With Natural Language Processing You Can Have Your Cake and Eat It Too

Within electronic health records, there is an inherent tension between the need for discrete data and natural language.

Clinicians find value in both forms of information. The efficient administration of healthcare demands discrete, encoded data. Discrete data is ideal for quickly noting trends, graphing, encoding and machine processing for myriad uses.

On the flip side, narrative text tells the patient’s unique health story, including the more nuanced details of their symptoms and disorders. The narrative describes the patient’s preferences and care priorities, and provides an expanded picture of the patient’s social history. It provides a window into the mind of the diagnostician by expounding on the rationale for diagnosis and treatment decisions. These details are vital to good patient care and are required for effective communication between healthcare providers and between providers and patients.

What is Natural Language Processing?

Natural language processing (NLP) technology enables computers to derive computable and actionable data from text, especially when text is recorded in the form of natural human language (i.e., phrases, sentences, paragraphs). This technology allows humans to record information in the most natural method of human communication (narrative text), and then enables computers to extract actionable information from that text. NLP is also capable of analyzing the often non-standard grammatical constructions common in medical lingo. Natural language understanding (NLU) is a subset of NLP that uses reasoning, inference and semantic searching to help clinicians make decisions and take action.

The Cake

The Health Story Project was founded on the premise that a central component of medical records is the narrative, textual information that conveys the patient’s unique story as well as the clinician’s thoughts and rationale for treatment decisions. Check boxes and drop-down lists simply cannot capture these nuances.

RELATED: How Your Voice Can Shape Interoperability

When the record consists of only canned responses, all patients with diabetes, for example, end up looking the same on paper. User interfaces that limit a clinician to record information about the patient’s story using check boxes and drop-down lists is not cake. It is a problem that creates burden and dissatisfaction for physicians who are forced to work in a way that is natural for computers, not humans.

Eating the Cake

What if we could have our cake and eat it too? What if we could have that rich narrative and the ability to machine process and encode important concepts? NLP bridges that divide and makes both possible.

NLP is not a new technology. In recent years, it has advanced sufficiently to be adopted in healthcare. Two drivers of NLP adoption in particular include the explosion of data analytics applications for both evidenced-based care models and administrative uses, as well as value-based care models, which require a more detailed patient record to support maximum payment levels. These market drivers crave data. However, clinicians are too time-starved to provide both discrete, encodable data and rich narrative. NLP is one technology that will aid clinicians in improving the quality, accuracy and completeness of documentation.

RELATED: Interoperability Takes a Village

NLP can be applied in numerous use cases in healthcare. It enhances the quality of documentation and enables secondary and downstream uses of the information captured during patient encounters. For example, in a post-encounter scenario, NLP can codify concepts in a standard terminology such as Systematized Nomenclature of Medicine – Clinical Terms, or SNOMED CT. In real-time documentation scenarios, NLP prompts clinicians to provide additional detail to assure proper reimbursement. In other use cases, NLP can be used to analyze records to detect care gaps.

The Health Story Project’s recent Roundtable explored three specific use cases for NLP. As our presenters demonstrated, NLP makes unstructured data relevant and actionable. It derives contextual meaning from the clinician’s textual notes and unlocks critical portions of the record for both clinical and administrative uses. In keeping with the Health Story Project’s mission, this roundtable highlights the value of narrative and shows how NLP can be used to boost its power.

The views and opinions expressed in this blog or by commenters are those of the author and do not necessarily reflect the official policy or position of HIMSS or its affiliates.

Interoperability Definition Call for Comments

HIMSS is updating our definition of interoperability and we’d love to hear from you. Review the latest definition as shared by HIMSS and provide your feedback by March 23.

Interoperability & Health Information Exchange Committee
Help advance interoperability and standards-based health IT systems that lead to meaningful health information exchange. Join the committee

HIMSS Interoperability Call to Action
Our Call to Action offers guiding principles to inform health policy and spur our nation’s health sector to action. Become a supporter

Originally published December 18, 2018


Article by channel:

Read more articles tagged: Natural Language Processing