The data released earlier this year by the non-profit ProPublica has unveiled a concerning trend in the health insurance industry, shedding light on the increasing number of denied claims and the utilization of Artificial Intelligence (AI) by major payors. In 2021 alone, insurance companies nationwide denied over 49 million claims. What’s particularly startling is the minuscule fraction of appeals – a mere 0.2 percent – lodged by patients against these denials. This low rate has emboldened payors like Cigna and UnitedHealthcare to adopt stricter auditing practices, leading to a spike in the rejection of medically necessary services. The concerning aspect is the alleged reliance on AI programs by these companies, such as Cigna’s PXDX algorithm and UnitedHealthcare’s naviHealth platform, to expedite claim denials without thorough medical reviews.
Cigna’s PXDX Algorithm and Batch Denials
Cigna finds itself entangled in a class-action lawsuit filed in the U.S. District Court for the Eastern District of California. The lawsuit accuses Cigna of utilizing the PXDX AI algorithm to automatically screen and deny hundreds of thousands of claims without individual physician reviews. According to the lawsuit, Cigna’s algorithm reportedly rejected claims for procedures deemed medically necessary by physicians, such as tests for Vitamin D deficiency and transvaginal ultrasound cancer screenings.
Although Cigna denied these allegations, stating that its technology verifies the accuracy of submitted procedure codes based on publicly available coverage policies to expedite reimbursement to physicians, the lawsuit contends that relying on PXDX allows Cigna to instantly reject claims without physicians reviewing patient files, leaving many without coverage and facing unexpected bills.
As part of their aggressive audit and recover strategy, payors are quick to allege providers committed fraud for being too productive and billing an “impossible day”. Yet, when it comes to claims denials, which are supposed to be claim-by-claim review of medical records, ProPublica’s investigative reporting found that one Cigna medical director denied 60,000 claims in a single month. The crux of the lawsuit is that state laws and insurance regulations require a careful review of claims, so the speedy batch denials performed by the PXDX program makes it impossible for claims to be reviewed individually.
The plaintiffs argue that many medical necessary services were denied because of the PXDX program, which Cigna was aware could happen but ignored due to the low appeal rate.
Cigna has defended the PXDX program, and most concerningly, highlighted that its program is a standard of review like what other payors use. ProPublica’s reporting focused on Cigna, but they have confirmed that they are looking into its more prevalent use across many other payors.
Cigna’s lawsuit was the first domino to fall as they are not an exception to the rule, but an alarming indication of what is happening behind the scenes.
UnitedHealthcare’s naviHealth Platform and Denial of Post-Acute Care
UnitedHealthcare faces a potential class-action lawsuit following allegations that it employed an algorithm through the naviHealth platform to deny claims for post-acute care services for Medicare Advantage beneficiaries. The lawsuit, filed by the families of deceased MA members, claims that the insurer illegally used technology with a reported 90% error rate to deny necessary care to seniors. Despite the high error rate, the lawsuit alleges that UnitedHealth continued deploying this technology, banking on the low appeal rates (less than 1%) among members who were denied care. The plaintiffs argue that the insurer leverages patients’ conditions, lack of resources, and impaired states to dissuade appeals against the purportedly erroneous AI-driven decisions, leading to potentially detrimental consequences for seniors requiring essential care.
The lawsuit against UnitedHealthcare, accusing it of employing the naviHealth platform to deny post-acute care services, coincides with an investigative report from Stat, shedding light on the internal pressures faced by UnitedHealth employees to utilize naviHealth for cost-cutting purposes.
In response to the lawsuit, UnitedHealthcare vehemently denied any merit to the claims and stood by its program. The lawsuit contends that the naviHealth prediction tool imposes rigid and unrealistic predictions for recovery, overriding physicians’ assessments. It further accuses UnitedHealthcare of significantly increasing post-acute care claims denials through this AI-driven model, denying necessary care to seniors and gaining financial benefits while neglecting promised care.
Current State of AI Tools in Auditing
The lawsuits against Cigna and UnitedHealtchare have emerged amidst heightened scrutiny regarding payors’ potential utilization of AI and algorithms to deny member claims. What should concern providers, aside from the fact these programs appear to be wildly inaccurate, is that when caught red handed, both Cigna and UniteHealthcare have defended their use of these flawed AI programs and continue to use them.
It is a certainty that other major payors are either currently using some form of AI tool in their auditing, or contemplating doing so. The fact that a fraction of a percent of claims are actually appealed by patients leaves the burden on providers to timely appeal claims or face the unpleasant task of collecting from patients directly.
How Can Practices Prepare for the AI Era
There is no going back to the pre-AI assisted auditing process. Payors will likely continue to use these tools and refine and expand them going forward. For instance, in ProPublica’s investigation into Cigna, they uncovered internal documents that revealed areas where Cigna planned on expanding its PXDX program beyond its current scope. Part of the analysis was Cigna’s understanding that if they expanded, they only expected a 5% appeal rate from patients.
With the expected expansion of AI auditing tools, providers should begin to make preparations:
Self-Audits
Providers need to understand what codes they bill the most often and which procedures are the most profitable. When payors determine who to audit, they focus on outliers. Outliers are those practices that either make more money than their peers, or that bill a code outside of the expected distribution. For instance, if a practice bills too many high-level evaluation and management codes (“E&M”), they are likely to be audited.
At times, there are legitimate reasons for a provider to be an outlier. However, a provider must understand that being an outlier means that they will likely be audited at some point in the future. Audits are nothing to fear if your documentation is clear, concise, and meets the requirements of the code billed. Audits focus on weaknesses in documentation, so by understanding where your weaknesses lie, you can determine your level of exposure, and more importantly, enact immediate corrective action.
Appeal All Denials
Auditors are human, and everyone knows that humans are not perfect beings. When a human is auditing a claim, they may carelessly overlook a document or accidentally miss some verbiage that would have supported the code used. What we are learning now is that in an effort to be faster, not more accurate, major payors are using AI as part of their auditing protocol. Our firm has overturned 100% error rates and reduced them to 0% after conducting a rigorous independent review of both the documentation and the authority cited by the payors in their overpayment demand.
It is crucial that no matter how bleak the findings may seem, or how strongly a demand letter is worded, that providers appeal unfavorable findings and retain healthcare counsel to fight the demand from a legal, coding, and/or statistical angle.
Audit Response Protocol
Every practice should have an “audit response protocol” that they can turn to if they receive notice of an audit or an overpayment demand. The protocol should include the name of your healthcare attorney who can open a dialogue with the payor and advise providers and staff as to critical deadlines that must be met. Records should be gathered but not amended in any way to avoid allegations of submitting altered charts. The protocol should also designate someone at the practice as the contact person who will ensure that demand letters are received and addressed promptly, not ignored, or misplaced. Many times, audits fail from a lack of preparation or a cavalier attitude towards the audit process.
Conclusion
The AI era is still in its infancy by many measures, but it is being beta tested in real time by payors with providers across the nation. Some payors are now facing class action litigation for their misuse and overreliance on flawed AI programs, yet they continue to use those programs due to the sheer lack of appeals.
If your practice is facing a pattern of consistent denials for services or procedures that were historically paid without issue, or facing consistent denials from just one payor while others continue to reimburse you without issue, you may be the target of an algorithm or AI audit tool. Regardless of the reason for the denial, it is crucial to appeal denials.
Frier Levitt’s attorneys regularly handle fraud, waste, and abuse investigations, subpoenas, audits, and overpayment demands from government and private payors. If you are under investigation, receive an audit letter, or are experiencing frequent denials for a specific service, contact Frier Levitt.
