Natural Language Processing Turns Text into Actionable Data
The next 12 to 18 months are going to be rocky ones for health information management (HIM) professionals, as they work with physicians on projects such as meaningful use and the ICD-10 transition. However, technology solutions like natural language processing (NLP) may prove beneficial in helping healthcare sort through the waves of data headed their way.
In an AHIMA-sponsored webinar “Building Data-Driven Workflows: More Than Just an EHR,” Steve Bonney, executive vice president, business development and strategy at RecordsOne, noted that while 80 percent of physicians are using EHRs, interoperability has been—and will continue to be—a frustration for them. A key barrier to EHR interoperability, he noted, is that 20 percent of EHR data is unstructured, and much of that comes from the free-text portion of the medical record.
“Being able to get the specificity of the documentation from the physicians today is the challenge. Codifying it and making it searchable is a completely different challenge,” Bonney said.
NLP is a technology that can take free text, process it, and turn it into actionable data points. Then, NLP software delivers the data to clinical documentation improvement (CDI) specialists and their associates to improve clinical decision making. Additionally, NLP has the ability to process all of this unstructured data in real time, whereas in the past CDI specialists have reviewed documentation retrospectively, slowing down billing and the querying process.
Bonney concluded by noting that overcoming barriers to interoperability is going to require more than NLP.
“Despite EHRs being everywhere you’d probably all agree with me that there are still significant gaps that need to be addressed. There’s a big difference between text data and information. Text can be formed into information, but it’s not enough. We need multiple technologies, NLP is just one,” Bonney said.