Why is UnitedHealthcare Denying Claims?

UnitedHealthcare, one of the largest health insurance providers in the United States, is facing allegations that it has been wrongfully denying claims using an artificial intelligence (AI) algorithm. Two recent lawsuits accuse the company of deploying technology to automatically reject claims for Medicare Advantage members’ post-acute care.

Lawsuits Allege AI Use to Deny Claims

In November 2022, a federal lawsuit was filed in Minnesota alleging UnitedHealthcare used an AI-powered tool called “nH Predict” to improperly deny coverage for extended stays at skilled nursing facilities for elderly Medicare Advantage members.

The lawsuit argues this violates the terms of members’ plans and Medicare guidelines. It states that by using this predictive algorithm, UnitedHealthcare increased denials for post-acute care services, knowing that few members would have the ability to appeal the AI-powered decisions.

A similar lawsuit was filed in California earlier in 2022 against UnitedHealthcare competitor Cigna, also accusing the insurer of using AI to systematically deny claims without proper manual review.

How Does the AI Tool Allegedly Work?

According to the lawsuit, nH Predict makes rigid predictions about how much care a patient “should” need during recovery based on past data. It then recommends denial of coverage when a patient’s proposed treatment plan exceeds the algorithm’s predetermined timeline.

The lawsuit argues this results in denying coverage against doctors’ orders, overriding physicians’ judgments of patients’ required care. It alleges the algorithm is deployed to cut costs by reducing approved days of coverage, not for clinically appropriate reasons.

Why Are Denials Increasing?

The Minnesota lawsuit states that by using nH Predict, UnitedHealthcare has significantly increased denials of coverage for post-acute rehabilitative services like skilled nursing facilities.

It argues the insurer knows most members will not have the knowledge or resources to appeal wrongful AI-powered denials. As a result, UnitedHealthcare can achieve cost savings through more rejected claims.

Government reports support allegations of increased denials. A 2022 federal audit found that Medicare Advantage plans improperly denied prior authorization requests for covered services in 75% of cases involving post-acute care.

Effects on Medicare Advantage Members

When insurers deny coverage for post-hospital care against medical advice, it can have serious consequences for Medicare Advantage members.

Being discharged prematurely can lead to readmission when complications arise. Out-of-pocket costs for nursing home care average $7,500 per month, which is unaffordable for most seniors without insurance coverage.

The lawsuit alleges UnitedHealthcare profits from wrongful denials while vulnerable, elderly members suffer harm. Critics argue this violates the purpose of insurance to cover medically necessary care.

Why Are AI Tools Controversial?

UnitedHealthcare is not alone in testing AI to automate processes like claims management. But use of algorithms to adjudicate healthcare claims remains controversial.

• Lack of transparency: Critics argue AI claim review processes are opaque “black boxes”, lacking transparency and accountability.

• Ethical issues: There are concerns AI could perpetuate biases or unfairly impact minorities and lower-income members.

• Accuracy concerns: Studies have found AI models trained on past claims data inherit problems and inaccuracies in that data. Skewed datasets can lead to flawed AI predictions.

• Lack of human oversight: Automated claim reviews may involve insufficient human supervision, controls, and quality assurance. Rapid AI claim denials limit individualized review.

Potential Solutions and Safeguards

In response to criticism of AI claims management, insurance experts recommend measures to improve accountability, fairness, and accuracy:

• Health plans should conduct regular algorithm audits for bias and inaccuracies. AI models need ongoing monitoring, not just one-time checks.

• AI coverage decision processes must have meaningful oversight and opportunity for timely case-by-case reviews. Rapid automated denials should face greater scrutiny.

• State regulators could require transparency for how health plans use AI, including making algorithm models and key data inputs public for independent evaluation.

• Claims decision processes should account for social determinants of health impacting patients, not merely analyze past care usage patterns.

• Health insurers should involve frontline clinicians in developing and assessing AI tools to improve reliability and medical appropriateness.

What lmpact Could the Lawsuits Have?

The UnitedHealthcare lawsuits highlight concerns about AI’s growing role in claims management. If found liable, the company could face substantial penalties and be forced to change its practices.

Beyond financial consequences, legal challenges can draw public and regulatory attention to AI oversight issues. However, some observers argue more explicit state and federal laws governing use of healthcare algorithms may be needed.

It remains to be seen how courts will view use of predictive analytics tools to adjudicate complex care decisions traditionally made by doctors and nurses. But the UnitedHealthcare litigation makes clear there are mounting legal risks for insurers deploying AI without sufficient accuracy, transparency, and accountability safeguards.

In Conclusion

Recent lawsuits targeting UnitedHealthcare’s algorithm-driven processes raise important questions about AI’s appropriate role in claims management:

• Can machine learning responsibly automate highly complex, individualized health coverage decisions?

• How can payers ensure AI tools properly reflect accepted standards of care and medical necessity?

• What safeguards and transparency should regulators require for AI claims review models?

UnitedHealthcare maintains its technology improves healthcare delivery and follows applicable laws. But the lawsuits exemplify rising concerns that AI claims management can improperly deny needed care if not thoughtfully designed, validated, and monitored.

As AI takes on increasing responsibilities in healthcare, these cases will likely pressure insurers using such tools to ensure they are deployed safely, fairly, and accountably. Robust human oversight, rigorous auditing processes, transparency, and proactive bias checks will be key to addressing growing legal, ethical, and medical risks posed by AI automation of crucial coverage decisions.