Artificial intelligence (AI) is rapidly entering medicine. Discussions have focused on how physicians may use AI: ambient documentation, clinical decision support, imaging interpretation and remote monitoring.
However, there is another use of AI that physicians may encounter without ever realizing it. Health insurers and the companies that work for them are increasingly using sophisticated technology to determine whether physician claims should be paid, changed, reviewed or audited.
For allergy practices, this may ultimately be one of the most important applications of AI to understand. Insurers do not necessarily develop all of these systems themselves. They may contract with companies specializing in what the insurance industry calls payment integrity.
One example is Lyric, formerly known as ClaimsXten. ClaimsXten was originally part of Change Healthcare. When UnitedHealth Group acquired Change Healthcare, ClaimsXten was divested and sold to private-equity firm TPG for approximately $2.2 billion in 2022. It subsequently became Lyric.
Today, Lyric provides claims editing and payment integrity technology to health insurers. Reuters recently reported that UnitedHealth, CVS and Humana are among the insurers that use Lyric to identify and prevent inaccurate medical claims payments.
TPG is now reportedly considering selling Lyric in a transaction that could value the company at approximately $5 billion. These numbers illustrate how valuable payment integrity technology has become.
What does payment integrity mean?
The term sounds benign. Everyone agrees that claims should be accurate and that true fraud and billing errors should be identified. However, payment integrity technology can do much more than look for obvious mistakes.
Systems can compare claims against coding rules, payer policies and clinical criteria. They can identify claims for additional review, apply coding edits, analyze patterns across enormous claims databases and potentially identify claims for prepayment or postpayment scrutiny.
Lyric describes its ClaimsXten heritage as more than 35 years of prepayment claims editing. The company now markets a broader payment accuracy platform incorporating emerging technology, including AI. Lyric states that AI supports rather than replaces human expertise and that its systems incorporate clinical validation and client-controlled implementation.
That distinction is important. The concern is not that every claim decision is being made autonomously by an algorithm. The larger issue is that increasingly powerful technology can determine which claims are questioned and what information a human reviewer sees.
Why allergists should care
Consider some situations already familiar to allergy practices:
- A claim is unexpectedly downcoded.
- A payer requests large numbers of medical records.
- An immunotherapy claim that was previously paid is reviewed months later.
- A practice receives a recoupment request.
Similar coding edits appear across seemingly unrelated insurers. None of those situations prove that AI or a particular payment integrity vendor was responsible. We should not assume that it was. But the growing role of common payment integrity vendors raises an important question: When similar claims problems occur across multiple insurers, are the insurers independently reaching the same conclusion, or are they sometimes using the same underlying technology, policies or vendors? Physicians frequently cannot tell.
The transparency problem
When a physician receives a denial, the letter usually identifies the insurer. It may not identify the technology or outside company involved in generating the decision, which creates a significant transparency problem. These questions will become increasingly important as AI becomes embedded in payer operations.
- If an algorithm flags a claim for downcoding, who developed the rule?
- If an outside company recommends that a previously paid claim be recouped, who is responsible for the decision?
- Was the claim reviewed by a physician in the same specialty?
- Can the practice challenge the underlying rule?
- If an algorithm is producing inappropriate results repeatedly, how would physicians, or regulators, recognize the pattern?
AI deserves scrutiny on both sides of health care
Physicians are appropriately being asked difficult questions about their own use of AI.
- How was the system validated?
- Is a physician reviewing its output?
- Who is responsible when it makes a mistake?
- Are patients adequately protected?
Insurers using AI to make decisions affecting patient care and physician payment should be expected to answer similarly rigorous questions.
The College will continue monitoring the use of AI, automated claims editing and payment integrity systems by health plans, particularly where these technologies affect allergy services.
AI may ultimately make health care more efficient. A practice that is not using AI to respond to an insurer’s claims analysis systems may be at a distinct disadvantage. Asking physicians and their staff to battle AI-driven bots one claim at a time may ultimately be a futile exercise. Efficiency cannot come at the expense of transparency, appropriate physician review or patients’ access to medically necessary care.
The Advocacy Council – ADVOCATING FOR ALLERGISTS AND THEIR PATIENTS.
