AI
Combining Multiple NLP, Machine Learning (ML), and AI techniques to Achieve the Best Accuracy
Astrata’s AI platform is based on a mature but highly innovative technology that we developed over almost two decades, funded by the National Institutes of Health. It combines multiple NLP, Machine Learning (ML), and AI techniques to achieve the best accuracy in the industry for healthcare quality measurement.
NLP THAT’S OPTIMIZED FOR HEDIS®
Astrata’s proprietary technology starts with the HEDIS® value sets but greatly expands them to include terms and phrases that do not exist in other terminologies. For example, a colonoscopy may be called “c-scope”, or a FIT-DNA might be referred to by a brand name like “Cologuard”. Misspellings such as “Colonoscopy” occur frequently. Astrata’s NLP can pick up all these synonyms and misspellings more accurately with the full range of synonyms and common misspellings in its dictionary.
Astrata’s NLP can also understand temporal statements and reason on them. For example, mentions of procedures like colonoscopy that are referenced by temporal expressions such as “three years ago” are acceptable as evidence of gap closure as long as they are within the required range of ten years. But lab tests such as FIT DNA require different reasoning because the entire date (day, month, and year) must be explicitly mentioned. Astrata’s NLP Insights product provides the power to identify when the event occurred and to evaluate it against the HEDIS® specification to know whether the gap can be closed with current documentation.
Key Benefits
Optimized for Clinical Quality
Sophisticated, beyond-industry-standard NLP capabilities deliver superior value for quality measurement.
Bespoke NLP that Speaks Your Language
It’s a big country with lots of regional and local variations in data-entry requirements for EMRs and HIEs. That’s why clinical NLP systems’ accuracy typically falls when implemented in a new location. Astrata’s NLP Insights is designed for efficient, speedy localization. As part of our engagement with you, we train our NLP on your data, tailoring for variations and quickly adjusting to learn your EMR’s language.
Modern and Scalable
Designed from the ground up to be deployed in a modern scalable cloud infrastructure, NLP Insights can process hundreds of millions of documents in short order. Yes, you can calculate HEDIS® measures on your entire population across all business lines.
Constantly Learning & Improving
We don’t stop when the software is deployed. We monitor our NLP accuracy by analyzing cases where your abstractors disagree, and we use these disagreements to inform improvements
UNDERSTANDS THE CONTEXT TO ACCURATELY CODE HITS, EXCLUSIONS, LEADS AND GAPS
Similar to expert chart reviewers, Astrata’s NLP makes complex judgments to determine whether a specific case has some or all of the evidence to close the gap, or meets the criteria for exclusion. Your HEDIS® NLP system should be able to do the same. For example, not all colonoscopy reports provide evidence of a completed colonoscopy. But based on the current HEDIS® specification, only completed colonoscopies can be used to close the COL Gap. Astrata’s NLP Insights understands the context and nuances of HEDIS® to deliver highly accurate classifications across your populations.
TUNED FOR YOU
Astrata NLP is highly portable and works equally well across various types of provider EMRs and locations. We have a built-in process for “tuning” the NLP to your environment to boost accuracy in your setting. At Astrata, our customer agreements build in this “tuning” phase as part of our engagement with you. Our software makes it easy to adjust to the nuances of your providers.
ALL ABOUT ACCURACY
Astrata’s NLP was developed to exceed rigorous academic standards for measuring accuracy. We measure NLP performance at all stages of development from measure development to tuning and performance in the wild. We provide the reports and tools to give you line of sight into our accuracy and empower you to meet your auditing requirements.
IN SYNC WITH NCQA
Recognized as leader in the market for this technology, Astrata was one of only three companies invited to participate in NCQA’s NLP Working Group. Astrata is the only recipient of the NCQA DAV Data Partner certification for NLP-assisted CCD enrichment. Our customers have had a 100% success rate with their NCQA auditors.