Integrating artificial intelligence (AI) and automated workflows have significant potential to improve health care operations, particularly in revenue-cycle management (RCM). And with third-party payer denials and the rising cost of collections, providers increasingly are exploring solutions.
About 46% of hospitals and health systems now use AI in their RCM operations, according to an AKASA/Healthcare Financial Management Association (HFMA) Pulse Survey.
This trend is part of a broader movement toward automation in the health care sector, with 74% of hospitals implementing some form of revenue-cycle automation, which includes AI and robotic process automation (RPA). However, use of AI for RCM is limited to specific functions and health care operations.
Health care RCM is a prime area for innovation and disruption, with AI being a promising solution, according to a 2023 McKinsey & Company report. By streamlining these tasks, technology can reduce administrative burdens and expenses while enhancing efficiency and productivity.
For example, the McKinsey report states that call centers already have increased their productivity by 15% to 30% using generative AI. Some health care organizations have adopted technologies such as RPA, natural language processing (NLP) and, more recently, AI to achieve these goals.
Use Cases for Generative AI
Practical applications for generative AI are emerging, such as generating appeal letters for claim denials and handling prior authorizations. However, more significant applications, like improving front-end processes and data validation, are not yet ready for widespread use.
Generative AI can prevent avoidable errors by analyzing extensive documentation to identify missing information or potential mistakes, thus optimizing coding and other processes, proponents argue. It also can enhance communications within the revenue cycle, aiding in staff training and improving interactions with payers and patients. Here are some key applications in use today, according to a 2024 HIMSS report:
Automated Coding and Billing
- AI-driven NLP: These systems automatically assign billing codes from clinical documentation, reducing manual effort and errors.
- Claim scrubbing: AI identifies and corrects claim errors before submission, reducing denials.
Predictive Analytics for Denial Management
- AI predicts likely denials and their causes, allowing proactive issue resolution.
- Machine learning models analyze denial patterns to implement corrective actions.
Revenue Forecasting and Financial Planning
- AI-powered analytics provide accurate revenue forecasts, aiding in budget planning and resource allocation.
- AI simulates financial scenarios for informed decision-making.
Patient Payment Optimization
- AI personalizes payment plans based on patients’ financial situations.
- Chatbots remind patients of payments and facilitate billing queries.
Enhanced Data Security and Compliance
- AI detects and prevents fraudulent activities, ensuring data security.
- AI updates coding standards and guidelines for compliance.
Operational Efficiency
- RPA automates repetitive tasks, freeing staff for complex duties.
- AI optimizes scheduling and resource allocation.
3 AI Applications Improving RCM
Many hospitals and health systems already are harnessing the power of AI to enhance the effectiveness of RCM, leading to better financial outcomes and operational efficiency. Some examples from an October 2023 HFMA report include:
1 | Auburn (New York) Community Hospital
Auburn (New York) Community Hospital, an independent 99-bed rural access hospital, leverages RPA, NLP and machine learning in its revenue-cycle management. The hospital embarked on this path nearly a decade ago. Since then, it has experienced a 50% reduction in discharged-not-final-billed cases, a more than 40% increase in coder productivity and a 4.6% rise in case mix index, which the hospital credits to AI.
2 | Banner Health
Banner Health, a system with locations in California, Arizona and Colorado, has automated a significant portion of its insurance coverage discovery by utilizing a service that identifies each patient's coverage, coupled with an AI bot that integrates this information into the patient’s account across various financial systems. Another bot manages requests from insurance companies for additional information. Additionally, the health system employs a bot to automatically generate appeal letters based on specific denial codes. They also have developed a predictive model to determine whether a write-off is justified, based on particular denial codes and the low probability of payment.
3 | A Fresno, California-based Community Health Care Network
A Fresno, California-based community health care network uses an AI tool to review claims before submission and flags those likely to be denied based on its historical payment data and payer adjudication rules. The tool proactively addresses two types of denials: lack of prior authorization and services not covered. Since deploying the tool, the health system has experienced a 22% decrease in prior-authorization denials by commercial payers and an 18% decrease in denials for services not covered, all without hiring additional RCM staff. The health system estimates that it saves 30-35 hours per week by reducing the need to write as many back-end appeals.
Key Takeaways for Leveraging AI’s Potential in RCM
1 | Optimize staff time.
Generative AI can reduce health systems’ reliance on resource-intensive processes that are often understaffed or staffed with inadequately trained personnel, according to the 2023 McKinsey report. Early in the patient journey, AI could help identify duplicate patient records, automate eligibility determination based on payer policies and contracts, coordinate prior authorizations from health insurance companies and suggest solutions to address any administrative gaps identified.
2 | Improve accuracy.
In the midcycle, AI can enhance clinical documentation accuracy and reduce the time clinicians spend on recordkeeping, experts note. At the end of the patient journey, generative AI can assist accounts receivable with automated follow-ups and create fact-based appeals to health insurers using historical insurer performance, policy manuals and contracted terms.
3 | Assess risk factors.
New technologies like generative AI come with risks. Health systems should invest in mitigating these risks by establishing guardrails in data structuring to minimize bias or inequitable impacts on different populations, and by having humans validate computer-generated outputs to prevent closed-loop automation.
Generative AI faces a long journey ahead in health care, with some experts forecasting significant adoption within two to five years. Initially, it will address simpler tasks like prior authorizations and appeal letters. Over time, generative AI will tackle more complex aspects of the revenue cycle, potentially transforming the field.
Learn More
A new AHA Center for Health Innovation Leadership Scan episode, “Revolutionizing Revenue Cycle Management Efficiency with Artificial Intelligence” will provide insights on how AI solutions can help hospitals and health systems overcome staffing challenges in critical revenue-cycle management areas to respond more efficiently to payer denials. Register now for the June 27 program that runs from 1 to 2 p.m. ET.
