From my experience working with healthcare providers and payers, there’s one thing that stands out to me: the most important insights for compliance are almost never in the data fields you expect. Instead, they’re buried somewhere in the mounds of clinical data amassing in healthcare every day. As a practicing radiologist, I see firsthand just how much of the most meaningful information never makes it into traditional reporting. In my experience working with payers and providers, the difference between missing reporting targets and exceeding them usually comes down to how well you can access what’s hidden in the rising mountains of clinical narrative. I’ve also found that with the help of natural language processing (NLP), it’s entirely possible to find every needle in the haystack and consistently exceed reporting requirements.
The Unstructured Data Gap is Too Wide to Cross Alone
More than 80% of healthcare data is unstructured, meaning it exists as physicians notes, referrals, discharge summaries, radiology images, and other documents in a patient’s electronic health record that defy organization by pre-defined structure. That leaves just 20% of health data able to fit into neatly structured fields. Most of the time, the details that matter for HEDIS and accreditation are locked away within this narrative text. That can make it tempting to focus on easy-to-use structured data. But relying on structured data alone is like reading every fifth page of a book. You miss the context, the crucial facts, and sometimes the entire plot. I’ve watched health plans miss out on better HEDIS scores, risk their accreditation, and lose millions of dollars in revenue due to decreased Star Ratings, all because their systems couldn’t see, and report on, the whole story.
Medical NLP is Like a Bridge to Access Unstructured Data
Medical NLP has the potential to change everything. That’s because it can extract the exact needed details from thousands of pages of notes without the time-consuming manual labor – which is not just slow, it’s financially crushing. Medical NLP transforms HEDIS from a labor-intensive, retrospective fire-drill costing millions in temporary staffing to a continuous, automated process that minimizes annual expenditures.
That’s not just a theory or a dream. It’s already happening. A recent peer-reviewed study of clincial NLP found that 81% were using NLP for data collection, and that in multiple cases, it achieved 90% of higher accuracy in information extraction and classification. That accuracy stems from NLP being much more than a keyword search that misses out on context and relationships. Modern NLP understands language, chronology, negation, and nuance, making it possible to find the evidence needed for HEDIS measures, accurately codify events, and create a trail that auditors can follow. That means more complete quality reporting, fewer missed care gaps, and data that holds up under scrutiny. That’s more than just excellent compliance. It’s trust in the making.
How to Use NLP to Build a HEDIS-Ready Data Pipeline
Having an excellent NLP engine isn’t enough on its own. The organizations that have the most success with HEDIS compliance are the ones that think strategically about their entire data pipeline from start to finish. It’s not just about analytics; it’s about being audit-ready all the time. That means developing architecture that can bring together and make meaning of both structured and unstructured data, building in transparency from the beginning, and making sure that every insight can be traced back to the source. The goal is to create a foundation where technology simultaneously support the people doing the work while reducing manual effort and helps to improve outcomes without introducing or perpetuating information black boxes.
The Positive Impact of NLP on Business and Compliance
When— not if— you get this right, the results are real and immediate: better HEDIS scores, stronger Star Ratings, and the ability to maintain accreditation and financial performance without it always feeling like a struggle. Investing in explainable, auditable data pipelines will prepare you for not just today’s requirements, but for every change that comes.
Unlocking the full value of unstructured clinical data isn’t just a technical problem, it’s a strategic one. Taking the extraction and analysis of unstructured data seriously and using NLP to your advantage in pulling out buried insights will position your organization to excel in compliance, audit, and care quality without increasing your work burdens.
About Dr. Tim O’Connell
Dr. Tim O’Connell is the founder and CEO of emtelligent, a clinical-grade AI solution. He is also a practicing radiologist and the vice-chair of clinical informatics at the University of British Columbia.
