Are you ready for the changes that AI will bring to coding?
I am attending Healthcon this week and the buzzword is the increase in innovative technology as it relates to Artificial Intelligence (AI) and its impact on the role of the coder and healthcare in general. How much do you know about AI and are you ready for changes to your role?
Specific to coding, we first started seeing content regarding AI in CPT in 2022 although you will not see the term “AI” in the code set itself. Code 0691T (Automated analysis of an existing computed tomography study for vertebral facture(s), including assessment of bone density when performed, data preparation, interpretation, and report) was introduced that year. Since that time, there has been expansion of AI codes in radiology and pathology.
If you look at Appendix S, Artificial Intelligence Taxonomy for Medical Services and Procedures, you will find an explanation of the taxonomy which includes three categories to classify AI applications:
- Assistive – detection of clinical data without analysis or interpretation
- Augmentative – analysis of clinical data with quantification of results
- Autonomous – interpretation of clinical data with conclusions
Assistive and Augmentative AI processes require the interpretation and report to be completed by a provider (a physician or other qualified healthcare professional).
Autonomous AI processes have three levels of provider involvement:
- Level I – AI determines diagnosis and/or management options and requires the provider to authorize action.
- Level II – AI determines diagnosis and/or management options and alerts provider to authorize action or override.
- Level III – AI determines diagnosis and/or management and initiates management; requires provider action to override or change management.
CPT provides some examples of AI use in procedures in Appendix S:
Assistive:
- +0764T Assistive algorithmic electrocardiogram risk-based assessment for cardiac dysfunction (eg, low-ejection fraction, pulmonary hypertension, hypertrophic cardiomyopathy); related to concurrently performed electrocardiogram
- This code is used with 93000 and 93010.
- 0765T Assistive algorithmic electrocardiogram risk-based assessment for cardiac dysfunction (eg, low-ejection fraction, pulmonary hypertension, hypertrophic cardiomyopathy); related to previously performed electrocardiogram
Augmentative:
- 75580 Noninvasive estimate of coronary fractional flow reserve (FFR) derived from augmentative software analysis of the data set from a coronary computed tomography angiography, with interpretation and report by a physician or other qualified health care professional
Autonomous:
- 92229 Imaging of retina for detection or monitoring of disease; point-of-care autonomous analysis and report, unilateral or bilateral
Artificial Intelligence applications work in conjunction with Natural Language Processing (NLP) and Machine Learning (ML) to identify potential diagnoses in provider documentation, help reduce billing errors, and to determine if provider documentation fully supports procedures billed. Coders may see AI suggestions when coding or auditing and may see AI used to support denials of procedures in medical reviews based on the specific language (or lack thereof) in the provider documentation. These are just a few examples of how AI is utilized in healthcare today.
Because AI is relatively new in our industry, it is imperative that we educate ourselves in order to embrace evolving technologies. Again, the only constant is change so we must be ever learning, growing, and evolving as well. Here are some free resources for you:
- Intro to AI Concepts for Medical Billing and Coding (free to AAPC members)
- Navigating the Way Forward with Artificial Intelligence (AHIMA) (some for AHIMA members only)
- AI for Medical Coding – Enterprise-Grade AI (SullyAI) (scroll down)
- Free Guide: AI in Healthcare – AI Implementation Guide (PolyAI)
- AI for Medical Coding Explained (Semantic Health)
