backgroundImage
backgroundImage

Great Expectations Case Study:

How THINKMD improves frontline patient care with GX Cloud

the Great Expectations and THINKMD logos

THINKMD's data team responds to data issues faster and more effectively with GX Cloud


Quick facts about THINKMD

Location: Burlington, VT

Industry: Healthcare technology and data

Company size: 15-30

Data team size: < 5

Background

THINKMD’s CDS platform helps front-line healthcare workers in remote or low-resource areas get access to clinician expertise that complies with World Health Organization guidelines. Ultimately, this helps reduce the impact of the healthcare worker shortage on underserved communities.

Healthcare workers enter a patient’s current health information, symptoms, vital signs and physical exam information into THINKMD’s platform, where clinician-created algorithms are activated. THINKMD’s platform then displays a triaged list of clinical risk assessments and the associated treatment options.

The aggregated data created by these individual health encounters, in turn, is used by THINKMD’s partners for population health monitoring, disease surveillance, and other population-level health research and intervention.

These dual uses mean that the quality of THINKMD’s data can have major health impacts and the individual and population levels. With these stakes, having high-quality data is critical.

“We need to keep the data very clean so that anyone who uses the data will make good decisions.” - Kathy Coughlin, THINKMD software project manager

Challenges

THINKMD faced many of the data quality challenges typical in modern organizations, including:

  • Complex pipelines: THINKMD’s data moves through a multi-stage pipeline where it undergoes several transformation and cleaning operations.

  • Duplicate data: Much of the data coming in through THINKMD’s platform is from locales where data connections can be unreliable, meaning that duplicate records are possible.

With a relatively small data team, THINKMD needed a data quality solution that would allow them to deploy quickly and effectively.

Solution

With GX Cloud, THINKMD implemented a data quality process that rapidly deployed tests which provide fast and information-rich alerts when data quality issues occur.

Fast

GX Cloud’s iteration-focused approach for sustainable growth allowed THINKMD’s data team to maximize their limited resources. 

An initial set of Expectations allowed the team to implement the most critical testing immediately. Later, the team expanded their testing to encompass additional data.

User-friendly

THINKMD’s data quality efforts were initially headed up by Ollin Langle-Chimal, a data scientist. After Ollin decided to return to academic research, the data quality effort transitioned to Kathy Coughlin, a software project manager.

Kathy has considerable interest in clinical data, software engineering, and data quality, but limited experience as a data practitioner herself. But even with little experience in data quality work specifically, she found the transition and introduction to GX Cloud straightforward.

Ultimately, THINKMD experienced virtually no disruption in their data quality process even as primary responsibility shifted from a data-oriented role to a more business-oriented role.

Impact

GX Cloud’s Slack alerts mean that Kathy and the THINKMD team are notified right away when a data quality issue occurs. A single click takes them to GX Cloud with more information about exactly what failed, how it failed, and what the expected state was. 

With GX Cloud’s detailed context for issues, the data team can start their root cause analysis and remediation work immediately—no need to spend their limited resources identifying the error first.

Ready to see how GX Cloud can give you and your team confidence in your high-stakes data? Request a demo.

11 countries
Caregivers and patients in 11 countries depend on the quality of THINKMD’s data.
15 assets
THINKMD simplifies asset management with multiple customers per asset.
1m+ records
Millions of records per dataset make automated quality testing critical.
Components
  • Amazon S3 (Athena)
  • Snowflake
  • Google Cloud Storage
  • Tableau
  • Keen (keen.io)
"It’s easy to use and easy to determine which record failed. [...] Getting from ‘something went wrong’ to ‘what went wrong’ is fast, which allows us to dig into the data to understand the root cause."
-Kathy Coughlin, software project manager

Make your data work for you

Maintaining data quality doesn’t have to be hard. Get started with Great Expectations today and see value in minutes.

Pricing bottom hero
©2024 Great Expectations. All Rights Reserved.