The Benefits of AI in the Revenue Cycle: Real World Scenarios and Use Cases
This article is the last in a three-part series on artificial intelligence (AI) in the health information (HI) revenue cycle field. Part 1 focused on what AI is and how it can benefit the healthcare revenue cycle. Part 2 examined what HI professionals need to consider when evaluating and planning for AI in the revenue cycle. The third and final article in the series will focus on actual use cases within various touch points in revenue cycle where AI has been successfully implemented to bring efficiencies and accelerate revenue cycle.
With a clearer understanding of artificial intelligence (AI) and how it can be leveraged to enhance the revenue cycle, here are four real-life use cases spanning several different points in the process. (Note: The authors have direct knowledge of these case studies.)
Moxi
ChristianaCare was the first health system in the Philadelphia region to deploy Moxi, a collaborative robot, or “cobot,” that assists hospitals by making deliveries and performing other non-clinical tasks.
ChristianaCare was one of 10 grant recipients from the American Nurses Foundation Reimagining Nursing (RN) Initiative. As noted by ChristianaCare, “Before Moxi can become fully part of the care team, it must learn to navigate the hospital and respond to the nurses’ needs. Using artificial intelligence, Moxi is mapping out Christiana Hospital through sensors and other machine-learning technology so that the cobots can ultimately navigate and work autonomously.”
Moxi is leveraged by the health information (HI) team to transport paper-based documents to and from the nursing floor, freeing up both clinical staff and HI administrative staff. This time-saving benefit of working with Moxi allows HI professionals and nurses to dedicate their time to other tasks, which is a revenue cycle win and a patient satisfier.
Dragon Ambient eXperience (DAX)
Northern Montana Health Care onboarded Nuance’s Dragon Ambient eXperience (DAX) in late 2021 at the request of a provider to increase documentation efficiency.
Increasing compliance regulations on documentation and higher patient volumes add administrative burden. DAX is an AI-powered, voice-enabled solution that automatically documents patient encounters accurately and efficiently at the point of care. This allows the provider to return to facing the patient during the encounter instead of being in front of the computer inputting data. Providers engage in natural conversation with patients and other family members.
The DAX mobile app securely captures the conversation accurately and efficiently, allowing providers to connect with patients without using explicit voice commands. The flow of conversation creates a more focused patient care environment. DAX converts patient encounter conversations into comprehensive clinical notes tailored to each specialty with built-in compliance standards to allow for accurate coding. The AI‑generated notes are reviewed by dedicated HI analysts for accuracy, omissions, and appropriateness before being delivered to the provider for signature in the electronic health record (EHR), which creates an AI learning loop for continuous improvement.
Providers also can customize where in their notes they want DAX to place the transcribed conversation.
Epic Emergency Department (ED) Facility Charge Calculator (FCC)
The Epic Emergency Department (ED) Facility Charge Calculator from Epic supports the streamlining and automation of ED facility level charging in an organization’s EHR.
According to Epic, the tool requires customization by the organization, but with focused time and dedication by members of IT, HI, and revenue integrity, it can become an autonomous coding solution. Using automated technology to read provider notes and apply coding rules allows hospitals to achieve non-human interventions for coding hospital evaluation and management levels. In a large health system in Pennsylvania, two community hospitals use the tool autonomously with some routine auditing. For three downtown hospitals in the system with typically higher acuity patients, there is little auditing for lower levels and moderate auditing for higher levels.
The tool has been in use for over five years and remains regularly customized. With changes in coding rules and guidelines, as well as medical innovation and technology, a team of IT, HI and revenue integrity leaders meets quarterly to ensure the tool is still functioning efficiently to capture appropriate codes for ED visits.
The Federal Communications Commission (FCC) supports efficiency and promotes accuracy in ED facility level code assignment allowing for providers to see acuity trends across all EDs in the health system in almost real time.
Computer-Assisted Coding (3M) 360/SmarterDx
Computer-assisted coding (CAC) is used to describe the technology that coding professionals use to assist in facilitating efficient and accurate coding. These platforms use AI such as natural language processing (NLP) and machine learning (ML) to evaluate the clinical documentation and “call out” opportunities for the end user. Some platforms connect to the EHR and have a seamless interface while others are external portals accessed on the web.
Application of guidelines, context, and scenarios still require a human brain. Two AI tools were implemented in the last year at a South-Central U.S. academic medical center and Level 1 trauma center. The first was one of the largest CAC tools that was upgraded and integrated with its EHR.
The second, more recent tool, called SmarterDx, is a separate portal in which AI identifies clinical documentation integrity (CDI) and coding opportunities by using clinical indicators and clinical documentation. Combining the two tools would allow more efficient, thorough, and accurate coding while not sacrificing initiatives in coding, CDI, and quality areas (e.g., patient safety indicators, hospital acquired conditions, mortality, etc.) Each tool brought a learning curve and a new perspective to look at a health record.
Here are some lessons learned from these case studies:
- The tools are only helpful if used correctly as intended. Not addressing the presented opportunities may result in missing code capture. Conversely, accepting every presented opportunity without verification or validation can cause over-coding. These tools rely on user interactions to learn and sharpen the outputs.
- The tools can’t detect variations in abbreviations and often struggle with strings of text with certain punctuation. The result is that these areas of documentation will not be presented to the coding professional, so reading between the lines is still necessary.
- Having open communication and regular feedback with the technology vendor offers a mechanism for growth and development of the application. Just as the machine learns from user interaction, vendors learn from their client partners.
- Regular training and re-training on new features, with follow-up audits, are essential to the overall success of AI-type tools.
Conclusion
AI can be leveraged by HI professionals in every aspect of the revenue cycle to accelerate cash, increase reimbursement, and reduce denials. From reduction of clerical burden by use of cobots to supporting coding professionals with accurate code assignment, AI is being woven into the fabric of our profession.
It is the responsibility of HI professionals to engage with these evolving technologies and partner with vendors to have a seat at the table to continue to evolve the field. AI is here to stay so HI professionals should embrace it and be at the forefront of leadership through these changes with their skills and expertise.
Tami Montroy, MS, RHIA, CHPS, is a director of central fee abstraction.
Glenda Rakes, MHIIM, RHIA, CHPS, is the HIM director/privacy officer at Northern Montana Health Care.
Kimberly Seery, RHIA, CCS, CDIP, CHDA, CPC, CRC, is the associate director of coding and data quality.
Monica Watson, RHIA, CPC, CCS, CCS-P, CPMA, CIC, CRC, is the senior director of hospital and professional coding at the University of Arkansas for Medical Sciences (UAMS).