The client had data including 2,500 variables cataloged on over half a million patients. That’s an incredible amount of data that holds virtually no value unless it can be seamlessly analyzed and reported upon. Thankfully, the client had a robust understanding of what data they had available, what they needed the new data reporting system to do, and how they hoped their users would interact with it.
phData sat down with the client to create a plan of action for developing the new solution. This phase included pulling sample tables of data, identifying the data types, what the data represented, and how filters would be needed to be combined or separated to effectively analyze and report on the data.
Next, the teams determined the best way to combine and cross-reference data between the multitudes of tables. Finally, the team created and carefully curated a workbook to make searching as robust and easy to understand as possible. As you can imagine, with this volume of data and its intricacies, the data requirements for this project were complex.
In addition to improving on an existing flexible data modeling functionality, the team was able to offer a new functionality called longitudinal filtering. Longitudinal filtering allows the user to search for an attribute of a client at any point in that client’s medical history. For example, this means that a researcher can pull a report that showcases all patients who have ever had a hemoglobin lab result that fell within a certain set of values at any point in their healthcare journey. Previously, the researcher in this scenario would only have been able to pull from the most current (or present) hemoglobin values of the patient database.
This feature is incredibly beneficial in the world of researchers and investigators looking to follow the journey of cancer patients over time to determine and develop better, more successful treatment plans.
For the client, the success of their new data reporting system hinged on two critical aspects: user-friendliness and accurate reporting. To ensure these objectives were met, they recognized the need for a seasoned team of professionals – a team like phData.
With an established reputation as experts in Snowflake and Sigma, phData was the ideal choice. They not only possessed the technical expertise but also housed a dedicated team of data engineers specializing in crafting tailored data analysis and reporting solutions for various organizations.
With the new data reporting system, the client can now build more robust reports – including the added functionality of applying a longitudinal filter. Additionally, the client’s previous reports took upwards of 20 minutes to pull, whereas the new system can now produce the same – or more robust – reports in less than one minute.
Between the start of this project and the present, Sigma has added new functionality that has made complex data projects like this one even simpler. The newly-released Sigma feature is called “in workbook join,” and it essentially enables more flexible data modeling features.
Learn more about this solution here: This But Not That Filtering in Sigma Computing.
Lastly, phData recently developed a proof of concept similar to this one for another healthcare client in less than 48 hours!
Are you feeling inspired to start making data-driven decisions for your business? We can help!
Data Coach is our premium analytics training program with one-on-one coaching from renowned experts.