Guest Post: Natural language processing and clinical documentation, part 1
by Crystal R. Stalter, CPC, CCS-P, CDIP
Long before ICD-10-CM/PCS became a focus, working as a clinical documentation improvement (CDI) manager to improve physician progress and/or operative notes was a challenge—doctors either got it or they didn’t. But as the transition from paper charts to an electronic health record (EHR) began, providers started to understand how to better document their visits, since they had to choose from drop-down menus and multiple options to complete their notes. Then, as ICD-10 approached, a new awareness of medical necessity denials and revenue impact took shape. Providers began looking for ways to document better in less time.
Enter natural language processing (NLP) and front-end speech. Products and software came along that allowed physicians to dictate their notes into the EHR in real time. NLP software can learn a provider’s speech patterns, making each episode of documentation faster and more accurate.
Working with documentation software
Some providers embraced this technology with open arms, and others—well, they were less enthused. It has been interesting over the past six months or so to watch healthcare providers interact with their documentation software. Every day I strive to make technology better for the providers to give them more time with their patients. In addition, I want the technology to help communication between providers, clinicians, and coding staff.
Physicians have a love/hate relationship with EHRs. When I am with a physician typing away or dictating into a smartphone, it’s interesting to watch just how quickly the physician moves from field to field—and to see just how specific the created documentation is. At times, I have interjected that the provider may want to include type or acuity in the notes. Most of my providers know what I do for a living, so they are usually not offended by my suggestions. In fact, they laugh and say their coding manager thanks me every time I’m in the office.
At this point, though, we are more than a year and a half into ICD-10. So why are we still reminding providers about documentation? Without message overload, how can we develop content relevant to their specialty that is informative, efficient, and easily absorbed? Obviously it’s crucial to strive for the best documentation possible—correcting underspecified conditions at the time of the patient encounter has a direct ripple effect through the life cycle of the patient’s stay.
As an inpatient, the entire patient encounter can be reviewed by a CDI specialist, especially if the patient is a Medicare patient. The CDI specialist is tasked with reviewing the encounter in its entirety and ensuring the treated conditions are fully documented, which allows for CC/MCC coding opportunities that affect the DRG assignment.
With the transition of healthcare into a value-based reimbursement model, the capture of all conditions is imperative; it directly affects severity of illness and risk of mortality scores, which in turn affect the overall risk adjustment and quality scores of the hospital.
Technology like NLP software, which can help interpret the patient encounter and identify conditions with query opportunities, is paramount to making a CDI team more efficient. Partially automating the chart review process allows for more reviews during a CDI specialist’s day and potentially affects not just Medicare patients, but also other payers, creating an even greater impact on care quality and reimbursement.
Editor’s note: To read the second part of this article, return to the ACDIS Blog next week. Stalter is the CDI manager for M*Modal in Pittsburgh. Contact her at crystal.stalter@mmodal.com. Advice given is general. Readers should consult professional counsel for specific legal, ethical, clinical, or coding questions. Opinions expressed are that of the author and do not represent HCPro or ACDIS. This article originally appeared in JustCoding.