November 16, 2022

Our industry has seen a tremendous amount of change over the past decade. Electronic medical records, ICD-10 and computer assisted coding (CAC) to name a few. The latter, CAC, is a software intended to help with documentation and code assignment.

CAC is built two different ways. The first is natural language processing. This uses artificial intelligence to extract terms from a text-based physician document and is converted to medical codes to be validated and/or edited by the coder.

The second way is called the codified input method. It is based on menu items chosen from a template that is dropped into the medical record. Codified input method does not require coder validation as the code is automatically put on the claim. This method can reduce transcription costs.

The pros and cons of implementing CAC are numerous. A few of the pros and cons of the natural language processing are listed below:


  • Increases coding productivity and efficiency
  • Creates a coding audit trail
  • Can increase coding consistency


  • Costs of hardware and software
  • Templates not built properly to ensure accurate code selection
  • Coders not validating codes and just accepting what is given, increasing coding errors
  • Lack of industry standards
  • Technology limitations

The CAC is there to support complete and accurate coding.  There are times when the CAC will highlight and find a code that the coder missed. For example, finding hydrops of the gallbladder in the operative report when this was not mentioned anywhere else in the record. Without CAC, this may have been missed and the DRG could possibly be impacted.

The CAC can save time in the coding process.  Anemia is documented everywhere in the chart and a transfusion is done but the coder cannot find documentation of “acute blood loss”.  The CAC coded acute blood loss anemia and highlighted the documentation on one progress note that was buried among 67 progress note pages.  The CAC identified documentation impacting coding that was difficult and time consuming for the coder to find.

Codes should always be assigned to the highest level of specificity.  There are times when the CAC can fall short in identifying the most specific code. For example, fracture of femur is picked up by CAC. This does not tell you laterality, initial, subsequent, sequela, open or closed or specific type of fracture. The coder would need to investigate for further specificity for code assignment.

Coder training upon CAC implementation is critical.  Coders are moving from a pure coder role into a coder/code validator role.  Coders must understand that they cannot rely only on CAC and must validate that the clinical documentation supports the code choice. Codes prompted by the system must be validated.

There are pros and pitfalls to using CAC that vary by hospital, physician, patient type and most importantly quality of documentation. We as coders must remember that our job has not changed — the tools have.



Alicia R. Blamble, RHIA

Managing Auditor | Excite Health Partners


Lisa Marks, RHIT, CCS

VP of HIM | Excite Health Partners