Ask Erika about: Natural Language Processing
Why is NLP becoming more commonly used in healthcare?
Two major factors are driving market adoption of NLP. First, the computer science research behind NLP has advanced at an astonishing rate over the past 10 years, which means much higher accuracy. And second, the move to value-based care and importance of population health makes it essential to understand how we are providing healthcare in much more detail, and at much greater scale. That’s a great opportunity for NLP.
How does NLP work? What are some of the opportunities and challenges when using NLP?
Natural Language Processing “reads” clinical notes that are generated by providers and usually stored in an Electronic Medical Record (EMR) system. NLP tries to imitate a skilled clinical abstractor, but in a different way. And unlike manual abstractors, NLP can scale to entire populations and get you answers more quickly. NLP also has limitations. For example, it works best when the scope of the task is more limited (for example just Quality or just Risk Adjustment). Another limitation is that the technology may need to be customized to your data and maintained over time as your data changes.
What are the most important things to consider when selecting an NLP vendor for my quality program?
Three things are really important. First, how will your vendor address the NLP “portability problem” to make the system work in your organization. Second, how well does the vendor evaluate and track its accuracy for your use case and maintain that accuracy over time? Third, how well is the NLP technology integrated into real workflows, and how do they measure the value to you.
What is the NCQA NLP working group all about?
The NCQA NLP working group was established in early 2020 to help NCQA determine how to use NLP for quality measurement, and whether NCQA could play a role in certifying this technology. The Astrata team was proud to be selected to be part of the Working Group, while we were still being incubated as a company. We look forward to continuing to work with NCQA to help advance the field of digital quality measurement.
Is NLP part of the Future of HEDIS®?
As a big fan of the Future of HEDIS, we certainly hope so! In its Presidential Pitch, NCQA has included NLP as one of the methods that will be important for the next generation of HEDIS technologies (page 4). At Astrata, we see NLP as one of several technologies (along with FHIR, CQL and the QI Core data model) that will be critical to the future of quality measurement.
How can NLP help my organization increase our quality rates?
NLP can be an important first step in your HEDIS transformation, because it will help you change your operations and move to prospective, year–round quality reviews. But NLP can also help you improve your quality rating and unburden your providers.
How do I judge how good a vendor’s NLP is? How do I ask about accuracy?
NLP accuracy is one of the most important factors in determining how much value your organization will derive. Vendors that are actively monitoring their accuracy should be able to give you performance metrics for specific measures and should separately measure false positives and false negatives. They should have a way to assess how well the NLP is working after deployed and over time, and they should be able to make changes to keep the system functioning within acceptable limits. They should be transparent about all of this with you.
Ask Erika about: Quality Measures
What kinds of measures can Astrata NLP handle?
Astrata technology can be used for a huge range of measures, from population-measures used by payers (e.g. HEDIS), to provider-focused measures (e.g. MIPS). We also provide measures in a wide range of domains including primary care, subspecialties, and behavioral health.
Does Astrata cover measures for Star Ratings, state Medicaid, and QRS measures for ACA Marketplace plans?
Yes. Astrata technology includes many measures that are used in CMS Star Ratings, state Medicaid pay for performance programs, and the ACA Marketplace QRS program. If your organization is focused on specific measures or measure programs, we would like to learn more. We can provide you with a summary of measures that we currently include, as well as an analysis of which of your measures of interest we can produce quickly for you.
What if Astrata’s NLP doesn’t include the quality measures I’m interested in? Can Astrata add new and/or custom measures?
Yes. Astrata’s technology is designed to make it easy to create custom NLP measures. This can be added to our standard offerings for organizations that have more specific needs.
Ask Erika about: HEDIS® Operations
Our organization is interested in moving to year-round HEDIS review. Can Astrata help?
Congratulations! It’s a big step to move to prospective HEDIS, and an even bigger step to expand a small prospective program. We’ve done the math ourselves. And while year–round review may seem daunting, our proven technology can help you make this leap.
What parts of my HEDIS operation can Astrata help me with?
Astrata technology can transform your HEDIS operations: (1) move your team to year-round review across multiple lines of business, (2) empower your quality intervention efforts with more timely data, (3) reduce the burden on your providers and increase the value of quality data you provide them. Astrata’s integration capabilities help you transform gently, by working within your existing work processes and technologies.
What are FHIR, QI-Core, and CQL? Why do they matter?
Fast Healthcare Interoperability Resources (FHIR) is a widely used data standard that helps systems and organizations interchange and use healthcare data with less manipulation. Quality Improvement Core (QI-Core) is the quality data model that NCQA is moving the industry towards. And the Clinical Quality Language (CQL) is a programming language that is designed to help automate and standardize the logic of quality measurement and decision support. Together, these new technologies will significantly change the way we measure quality. As CMS, NCQA start requiring these technologies, your organization will need to prepare for change.
Ask Erika about: Data
What type of data does Astrata work with?
Astrata NLP technologies work with “unstructured data” – clinical notes that providers write or dictate during and after encounters. Astrata’s FHIR and CQL technologies work with “structured data”. These are the data collecting in discrete form such as claims data and codes, observations, and laboratory results.
What kinds of data formats does Astrata allow?
Astrata’s NLP technology can take documents in many different forms, including CCD, HL7, PDF and plain text. As part of your advancing data strategy, we’ll work with what you have in place already and help you evolve to better, faster and more scalable ways to consume this data.
Manual review can only get you as far as the naked eye; to truly understand your data, you need technology that can reveal its full scope. Natural Language Processing — the decades-in-the-making, machine-driven processing of unstructured data — is a paradigm-shifter for Quality and value-based care, enabling a focused view of real-time and historical data at the scale of your entire population. But it’s not magic: setting up your data, and installing effective, accurate NLP, is a journey. Astrata has decades of experience developing NLP technology and a track record of successfully deploying it in clinical environments. We can help your organization navigate the journey to Quality-driven population health.