results, and other third-party data sources gives our medical teams and physician customers more information than ever before about treatments’ efficacy and adoption. However, as we’ve added data sources, we’ve learned that more data doesn’t automatically translate into better insights.

The need for raw data access and regulatory compliance 

Because Sumitovant is a biopharmaceutical company, the scientific process directs everything we do. The proof or disproof of every hypothesis involves a sequence of ever-more-specific questions answered with available data. Because each question depends on the previous inquiry’s answer, it’s not possible to compile a complete list up-front. This meant that my team had to run a custom query to answer every question every researcher had as they refined their hypotheses.      

For years, our back-and-forth query process worked, but as we added data sources, the process became too time-consuming. We had to write increasingly complex queries, and researchers waited several days to get answers to their questions. Realizing our ad hoc query process was no longer sustainable or scalable, we looked for a solution that enabled researchers to answer their own questions using raw data while meeting our global data security and privacy regulations. 

We evaluated solutions and chose Looker because it:

Reduces the time to develop and scale flexible queries on massive data sets. In one day of development with Looker, we can accomplish what used to take two weeks of manual back-and-forth analysis. 

Makes it easier to hook into and integrate more data sources, such as researchers’ Redshift and Athena cloud databases and our partners’ multicloud enterprise data lakes.

Supports our technology agnostic strategies by providing APIs that enable deep integration with our Infrastructure as Code data platforms.

Simplifies the creation of governed dashboards that give people permissioned access to data and the freedom to explore it on their own.

More than twice the research output

For our initial Looker use case, I built dashboards that hook into raw third-party electronic health record data for a health economics and outcomes research (HEOR) team. The dashboards meet the diverse needs of the team’s researchers, pharmacists, medical professionals, and data scientists, and they give them answers in seconds rather than days. Enabling each team member to drill down into data themself is also vital for greater exploration into treatment outcomes and generating data for publications that physicians rely on for treatment decisions.

Answers from more sources including unstructured text

The HEOR team’s early success with Looker is driving broader organic adoption across Sumitovant and its subsidiaries. That’s because in addition to saving time, Looker scales. I can easily build new dashboards that hook into raw data from Sumitovant, our subsidiaries, and our partners—and provide filters so that people can query the data for their specific disease research purposes. I recently created a natural language processing framework with Looker that uses topic modeling to analyze field notes from physician conversations about prostate cancer and women’s health treatments. This Looker dashboard allows our field medical team to quantify and identify the most pressing topics by understanding what physicians are saying across thousands of unstructured text comments and quotes.

Looker dashboards are a critical innovation for our company and our partners. Researchers can answer questions as they arise, work seamlessly with fewer delays, and accelerate the generation of data that providers rely on to improve patient outcomes and save lives.