Get Involved With Computer-assisted Coding Decisions
If your organization is considering computer-assisted coding, don’t panic! Technology is changing how people work in virtually every industry and medical coding is no different. Keeping up with the technology means you will continue to succeed.
Consider your job to be part of an evolution—computer-assisted coding enhances the role of a coder, making you more productive and giving you more time to spend on challenging reports that need your expertise. Rather than code your 300th screening mammogram of the day, you’ll be able let the technology handle it. The system will give you the chance to focus on more advanced coding work, auditing and quality control. And remember: While computers excel at quickly moving through well understood and repetitive aspects of coding, they can be quickly derailed when new concepts are introduced. The continual flow of new language and challenging reports will always preserve the role of the coder.
Instead of fearing the unknown, take the initiative to get involved with your organization’s decision-making process. Learn all you can about computer-assisted coding and know the right questions your organization should ask of any vendor.
What differentiates “computerized coding” from “computer-assisted coding?”
Sophisticated computer-assisted coding systems are built by software engineers and coders—continuously weaving coders’ expertise and experience into the technology. Computerized coding, on the other hand, cuts human coders out of the loop—risking their experience, as well as compliance and accuracy.
How does the system tackle code generation?
Several systems are available and branded as computer-assisted coding, but their approach to coding leads to widely varied results and risky compliance issues. Be wary of “word suggestion” systems that propose codes based on what’s entered in a medical report. Likewise, ontology-based systems attempt to assign codes based on language specific to a particular practice of medicine.
Systems using rules-based natural language processing, or computerized language analysis, are programmed to detect phrases such as “consistent with pneumonia” to code for pneumonia, but may not always catch later findings that would change a diagnosis.
The most sophisticated systems use a combination of natural language processing, statistical analysis and human coder decision making. With natural language processing at its core, these systems take into account the codes assigned, as well as the documentation for the codes, and the decisions people would make if they reviewed the same report.
Can the vendor’s system handle challenging concepts?
Vendors should be able to prove how their computer-assisted coding systems handle difficult concepts such as negation. Can they differentiate between “no evidence of pneumonia,” “consistent with pneumonia” and simply “pneumonia?” Another test would be to examine how a system works within the context of the report. For example, can it decipher that “broken neck of femur” means a broken leg should be coded, rather than a broken neck? Finally, determine if the vendor can maneuver through time references in medical reports, distinguishing chronological references to appropriately code historical X-rays vs. new ones or comparative studies.
By learning the intricacies of the myriad systems on the market, you can help your organization make an effective choice not only for the company as a whole, but for you and your fellow coders.