CAC Takes Coding
Into the Future
By Kristi Fahy, RHIA
TECHNOLOGY IS THE segue to the future. It is the very thing
that will transform the landscape of today’s world into one
with endless possibilities. In healthcare, that transformation is being accelerated by the transition to electronic health
records (EHRs), which has helped providers to digitize and
drive the organization with data every day. Technology is the
catalyst to unlocking new opportunities, many of which will
help push healthcare organizations to a more effective future.
Companies that embrace technology will surpass their competitors and stay ahead in the marketplace. Meanwhile, companies that are content to wait will be left behind.
Given today’s shift to value-based reimbursement methodologies, both clinical documentation improvement (CDI) and
coding accuracy are more important than ever. These specialties are necessary for improving patient outcomes, ensuring
continuity of care, and attaining appropriate reimbursement.
The availability and quality of clinical documentation directly
impacts which codes can be assigned. Additionally, the codes
assigned for each patient encounter directly impact reimbursement—a sentiment often expressed in the coding mantra: “if it isn’t documented, it didn’t happen.” That’s why many
healthcare organizations have sought to streamline their coding and CDI processes with technology through computer-assisted coding (CAC).
AHIMA defines CAC as the “use of computer software that
automatically generates a set of medical codes for review, validation, and use based upon clinical documentation provided
by healthcare practitioners.” 1
CAC technologies, designed to be a productivity tool, have
AI Meets Coding
come a long way since they first made their debut in the ear-
ly 1990s. In more recent years, before CAC programs were
broadly adopted in the mid-2000s, many organizations be-
gan to fear what was to come from the long-anticipated shift
to ICD- 10. Although the transition was still several years
away—ICD-10-CM/PCS was implemented in 2015—health
information management (HIM) professionals questioned
how they were going to manage the new complexity of
ICD- 10. With over 75,000 additional ICD-10-CM codes and
increased specificity requirements for ICD-10-PCS codes,
healthcare organizations needed an approach that would
consistently capture accurate codes without hindering pro-
ductivity. For many, that approach was CAC.
Traditional CAC technologies utilize natural language processing (NLP) to gain insights from large amounts of plain
text data extracted from the EHR. NLP is used to scan documents and provide code suggestions for coders to review and
validate. The power of NLP lies in its ability to help coders
quickly find key words or phrases and suggest affiliated codes
within the documentation for more efficient coding.
Still, as a rules-based algorithm that is dependent on the
context of the documentation, NLP has its limitations. If the
rules don’t accommodate the variances in the way the documentation is written—“left ankle fracture” vs. “the patient has
a fracture of their left ankle”—NLP may suggest unspecified
codes. In addition, clinical findings can be written across several sentences, paragraphs, and documents due to NLP’s reliance on consistent patterns. When patterns change, the engine needs a human to program new rules.