Better AI + Correct Provider Guidance = Fewer Denials 

The fewer healthcare claims that are denied, the better. Could there be a more agreeable sentiment in healthcare? Nearly 15% of all claims submitted to private payers are initially denied. In 2023, insurers of qualified health plans sold on Healthcare.gov denied 19% of in-network claims. 

What can be done to start moving that needle in the right direction? Honestly, we need better AI. Hear me out! 

Few things are more personal than one’s health, so when a computer algorithm run by artificial intelligence denies care, it is maddening for both patients and providers.  Humans want other humans to make decisions about claims, but that’s rarely how those decisions are made anymore. In a perfect world, denials wouldn’t exist, but they are necessary in the current U.S. healthcare system. Insurance fraud and excessive claims would run rampant if payers covered every claim. There must be a balance between what payers cover and what they don’t. With these elevated denial rates, there’s an obvious imbalance.    

How do we tip the scale to lower the rate of denials? While AI has been the initial decision-maker for years, it must become better trained. And to do that, it needs help.

AI in denials:

Over the past several years, there have been massive investments in automating the claims and review process in the health insurance industry. The technology has already led to improvements for patients and providers.  

• Fewer tedious tasks: Adding AI to the claims process allows clinicians to spend more time caring for patients through peer-to-peer review instead of reading every claim line-by-line.  It frees up time to review complicated claims that may need that additional human touch. 
• Faster claims: Instead of waiting days or even weeks to learn a claim status, AI can make a decision on a claim nearly instantaneously.  If a claim is denied, AI models can provide immediate feedback, allowing providers to resubmit it for approval. While 15% of claims are initially denied, 54% of those denied claims are eventually paid. If the denial stems from something missing in the claim, the faster you know that, the faster you can correct it and get paid.
• Better revenue cycle management: Time is money. When a claims process is slow, providers may not be paid on time. Speeding up the process with AI allows providers to better understand when they will be paid.

Improving AI for healthcare denials

Of course, you can’t just throw AI at a problem and expect it to be immediately fixed. It takes time. It takes work to hone an AI model to be as close to perfect as possible. Think back to when generative AI models first produced images. People barely looked like people  — hands would sometimes have eight fingers, and torsos might be as long as the person’s legs. But with years of data and human adjustments, these models learned to create photo-realistic images. AI for healthcare claims, in some instances, is at the goofy-image stage of development  — it needs more time and data. Fortunately, the U.S. healthcare industry has mountains of data. What these models require is more human correction, and no one is better suited for assisting AI models than the clinicians who deal with them every day.

When clinicians insert their own insurance claim knowledge into these models, it improves. Information like why they made their care decisions, such as consideration of a patient’s medical history or evolving clinical guidelines. The more an AI model understands the decision-making process, the better the AI becomes. This human fine-tuning of these AI models will take healthcare claims AI from eight-fingered pictures to a photo you’d swear is real. 

Bridging the payer-provider denials gap

Payers are working to keep the healthcare industry’s finances afloat. Providers are working to treat patients. Those interests don’t always align. Hence, the reason for denials. So, even with the increased use of AI in claims, there are ways for providers to stay one step ahead of the process.

• Predictive analytics: Healthcare systems should create or incorporate their own predictive AI models to better understand and prepare for a payer’s claims model. Think of this like laminating a treasured family recipe. It’s an extra step a provider can take before submitting a claim to protect it from being denied.
• Education: With predictive analytics, more providers will better understand the playbook of a payer’s claim model and what needs to be done to ensure approvals. If a clinician continues to have the same claims denied, the healthcare organization must find a way to reach and educate them on how to decrease denial rates.

Artificial intelligence will not eliminate denials, nor should it. However, AI can reduce them and continue to speed up the approval process. As the technology advances, becomes more prevalent, and more successful, trust will follow. Instead of looking line by line at a claim to ensure it’s approved, clinicians can focus more on their patients. Patients will worry less about a denial and put their energy toward healing. AI is here to stay, and as long as experts with the best intentions guide it on the right path, healthcare will improve, and denial rates will fall.

Photo: utah778, Getty Images

This post appears through the MedCity Influencers program. Anyone can publish their perspective on business and innovation in healthcare on MedCity News through MedCity Influencers. Click here to find out how.