Q&A with Perspecta’s CTO: Behind the Scenes of Provider Data Clean-Up

Provider data is messy. And if you’ve ever wondered what it takes to turn that chaos into clean, accurate, actionable information — we’ve got a peek behind the curtain for you!  We sat down with Bo Roff-Marsh, Chief Technology Officer at Perspecta, to get an insider’s look at what really happens when provider data lands in our hands, the headaches it causes, and why good data hygiene isn’t just a nice-to-have, it’s critical.

When provider data first lands in your team's hands, what is the typical state it’s in?
Bo: It’s all over the place. We get data through everything from SFTP and APIs to clients’ proprietary systems, and no two are alike. The biggest issue is typically the format: same type of data, completely different layouts. Beyond that, we deal with incomplete files, mislabeled columns, truncated values… a mix of human and machine errors. Every new file is like a mystery box.

 Have you seen any “horror stories” or examples of data so ugly it made you stop and shake your head?
Bo: Plenty! The surface-level stuff is annoying — bad formats, broken rows, missing fields. But the real horror is when bad data directly impacts people. Imagine an injured worker using a directory to find a provider, only to discover the phone number is wrong, or the address leads them to an empty lot. That’s not just inconvenient, it’s unacceptable. Data inaccuracy at that level discredits the entire system and creates real-world problems for the people depending on it.
 
What are the most critical attributes often missing or incorrect in provider data sets? Why do those gaps matter?
Bo: Operational hours, email addresses, website links, provider specialties, languages spoken… these are the details that make a provider listing truly usable. Without them, a directory is half-baked. People shouldn’t have to jump through hoops to get the info they need when they’re already dealing with a healthcare issue. Comprehensive, accurate data saves time, reduces frustration, and improves outcomes.

How do duplicates sneak into provider data, and what kind of problems can they cause?
Bo: Duplicates usually start at the point of entry, such as typos, lazy shortcuts, or inconsistent processes. Sometimes it’s a human copying data for convenience. The trouble is that duplicate records can misrepresent provider information. A doctor might be listed twice under slightly different details, causing confusion, misrouted calls, or incorrect mailings. Once bad data spreads to external systems, it’s nearly impossible to retract, and correcting it later can take weeks or months.

What is the triage process your team follows when new data arrives? 
Bo: First, we check for missing required fields. Then we validate the file format against what's expected. From there, we standardize data, like mapping various speciality descriptions to our internal standards. We normalize attributes too, ensuring things like phone numbers all follow a single format. It's about bringing order to chaos, quickly. 

What tools and technology do you use to handle this at scale? 
Bo: We rely on MS SQL, T-SQL, Azure Data Lake, and SSIS packages for ETL processes. These tools help us cleanse, normalize, and validate provider data efficiently and accurately. 

What's the difference between making data "usable" and making it "accurate"? How far does your team go to bridge that gap? 
Bo: Usable data is clean enough to work with. But that doesn't mean it's reliable. Accurate data give you confidence in decisions made from it. We make data usable by standardizing and formatting it, then take it further by cleansing, augmenting, and validating it against trusted sources. Our value is not just delivering what clients ask for but improving it so it's truly actionable. 

Finally, what would you say to payers or networks who don't fully grasp the importance of provider data quality? 
Bo: I'll give you two analogies. First... it's like personal finance. When you have a little, it feels manageable. As it grows, it gets harder to track. Ignore it too long, and things get out of control. Data quality is the same — the longer you neglect it, the harder it becomes to fix.

Second... think of it like cancer. Early detection is key. If you address data quality issues while your datasets are small and young, it's fast and cheap to fix. Wait too long, and problems spread, costs skyrocket, and your system never fully recovers. Prevention is always better, and cheaper, than remediation. 

Want to know more about how Perspecta transforms messy provider data into clean, accurate, reliable directories? Contact us today to start the conversation. 

Learn more about Perspecta's Data Cleansing Solutions

DOWNLOAD