Over the years, computer-assisted coding (CAC) has proven its ability to boost revenue team productivity and accelerate critical decision-making while reducing denials, missed charges, and low-risk scores. CAC also increases coder productivity and code capture, and enables flexible and scalable coding to increase accuracy, efficiency, productivity, and flexibility.
Indeed, these benefits are why we are seeing a rapid uptick in adoption of CAC on the professional side of the healthcare house. Once reserved primarily for facility-based coding, today’s CAC solutions are stepping up to fill the void created by the chronic coder shortage that is impacting both facility and professional fee coding while helping to maximize the performance of any healthcare organization’s coding operations with improved throughput and quality – increasing coder productivity by 25%-45% and decreasing Discharged Not Finally Coded (DNFC) by between one and three days.
Much of the credit for CAC’s substantial impact on coding and the revenue cycle goes to the integration of advanced technologies including artificial intelligence (AI), machine learning (ML), natural language processing (NLP), and predictive analytics. Not only have these technology tools enabled today’s CAC to make the coding process easier and more accurate and efficient, but they have also set the stage for transforming coding into a fully autonomous process in which coders assume the role of validators or auditors.
Technology-Enabled Coding Today
It was the transition to ICD-10-CM/PCS in 2015 that drove many of the advances that brought us to today’s CAC solutions. The vastly increased number of codes and completely new classification system required system upgrades and enhancements to ensure ICD-10 compliance. The more granular code set also exposed weaknesses in documentation processes that needed to be addressed before they impacted patient care, reimbursements, and quality metrics under value-based care.
These realities stepped up demand for technology-enabled tools that would make the coding process easier and more efficient while also ensuring compliant documentation. The result was the enhancement of CAC software with NLP capabilities that allowed the technology to electronically review notes within the EHR and apply system logic and standard coding rules to propose and group diagnostic related group (DRG) codes based on the presence of diagnostic words and/or phrases. It is a level of automation that makes it possible to achieve higher coder productivity without an increase in staffing levels.
A growing number of providers and coding outsource vendors are leveraging the cloud to integrate CAC, CDI, and auditing tools into a single cohesive platform. Doing so allows coders, providers, CDI staff, and auditors to collaborate across a shared workflow that follows the patient throughout their encounter. Importantly, integrated cloud-based CAC and auditing platforms enable provider organizations to hone coding and charge capture processes to reduce denials, accelerate payments, and capture correct revenue.
The addition of NLP and AI technologies to the mix enhances these benefits by taking some of the workload and decision-making burden off coders. When these tools are embedded into encoder and CAC software to suggest the most likely codes based on clinical indicators, it allows coders to focus on validating or adjusting recommendations based on their review of appropriate chart elements. The resulting coder-CDI collaboration also generates a higher degree of quality, accuracy and, subsequently, reimbursement. Coder and CDI staffs become more efficient and able to focus more on complex cases while the software manages the routine cases and more easily automated tasks.
Today’s advanced CAC solutions also allow coders to leverage the power of data analytics to improve productivity, accuracy, and revenue cycle outcomes. For example, the coding team can tell if optimal productivity has been achieved by monitoring everything from begin/end times, average number of charts coded per hour, percentage of charts exceeding standard minutes to code, case assignments, how many systems are accessed per case, frequency of and turnaround times for physician queries, and the volume of coding and non-coding tasks assigned to each coder.
Analytics can also be a powerful tool in reducing claim denials due to coding and documentation miscues, which are behind an average annual loss of $5 million for hospitals and write-offs of up to 5% of a physician practice’s net patient revenue. Identifying denial trends is critical as rates are on the rise, having already increased more than 20% over the past five years to about 10% for hospitals and approximately 20% for practices.
Predictive analytics can help healthcare organizations reduce costs and increase revenues by providing insights into performance gaps and coding/documentation trends that are driving down reimbursements and driving up denials. Predictive analytics leverages historical data, AI algorithms, ML, and NLP to identify denial trends and documentation and coding miscues that lead to over- and under-coding.
On the CAC Horizon
Advances in AI- and NLP-enabled coding technologies also allow healthcare organizations to address improper documentation and coding that currently cost the U.S. healthcare system about $54 billion annually. For example, adopting a single path coding model that merges professional and facility coding into a single workflow can significantly boost productivity while increasing clean claims and reducing denials.
When predictive analytics is added to the mix it enables another emerging solution – “coding alert” technology that works in the background of CAC software, scanning documentation against a database of procedures and terminologies, alerting coders to possible opportunities, and pointing back to where in the record it found the information. This kind of predictive insight helps reduce revenue losses due to missed procedure coding. When coupled with AI and ML to boost accuracy over time, coding alerts can also accelerate the overall coding process and advance automation beyond routine cases.
Finally, as the accuracy of automated coding technology expands, it brings the process closer to full autonomy. Pushing CAC across the last mile to true autonomous coding – a fully automated solution that rapidly and accurately codes charts without any human intervention – requires technology capable of “understanding” unstructured clinical notes. This is where coupling Clinical Language Understanding (CLU) with AI/ML and NLP comes into play.
CLU analyzes the free text within clinical documentation and extracts appropriate data for use in a variety of healthcare applications, including coding. It draws upon clinical knowledge and computational linguistics to create a digital narrative of the physician’s documentation. It then applies this understanding to determine what within the documentation is relevant and which codes are most appropriate to assign to the case.
Autonomous coding technology also understands what it does not know and flags those charts for human review. As such, it promises to significantly accelerate the end-to-end coding process, completing charts in seconds and the full process in minutes, while pushing accuracy levels to near perfect and generating astronomical productivity gains – a 700% increase in one pilot program. It also has the potential to enhance and accelerate the overall revenue cycle by eliminating missing reimbursement opportunities, backlogs, delays, and claims errors that plague human-centered coding processes, all while elevating coders by transitioning them into the role of auditor.
CAC has come a long way in the years since the transition to ICD-10. Powered by AI, ML, NLP, and CLU, today’s coding technologies are evolving along with the healthcare industry’s need to accelerate the coding process, improve accuracy, boost productivity despite ongoing coder shortages, and enhance the revenue cycle by reducing denials and audit risk.
As CAC technologies merge into their future state, they will continue driving improvements to the financial stability of provider organizations and forming a solid foundation for success under current and future value-based models of care.
About the author: Suhas Nair, a product manager for AGS, has more than 15 years of experience in healthcare technology. Suhas has delivered several SaaS products from concept to market. Nair combines his passion for AI and knowledge-infused learning with user experience design to create products that simplify and optimize healthcare processes and outcomes.