Computer-assisted coding (CAC) can live up to its namesake when paired with a credentialed coder to speed up the process without losing accuracy, according to a study in the Journal of the American Health Information Management Association. It doesn't fare so well when used alone.
The study looked at the ICD-9 procedure and diagnostic codes on 25 Cleveland Clinic cases, all of which were complex with an average length of stay of 16 days. Six coders assigned codes with the assistance of (CAC) technology and six without. Those with the technological help reduced time per record by 22 percent, according to an announcement.
Cleveland Clinic requires its coders to be 95 percent accurate, and it was able to achieve that level for both procedures and diagnoses with oversight of a human coder, but accuracy fell off when CAC was used alone.
Since the system "learns" over time--known as tuning--the CAC recall rate improved for coding both diagnoses and procedures between implementation and six months later. CAC applies natural language processing to electronic notes and reports, and cannot extract meaning from scanned documents, FierceHealthIT explained previously.
Hospitals have been looking to CAC technology to help alleviate what they foresee as a loss of productivity when they move to the more complex coding system ICD-10.
CAC initially gained traction in outpatient departments such as radiology and pathology where the codes are repetitive and related to a limited number of terms. Coding for inpatient stays is much more complex.
While many organizations are lagging behind schedule--if they've begun at all--on implementing ICD-10, training will be a huge issue for staff. A previous study in the Journal of the American Medical Informatics Association found that mapping the 50,000 new medical diagnosis codes and 70,000 new procedure codes from ICD-9 is far from clear-cut and in some cases, quite convoluted.