By Suhas Nair, Director of Product Management, AGS Health
LinkedIn: Suhas Nair
X: @AGS_Health
Autonomous coding is an AI-driven leap forward from traditional Computer-Assisted Coding (CAC) systems, promising transformation for labor-intensive and error-prone coding processes in healthcare, with substantial benefits in terms of cost savings, efficiency, and reduced denials. However, to optimize its value and protect investments in coding solutions, healthcare organizations must understand the realities and limitations of this evolving technology.
The Emergence of Autonomous Coding from CAC
Historically, the coding process in healthcare has been a multi-step, human-driven task with inherent risks of errors. The advent of CAC brought notable strides in improved accuracy. However, the integration of artificial intelligence (AI) has propelled this further, enabling a shift in the coder’s role. Coders are now moving away from routine coding tasks, focusing more on vital areas such as validation and auditing.
Autonomous coding utilizes a blend of technologies like Machine Learning (ML), Natural Language Processing (NLP), Clinical Language Understanding (CLU), computational linguistics, knowledge graphs, and Large Language Models (LLMs). These technologies work in tandem to analyze and categorize extensive patient data sets, including electronic health records, physician’s notes, laboratory reports, and diagnostic data. This advanced AI capability not only assigns diagnosis and procedure codes with high accuracy and efficiency but also allows human coders to concentrate on performing more complex chart reviews and providing feedback crucial for the continuous learning and adaptation of the AI system.
This transition reflects the broader trend in healthcare toward automation and data-driven decision-making that aims to pave the way for streamlined operations, improved financial outcomes, and, most importantly, a better patient experience. The efficiency and precision offered by autonomous coding are particularly advantageous for healthcare organizations facing staff shortages or those seeking to reallocate their workforce to higher-value tasks.
The Current State and Benefits of Autonomous Coding
Today, autonomous coding is positioned as a transformative tool in healthcare revenue cycle management. Its primary advantage lies in its ability to expedite the coding process, thereby accelerating billing and reimbursement cycles. This efficiency leads to a substantial reduction in labor costs, as the need for extensive manual coding is greatly diminished. Furthermore, the precision of AI algorithms in autonomous coding systems minimizes human errors, enhancing the accuracy of coded data.
Beyond operational efficiency, these systems offer scalability, making them suitable for handling large volumes of data and adapting to the growth of healthcare organizations. They also provide valuable insights through data analytics, contributing to strategic decision-making and the organization’s improved financial health.
The benefits of autonomous coding, as reported by early adopters, include:
- A 50% reduction in coding costs.
- A decrease of 3 to 5 days in Accounts Receivable (A/R) days or Days Not Final Coded (DNFC).
- A 25% to 50% reduction in denials related to coding.
- A 50% reduction in Full-Time Equivalents (FTEs) required for coding tasks.
The Misconception of Full Autonomy
Perhaps the greatest limitation of autonomous coding is that despite advancements, it has not yet achieved complete autonomy. While autonomous coding technology has advanced significantly, its current state requires supplementary systems and human oversight. This technology excels in areas with repetitive data but struggles with variability, leading to higher automation levels in certain specialties like radiology and emergency departments.
Furthermore, most autonomous coding systems in the market do not have (pre-bill) audits or the abstraction of additional parameters – such as those required by the Merit-based Incentive Payment System (MIPS) in their scopes of work. This means customers require add-on tools like CAC, Clinical Documentation Integrity (CDI), and encoders for processing charts with lower confidence scores. Organizations relying on autonomous coding may find themselves in the challenging position of managing multiple vendors to address these gaps.
The potential lack of integration between autonomous coding solutions and these additional tools presents another challenge. One which can lead to fragmentation across different coding, auditing, and compliance functions. These siloed processes not only complicate vendor management but also hinder unified reporting.
The Road to True Autonomy
Today’s autonomous coding solutions are not entirely autonomous due to several factors:
- AI-based solutions require access to a comprehensive body of quality data for proper training and continuous learning.
- Coding often involves guidelines open to interpretation by human coders.
- Disparate systems used for overflow coding and auditing can limit the feedback loop, hindering the AI system’s ability to learn.
- Vendors solely focused on autonomous coding may lack foundational knowledge derived from work in key areas like CAC and CDI, impacting training quality.
The good news is that AI is constantly improving. With the right data and training, autonomous coding technology can move closer to true autonomy over time. In the meantime, organizations can still derive significant value from the current level of autonomous coding maturity, especially when addressing specific scenarios such as physician burnout or specialty coding.
Choosing the Right Vendor Partner
Once an organization decides that autonomous coding aligns with its strategy, selecting the right vendor is crucial. To reduce the need for multiple vendors and ensure a smooth implementation, healthcare organizations should look for a vendor offering an autonomous coding solution that includes one platform with fully integrated supplementary tools and services.
To identify the most suitable vendor for autonomous coding, consider these key questions:
- What are the primary goals for implementing autonomous coding? Objectives may include reducing costs, alleviating physician burnout, enhancing coding audits and quality, addressing staffing challenges, or resolving vendor-related issues.
- What are the current audit processes in place? Understanding these will help determine the integration and enhancement needs for autonomous coding.
- What is the extent of automation required to add value? Determine if partial automation or automation focused on a single specialty would meet the organization’s needs.
- Will there still be a need for supplementary systems like encoders, CAC, CDI, etc.? If so, how will these be managed in conjunction with autonomous coding, and what will be the associated costs? Consider whether other vendors will be necessary to provide supplementary coding services beyond the scope of the autonomous coding system.
By aligning expectations with the realistic capabilities of autonomous coding, organizations can make informed decisions, choose the right vendor, and implement a system that truly aligns with their needs. The future of autonomous coding is a path toward transforming healthcare coding processes, enhancing accuracy and efficiency, and paving the way for improved patient care.