4 questions to ask when evaluating Natural Language Processing (NLP) for risk adjustment

Natural language processing (NLP) has proven to be a highly effective way to hone in on specific, relevant clinical information to fuel effective risk adjustment. But there are a range of NLP strategies and options. Which best serves your needs? Here are four questions to guide your evaluation.

1. Is NLP integrated into the risk adjustment workflow?

Why integration may matter to you: Often NLP is a separate activity, taking place outside of the other risk adjustment process steps. A more efficient model strategy embeds NLP as part of the workflow to maximize identification of the records most likely to have relevant clinical information.

2. Does the NLP offering employ computer-assisted coding?

Why computer-assisted coding may matter to you: This emerging capability can speed coding and elevate accuracy by suggesting codes based on the chart. The coder can then, at his or her discretion, quickly approve, correct or reject the suggestion. Think of it as a way to improve the coder’s productivity and accuracy.

3. Is the NLP algorithm designed to reveal the charts most likely to merit a second look?

Why a more precise approach may matter to you: Laterality. Severity. Acuity. Exact body part affected. Coders find codes based on these and many more granular elements. A robust NLP strategy uses smart suspecting analytics to identify the charts most likely to contain documented conditions lacking correct specific coding.

4. How confident are you that your NLP strategy can extend to prospective risk adjustment?

Why prediction and early identification may matter to you: NLP is quickly evolving to provide early identification of new disease conditions and disease progressions, documentation trending across physician, hospital and specialist charts, and other advanced insights that help support correct interventions earlier than ever before. 

At a time when coding is more important than ever to help you improve clinical outcomes and support appropriate reimbursement, NLP – particularly when integrated into risk adjustment workflow – is increasing in prominence as a driver to maximize retrospective and prospective risk adjustment success.


Learn more about risk adjustment workflow-based NLP:


About the Author: eli wolter headshot

Eli Wolter, Strategic Product Director, Optum

Eli currently leads product and technology strategy for retrospective risk adjustment at Optum where he is focused on delivering innovation and value to clients through comprehensive, fully integrated retrospective risk adjustment programs. He has 16 years in information technology, much of that time with a healthcare focus. Eli has previously held roles in management information systems, healthcare process automation, software development, project management, solution delivery, and product strategy.

Leave a Reply