by Aimee L. Wilcox, CPMA, CCS-P, CST, MA, MT
Dec 13th, 2022
Working with computer-assisted coding (CAC) opens our eyes to the many facets of coding not often considered. When trends are noted on a larger scale due to an engine processing high quantities of the same type of encounter or service, it is easier to see formatting, documentation, and coding issues than if you were evaluating significantly smaller numbers of the same document types. CAC is a process where the engine consumes formatted provider notes and based on programming, recommends specific ICD-10-CM/PCS, CPT, and HCPCS codes for coders to choose from. When coders accept the autosuggested code, the accuracy rate or success rate of engine functionality increases. When the coders reject the autosuggested codes, the accuracy rate is reduced. Analysts review engine performance and metrics to determine how successful and accurate the engine is in identifying and suggesting the correct code.
The engine is programmed to apply coding rules for successful code suggestions. It consumes the medical report and applies the programmed logic to determine which codes to recommend. Coders can look at the evidence the engine used to determine the suggested code and either accept it or reject it and then can add any codes the engine may have missed. This is a field of artificial intelligence (AI) that is ever expanding and very exciting.
CAC requires analysts to review hundreds of encounters, very quickly to identify coding trends associated with engine functionality, note formatting, provider documentation, and coder behavior, and to then provide feedback to both the engine and the coder about how to continually improve performance outcomes.
Identifying Coding Trends Improves Engine Performance
Identifying coding trends of correct or incorrect code suggestion and code reporting, are essential to ensuring accurate engine autosuggested codes; however, coder behavior plays a significant role in the process. Here are a few examples:
- Issue: The engine autosuggests ICD-10-CM codes from the history sections within the patient encounter but the coders do not accept the codes.
- Trend: After analyzing a percentage of the encounter notes where certain condition codes were rejected, it was noted the patients were often seen in the Emergency Department for an acute, traumatic injury. The suggested codes were correctly autosuggested from the patient’s history (i.e., HPI, ROS, PMFSH) but coders will only accept the suggested codes related to the actual injury and ignore or reject the others.
- Resolution: Additional research highlighted an article by the AHA Coding Clinic that recommended that coders only report codes for conditions documented as addressed during the encounter, which aligns with the ICD-10-CM coding guidelines and recommendations are for coders to adhere to these guidelines and choose which codes they’ll accept and which should not be reported.
How To Know Which Conditions Meet Reporting Criteria
A common acronym has been floating around related to identifying reportable conditions. MEAT is an acronym that stands for Monitored, Evaluated, Addressed, and Treated. It indicates that if one of the document supports a condition meets the criteria of MEAT, it is considered a reportable condition.
The following is a summary of actions that are identifiable for each of these four areas of assessing whether MEAT has been met:
- Monitoring: The documentation includes signs, symptoms, disease progression or regression, continued surveillance or observational findings.
- Evaluation: The provider documented the current state of the chronic condition (e.g., stable, exacerbated, worsening, uncontrolled). There are examination findings, test results reviewed, discussion of medication or treatment efficacy or notations of the patient’s response to treatment(s). The provider has noted assessment of late effects of a condition or illness.
- Assessment: Documentation exists of a discussion of a chronic condition and how it should be monitored, followed up with the recommended specialist. Records and test results are reviewed. Counseling is provided or a discussion noted about additional evaluation or ordering of additional testing to determine the status of the disease, progression, or regression. Patient care or plans are discussed with another provider on the patient’s care team, internally or externally.
- Treatment: The current treatment is documented, reaffirmed, or changed. There is an increase or decrease in medication dosing. A referral to a specialist is made, additional testing, imaging, studies, or therapies are ordered. Patient education is provided related to a specific condition, recommended or ongoing treatment. Additional treatment options are recommended, discussed, or summarized. The treatment plan is reviewed and updated.
Understanding what to look for when trying to determine if a historical code should be reported for a specific encounter takes practice and having reference information at your fingertips can ensure coding and auditing success.
About Aimee L. Wilcox, CPMA, CCS-P, CST, MA, MT
Aimee L. Wilcox is a medical coding, billing, and auditing consultant, author, and educator with more than 30 years of clinical and administrative experience in healthcare, coding, billing, and auditing. Medicine, including coding and billing, is a constantly changing field full of challenges and learning and she loves both. Aimee believes there are talented medical professionals who, with proper training and excellent information, can continue to practice the art of healing while feeling secure in their billing and reimbursement for such care.
Quick, Current, Complete - www.findacode.com