Artificial Intelligence for Medical Coding: Challenges & Considerations

Artificial Intelligence (AI) has made significant inroads into various industries, including healthcare and medical coding. With the promise of improved efficiency, accuracy, and reduced costs, AI has been hailed as a game-changer in these sectors — but does it actually live up to its healthcare hype?

The answer may surprise you. Yes, AI can automate certain tedious manual tasks within healthcare and medical coding. However, it’s critical to recognize that relying solely on AI, without any human oversight, brings with it its own set of drawbacks warranting careful consideration. Understanding the potential negative impact AI can have in the healthcare industry can help separate fact from fiction.

Artificial Intelligence Brings With It a Hefty Price Tag in Healthcare

One of the biggest considerations with Artificial Intelligence in healthcare is overall cost. Recent reports revealed that AI implementation in healthcare often costs up to $1,000,000. While the potential benefits of AI in healthcare may seem impressive on paper, the actual cost of implementing and maintaining such technology can be prohibitive — and cannot be overlooked.

Beyond implementation and maintenance costs, there is a wide range of periphery costs that come into play with AI technology in healthcare. Some factors that drive overall expenses include:

Scope and Complexity of the Project
The scale and complexity of an AI project significantly impact its cost. Developing a simple AI system for a specific task may cost on the lower end of the spectrum, while more complex and multifaceted solutions can require a substantial investment.

Data Requirements
High-quality, diverse, and well-curated data are essential for AI algorithms to perform effectively in healthcare. Gathering, cleaning, maintaining, and updating such data can be a costly process.

Customization and Integration
Healthcare institutions often require AI solutions tailored to their specific needs. Customization and integration with existing systems can add to the overall cost.

Hardware and Software Infrastructure
The hardware and software required to run Artificial Intelligence applications in healthcare can be a significant cost driver. High-performance servers, cloud services, and software licenses contribute to the expenses.

Regulatory Compliance and Security Requirements
Compliance with healthcare regulations and data security standards adds a significant layer of complexity and cost to AI implementations in healthcare.

Training and Maintenance
Training staff, maintaining AI systems, and keeping them up-to-date with evolving medical knowledge and technology contribute to ongoing expenditures.

Beyond Costs: Other Potential Drawbacks to AI in Healthcare

Overall expense isn’t the only disadvantage to using artificial intelligence in healthcare and medical coding. Other potential drawbacks to wielding unmanaged, unsupervised AI to replace human interaction include: 

Limited Understanding and Judgment
AI systems are excellent at processing vast amounts of data and identifying patterns. However, they often struggle when it comes to handling complex and unique medical cases that require nuanced interpretation, judgment and decision-making. Patients with rare diseases, complex conditions, or atypical symptoms might be coded incorrectly, possibly leading to a misdiagnosis or inappropriate treatment if AI algorithms are solely relied upon.

Security and Privacy Concerns:
Healthcare data is highly sensitive, and maintaining the security and privacy of protected health information is paramount. A 2023 Medical Economics article stated that “One of the biggest risks is the potential for data breaches. As healthcare providers create, receive, store, and transmit large quantities of sensitive patient data, they become targets for cybercriminals. Bad actors can and will attack vulnerabilities anywhere along the AI data pipeline.” Relying on AI introduces new vulnerabilities, as any breach or misuse of AI systems can compromise patient confidentiality and expose personal data to unauthorized parties.

Lack of Empathy and Human Touch
In healthcare, empathy and the human touch are invaluable. AI lacks the ability to understand and respond to the emotional and psychological aspects of patient care. Patients often seek solace and reassurance from healthcare providers, and a complete reliance on AI can result in a cold and impersonal patient experience.

Data Bias and Discrimination
AI algorithms are only as good as the data they are trained on. If the training data is biased or incomplete, the AI system can propagate and even exacerbate healthcare disparities. For example, AI systems used in medical coding may struggle with State or insurance-specific requirements, leading to incorrect billing and insurance claim issues.

Dependency and Skill Erosion
The more healthcare professionals rely on AI, the more they may become dependent on the technology. This dependency can lead to a decline in the clinical skills and/or professional skills as they become less adept at interpreting and analyzing medical information themselves.

Incorrect Conclusions
AI systems are not infallible, and they can fail or produce inaccurate results for various reasons, including technical glitches, data corruption, or cyberattacks. Additionally, reports show that “AI is only as good as the information with which it starts working. Without near perfect historical data, it is incredibly easy to confuse the AI program into making some serious mistakes.” Often known as “garbage in, garbage out,” the theory holds that without perfect data entry, AI will create its own outputs and draw its own conclusions, which can lead to error. In a healthcare setting, such failures can have life-threatening consequences for patients.

Ethical Dilemmas
The use of AI in healthcare raises numerous ethical dilemmas. Who is ultimately responsible if an AI system makes a wrong diagnosis or recommendation? How do we ensure that AI decisions align with ethical and moral standards? These questions remain largely unresolved.

Pena4: Medical Coding Solutions With a Human Touch

Joe Gurrieri, President & COO of Pena4, a leading revenue cycle consulting company, states “AI can certainly assist medical coders in identifying key clinical documentation present in the medical record that may require ICD-10 diagnosis or procedure codes or HCPCS/CPT codes. Allowing AI to assign all of the codes without any human intervention or review is not recommended and is likely to lead to coding and billing errors. At this point in time, it is our opinion that nothing can provide better accuracy than a highly skilled, experienced human coder.”

Pena4’s team of experienced medical coding professionals delivers customized solutions with human touch. Contact us today to learn more about our personalized approach to medical coding.