AI-Enabled Clinical Claim Reviews: Improving Accuracy, Speed, and Trust 

Artificial intelligence (AI) is designed to make life better for humans, but many worry it will have the opposite effect. For example, a recent article by Pew Research cites that six-in-ten Americans say they would like more control over how AI is used. It’s an understandable reaction given the depictions of AI in popular culture. Many entertaining — if unrealistic — storylines have centered around sentient computers wreaking havoc while malicious robots rule the world.

It seems that movie directors have yet to seize upon an exciting and compelling scenario: using AI to review Medicaid claims for billing errors.

From Fiction to Fact 

AI has already found its way into nearly every industry, including Medicaid. Far from displacing humans, AI is actually benefiting them in multiple ways. For one thing, AI is working alongside humans to make their jobs more efficient. And rather than creating friction, AI can be instrumental in fostering better relationships through increased transparency.

Let’s look at the area of cost containment, where Medicaid programs are charged with being good stewards of public funds. Robust payment integrity initiatives are vital in combating fraud, waste and abuse. Payment integrity programs ensure that claims are medically necessary and correctly submitted, and this is where clinical claim reviews come in.

At their most basic, clinical claim reviews identify claims that were billed or coded in error. The reviews can be conducted in a pre-pay or post-pay environment, and a final determination is performed by an expert clinician.

With the introduction of automation, clinical claim reviews take payment integrity to the next level by adding predictive analytics. Incoming data is analyzed by combining machine learning (ML) technology with historical data for improved accuracy.When combined with  ML,  data can be used to create algorithms that easily identify the claims most likely to have errors. This reduces the administrative burden on providers by only reviewing claims with a high probability of inaccurate coding or billing.

Now, here’s where science fiction becomes science fact. With the use of ML and AI, algorithms constantly improve and get more efficient at targeting inaccurate claims. In other words, they not only help to speed the process, but they also continually work to make clinical claim review solutions become even smarter.  A sophisticated AI Reviewer enables intelligent extraction of complex documents to pinpoint the data that clinicians need much faster for clinical reviews, clinical programs and authorization workflows, with an accuracy of over 98%+. In addition, AI can be used to read payment policy documents and medical spend reports to help with ideation of new concepts for review. 

The Human Connection 

Unlike movie storylines, however, humans are not eliminated in the claim review scenario. In fact, they play extremely important roles.

First, the fine-tuning of algorithms is not accomplished by AI alone. It is actually driven by a synergy between AI and the collective knowledge of data scientists and clinical experts. Without human input, AI can only go so far.

The human touch is preserved as clinical experts step in to personally review records and validate results. These experts can be physicians, nurses, certified coders, dentists or behavioral health professionals, all of whom have the experience and insight necessary to determine whether fraud, waste and abuse are at play.

And here’s one of the best benefits. AI can actually be used to improve human experiences.  The new AI tools are used to help clinicians be decision makers instead of information finders.  The tools present information to the reviewer for validation with references on where the supporting information can be found in the medical record.  This leads to higher quality results and allows clinician reviewers to focus on the most critical aspects of the review without the busy work. Used correctly, AI can speed the payment process and actually reduce provider abrasion, saving time, money and frustration. 

Advancing the Clinical Review Process

When claims can be accurately assessed earlier in the review cycle, a significant barrier is removed. The process not only accelerates, but the experience becomes smoother for both payers and providers.

Predictive evaluation models can look at combinations of elements within a claim, such as length of stay, diagnosis and procedure codes, discharge status, member age, and provider-specific billing patterns. These insights help determine which claims are most likely to require deeper review based on historical outcomes. Using a calibrated tiering approach, claims with the strongest indicators of overpayment potential can be sent directly to clinical review without automatically requiring medical records. This shortens the time to determination, yet all findings are still finalized by a clinician before providers are notified. Claims with more moderate probability indicators follow the traditional process with medical record request and review, maintaining full safeguards for accuracy and fairness.

Embracing the Future 

AI can be elevated as an integral part of payment integrity, increasing innovation and helping states make the best use of healthcare dollars. With AI, clinical claim reviews become an even more valuable part of a digital health ecosystem, using smart technology to contain costs and drive innovation in the payment process.

The future of clinical claim reviews is here — and AI is enhancing the process.