The importance of coded data has increased over the years requiring high quality output more now than ever before. Robust compliance plans with monthly quality review schedules are no longer a recommendation, but a necessity. Due to the costs associated with staffing and continuing education, along with the difficulty in finding and retaining strong coding talent, many organizations are leaning towards artificial intelligence (AI) to improve the quality and productivity of their medical coding and billing.
Computer assisted coding (CAC), the first AI type of solution to hit the healthcare industry, is widely recognized among health care professionals today. Using a natural language processing (NLP) engine, CAC software scans medical record documentation, identifies key terms and suggests codes that support the treatment or service provided during that specific encounter.
Computer-assisted solutions automate specific activities which makes it easier to process more records in a shorter period of time. CAC software does, to some extent, alleviate a portion of the burden organizations face with volume and workflow fluctuations as it improves productivity. However, each chart still requires human time and attention and CAC software does not solve for coder scale issues.
CAC brought numerous benefits with its implementation, yet many felt the overall outcome fell short of expectations. Although users saw some improvements with productivity and quality measures, system limitations, template build issues, the appearance of cloning and increased denials are a few of the challenges that resulted in organizations feeling dissatisfied with their product.
Many organizations have been able to reduce discharged not final billed (DNFB) days, missed charges and incomplete documentation by leveraging CAC software, but the end result is that the solution cannot stand on its own. It does not possess problem-solving or critical thinking skills, and it cannot replace the coder as the medical decision maker and final code approver.
To improve accuracy, an organization must ensure that their templates, interfaces, and algorithms are built for accurate and complete code sets allowing the software to assign codes to the highest level of specificity. The days of assigning unlisted diagnosis codes are gone and those still reporting them will likely see an increase in claim denial.
Autonomous coding, a new and innovative technology, is a fully automated AI solution that codes charts within seconds with zero human intervention and with high accuracy. Built on the foundation of a new Clinical Language Understanding (CLU) technology, autonomous coding is unique in being able to understand unstructured physician language. CLU is a combination of medical knowledge and computational linguistics applied to clinical language. It allows computers to understand the logical relationships between various linguistic elements in the medical record, and to construct a model describing the narrative of the physician’s documentation.
Although there are many benefits of implementing an AI product, the main goal in the medical billing and coding space is to automate processes that enhance workflow and improve overall revenue cycle performance. Key performance indicators (KPIs), such as productivity and coding quality, remain areas of heavy scrutiny. Autonomous coding technology identifies every clinical aspect of the medical record. It understands relevancy as well as which codes are the most accurate to assign, thereby offering consistently accurate, high quality medical coding. Codes are assigned only for those charts the software fully understands. Unhandled medical records are flagged for manual review.
Autonomous coding solutions offer a scalable coding service without compromising accuracy. In fact, productivity gains are colossal. Autonomous platforms offer an unlimited capacity to process medical encounters at a speed of less than five seconds per encounter. Fluctuations in chart volumes are handled by simply allocating more or less compute power. The improvements in quality and productivity, combined with the system’s ability to stand independent of human intervention, has enabled organizations to overcome coder shortages, streamline the billing process, and reduce cost.
Autonomous coding is also receiving a lot of attention on its audit performance and ability to recognize and quickly correct coding errors, without the need to send encounters back to coders for resolution. In addition to the recognition enhancements, autonomous coding solutions offer a clear audit trail that provides a full explanation behind each and every code assigned. This technological advancement results from the core CLU capability to understand clinical narrative, and allows for quick validation of the assigned codes and a solid foundation to support denials management and the appeals process.
Customers who implemented autonomous coding technology realized an improvement in accounts receivable days (AR), missed charges, and incomplete documentation.. Furthermore, codes are reported to the highest level of specificity which has proven successful in reducing the number of denials and improving the clean claim rate.
The Bottom Line
Artificial intelligence has proven to be a powerful resource in the medical coding and billing space. While both CAC and autonomous coding offer a multitude of benefits, CAC solutions, as their name implies, are limited to being a human assistant, requiring a significant amount of human intervention by way of documentation assessment, code review, and template development to assure successful outcomes. In addition to the elements mentioned above, some organizations have expressed challenges in overcoming cloning non conformances and increased denials. In comparison, Nym’s autonomous coding solution offers a robust coding platform with the ability to work independent of human involvement. It is an added benefit that this new type of technology comes with a fully explainable audit trail enabling organizations to provide a listing of codes assigned and the rationale behind each one. Processing claims at computer chip speeds, with an unlimited volume threshold, and without sacrificing quality, has pushed autonomous solutions to the forefront of industry specific AI platforms.