By Amy Amick

At a time of rising costs, labor shortages, and operating margins still below historical levels, many hospitals are simply not in a position to absorb any loss of revenue. Yet that’s exactly what is happening when it comes to managing the influx of denied claims as part of the revenue cycle management (RCM) workflow.

Hospitals are grappling with a surge in claims denials, marked by an increase in the average dollar value per denial, extended resolution times, and reduced yield per claim across all payer types. A Kaiser Family Foundation study on Affordable Care Act (ACA) health plans found that, on average, insurance companies initially deny 17% of claims, even for in-network care. Prior authorization denials have also risen, as identified in commercially insured Medicare Advantage plans.

For hospitals still struggling to dig out from the financial hole left by COVID-19, the timing couldn’t be worse. According to an analysis from Crowe, the rate of claims denials by payers is frustratingly high, rising to 11% in 2022 from 10.2% in 2021. For the average health system, that’s 110,000 unpaid claims that require additional time and expertise to fight to overturn. The situation is even worse with aging AR. The same report found the proportion of hospital accounts receivable that have aged beyond 90 days has been trending upward and was at 37% as of August 2022. Compounding these challenges is a critical shortage of talent and expertise needed to navigate the complex denials landscape. According to an Experian Health survey, 100% of revenue cycle leaders recognize that the widespread healthcare workforce shortage significantly impedes payment collections.

Is AI-powered denials management a game-changer?
AI holds the potential to revolutionize the denials landscape, and its impact is already evident. Through the use of artificial intelligence, healthcare providers can streamline claims processing, enhance coding accuracy, and extract essential information from medical records and payer contracts to tackle the root causes of denials.

While the expertise of seasoned revenue cycle professionals remains crucial for denials resolution and prevention, the integration of AI and automated workflows empowers these professionals to operate at peak efficiency and effectiveness resulting from the following:

Selection models
In the context of AI, selection models are designed to choose or classify specific items or entities based on certain criteria. These models are a type of machine learning that can help automate the prioritization of accounts to increase yield.

Workflow acceleration
AI can also be used to simplify workflows and accelerate time to action by automating the extraction of essential information from claims, medical records, contracts, and guidelines, and then applying that information to accelerate activities. For instance, AI can generate draft appeal letters, minimizing manual effort, reducing errors, and improving the overall workflow for improved revenue outcomes.

Improved price accuracy
AI can be invaluable to RCM teams to unlock the value of unstructured data in medical records, managed care contracts, and other sources by identifying reimbursement rate discrepancies and ensuring contract compliance. This helps prevent denials and underpayments and arms providers with critical information which can help ensure accurate reimbursement in accordance with negotiated agreements.

Predictive trend analysis
As denial rates continue to climb, payers are also making changes to coding policies, creating more complexity around submitting claims and posing ever more challenges to short-handed provider organizations. AI can help by identifying and predicting trends in denials and payments to enable RCM teams to make operational updates to avoid denials, and when denials do occur, add the ability to predict which denials are more or less likely to be overturned so that RCM teams can focus their limited resources more effectively.

Considerations for AI-powered denials management implementation
When incorporating AI and automation into RCM, healthcare organizations will face a myriad of technical, data, talent, and operational considerations.

Building effective AI models is tough. It requires an enormous amount of quality data that can be used to train and mature models. It requires hard-to-find data scientists with sophisticated talent who understand how to leverage data in concert with isolating meaningful use cases, such that machine learning can establish meaningful learnings. It requires underlying platforms and engineering strength that allow models to be effectively deployed. It requires a sophisticated understanding of evolving standards and regulations to ensure ethical and permitted usage. Leveraging industry-leading RCM innovators who have already proven success in this emerging field can accelerate a health system's path to stronger financial health.

Large Language Models (LLMs) or foundational models represent a groundbreaking innovation, presenting both tremendous potential and considerable uncertainty. The regulatory landscape is catching up with the technology, and the rapid evolution of the technology itself introduces uncertainties. Collaborating with strategic partners skilled in this becomes a vital strategy. Notably, major players like Google and other cloud providers are developing healthcare-specific LLM platforms, contributing to risk mitigation in this evolving landscape.
Additionally, leaning on industry groups such as state hospital associations as well as enlisting multiple healthcare organizations and technology companies to set up pilot projects are also good ideas. Provider organizations would also do well to start small, leveraging AI for their most challenging problems, such as patient collections.

Further, interoperability is critical, and as with any innovation project at the start. The underlying data infrastructure must be capable of routing the right claim-level, clinical, contract, and operational data to the models, and then route the results into operational workflows. Implementation of these models also requires a degree of change management, training, and a product-driven mindset to ensure actual value for the business is generated.

When it comes to effectively managing claim denials, it's undeniable that AI can be a game-changer. With the ability to enhance RCM efficiencies and alleviate revenue pressures faced by healthcare providers, AI has transformative potential. Payers’ use of AI has already impacted the volume of denials and healthcare leaders can’t afford to wait any longer to use AI to benefit their revenue cycle.

About the author: Amy Amick is the chief executive officer of Aspirion, a company that helps healthcare providers maximize their hospital revenue recovery. At Aspirion, she leads a team comprising over 100 attorneys, over 30 clinicians, and other healthcare professionals in serving more than 1,000 clients in 45 states. She also serves as an independent director on the boards of urgent-care software company Experity and healthcare-revenue-cycle-management firm Pendrick Healthcare.