While supplies last! Free 2022 Essentials of Interventional Radiology Coding book with every ICD10monitor webcast order. No code required. Order now >

With the specificity of ICD-10, it is important to use tools and technology to assist with coding. Computer-assisted coding (CAC) is one method that can benefit a practitioner, but can also harm them if it is not set up or used appropriately. 

Technology is only as good as it is programmed, and like the old saying goes, “garbage in, garbage out.” This is very true when documentation is concerned.  Computer-assisted coding can be incorporated with the electronic health record (EHR), or it can be separate, depending on the capability to interface.

What is CAC?

CAC is a software tool intended to assist with documentation and code assignment that automatically generates a set of medical codes for review and validation based on documentation. Many physicians of various specialties are beginning to use this technology to improve efficiency in their medical practice.

CAC may be structured in two different ways. One way is natural language processing using artificial intelligence to extract data and terms from a text-based document in order to convert it into medical codes to be used and edited by a physician or coder. Physicians will develop preferred terms that are incorporated into the medical record. 

The second method is called structured input or codified input, and it is based on menu items chosen via a template that is dropped into the medical record. The physician will select a menu item based on the condition or diagnosis treated.

Structured input does not require the physician or coder to select the code. It is automatically dropped into the claim. One of the advantages of this method is the reduction of transcription costs, since once the encounter is finished, the documentation is complete.

Both methods employ a rules engine to assist with automated assignment of code(s).


CAC has become a valuable tool for hospitals and physicians to capture appropriate documentation to support both the procedure and diagnosis. It allows coders to review more complex coding cases, whereas more routine procedures and services are seamlessly submitted to the payor for payment.  However, CAC is still just a tool – if built or used improperly, it can be problematic.

CAS can be of great benefit to providers. Specifically, it can:

  • Increase medical coding productivity and efficiency
  • Increase medical coding consistency
  • Create a medical coding audit trail
  • Create queries
  • Allow for more comprehensive medical code assignment
  • Decrease medical coding costs
  • Use free text for recording documentation
  • Improve systems through feedback


Some of the pitfalls when using CAC include:

  • Costs of hardware and software
  • Templates not billed properly to ensure accurate code selection
  • Potential increase in errors in the coding process
  • Lack of industry standards
  • Technology limits
  • Appearance of cloning

How Templates and Workflow Affects CAC Success

CAC does very little for coding accuracy unless you do the work on the front end to ensure that templates, interfaces, and algorithms are built properly for accurate and detailed code assignment. No longer can we use an abundance of unspecified diagnosis codes when there are more specific codes we can select.  In addition, the EHR must be customized to support clinical documentation improvement (CDI).

In a structured CAC environment, each note is generated using predefined structured documentation, such as in templates. The coding engine assigns a code based on the documentation and code edits. The physician or coder confirms the code assignment. If the documentation lacks detail, the diagnosis code selected will not be accurate. For example, if the documentation states “fracture of the left femur,” category S72.- would be selected as unspecified fracture of the femur, but more information is required, such as type of femur fracture; whether it is the initial, subsequent, or sequela encounter, along with laterality; and whether the fracture is open or closed, type I, II, IIIA, etc. So the template that has been built will not be sufficient. In this instance, an unspecified category is selected, which could trigger payor denials.

When natural language processing (NLP) is used, the coding engine reads the document and selects potential codes, which may or may not be sent to the coder for manual coding review. Words, phrases, and sentences are analyzed by the software, and, based upon a set of underlying rules, codes are generated.  

Physicians and medical coders cannot rely on CAC systems alone. They need to verify the suggested codes even when clinical documentation is specific enough to support ICD-10 diagnoses. 


Are denials being generated due to CAC? If the workflow is structured, whether in a structure of NLP or not, the key to reduction of denials is, of course, complete and detailed workflows and templates. In addition, there must be a manual process as part of clinical documentation to review documentation and coding from a CAC system. In addition, one of the challenges is “cloning.” And how do you create more efficiency in the medical record without cloning?


You May Also Like

Leave a Reply

Your Name(Required)
Your Email(Required)