We recently had Melisa Tucker, Chief Product Officer at Nym, join Amy Faith Ho, MD, MPH on Becker’s Healthcare Podcast for a discussion on combining cutting-edge technology with deep clinical knowledge. While the discussion included many important points, we’ve selected one key insight to explore further in the format of a blog.
In this article, we’ll be taking a deeper dive into the complexity of clinical language, the challenges it creates when leveraging technology in revenue cycle management, and how emerging solutions are solving those challenges.
Listen to the full podcast episode.
When it comes to leveraging technology, healthcare is notorious for lagging behind other industries. The administrative side of healthcare is no exception, especially in revenue cycle management (RCM), where departments across the US continue to rely on manual processes that are error-prone, costly, and inefficient.
Why is this the case?
While constantly-changing regulations and complicated workflows can make it difficult to leverage technology for RCM, Dr. Ho emphasizes that it’s really the “messiness” and “obtuse vernacular” of the clinical data that presents the biggest challenge.
RCM departments oversee the medical billing process, which requires gathering and interpreting clinical data in patient medical records. This data - the patient’s medical history, the treatments and procedures they received, and so on - is written by providers using medical jargon. This clinical language is the “obtuse vernacular” referred to by Dr. Ho. It is essentially its own language, filled with abbreviations and idiosyncrasies, conveying complex medical concepts that are challenging to interpret.
For example, the abbreviation “O.D.” in a patient’s medical record can mean completely different things depending on the specialty. In ophthalmology, “O.D.” is short for “oculus dexter” which is Latin for “right eye,'' while, in the emergency medicine setting, “O.D” almost exclusively refers to “overdose.”
For another example, consider the following excerpt from a sentence in a medical record: “. . . Marshall Smith, MD patient…” On one hand, this excerpt could refer to a patient named “Marshall Smith,” who is also a doctor (MD). However, it could also refer to a patient who has Marshall-Smith syndrome and myelodysplasia (MD).
In RCM, especially during the medical coding phase, it is vital to correctly interpret clinical language in cases like the two above. It becomes clear that to successfully leverage technology in RCM, software must be capable of interpreting complex clinical concepts and deciphering the context embedded in the medical record. RCM technology must also be developed with sophisticated linguistic capabilities, such as understanding implicit vs explicit information, which is innate to humans.
Developing such technology is no easy task. There have been countless attempts to combine technology with deep clinical knowledge in RCM, but many solutions have fallen short of expectations.
Take computer-assisted coding (CAC), for example. Powered by natural language processing (NLP), CAC software scans medical record documentation, identifies key terms, and suggests reimbursement codes that support the treatment or service provided during the patient’s visit. In theory, CAC enables coders to process charts at a faster rate by automating the analysis of physician notes for code assignment.
In reality, the technology behind CAC does not have the ability to correctly interpret complex clinical language, often suggesting codes that do not accurately reflect the patient encounter described in the medical chart.1 Because of this, RCM departments that leveraged CAC still needed human coders to validate the accuracy of every suggested code.
CAC is a clear example of how the complexity of clinical language makes it challenging to successfully leverage technology in RCM. However, technology has developed since the emergence of CAC, and there are now solutions capable of interpreting clinical language with outstanding accuracy.
With its multidisciplinary team of medical doctors, linguists, computational linguists, and software engineers, Nym has developed technology that automates the accurate interpretation of clinical language. Through a combination of clinical expertise, computational linguistics, and explainable AI, Nym’s technology can decipher clinical language and unlock the vital data within medical records. Learn more about Nym’s clinical language understanding (CLU) technology.