Health Insurers’ AI: Denials and the Push for Oversight

Algorithms used to deny or delay care spark calls for FDA regulation.

Artificial intelligence is increasingly used by health insurers to make coverage decisions, potentially leading to delayed or denied care. As algorithms play a larger role, the need for robust regulation and oversight grows ever more critical.

AI’s Role in Coverage Decisions

Unlike medical professionals who use AI to assist in diagnoses and treatments, insurers employ these algorithms to determine whether they will cover treatments and services recommended by physicians.

A common instance is prior authorization, where doctors must secure approval from insurance companies before providing care. Insurers often use algorithms to assess whether the requested care is “medically necessary” and should be covered.

These AI systems also influence the extent of care a patient can receive, such as the duration of hospital stay following surgery.

The Appeal Process and Its Limitations

When an insurer refuses to cover a treatment, patients have limited recourse. Appealing the decision can be lengthy, costly, and require expert assistance. Data show that only 26% of appealed claims are overturned.

Alternatively, patients may opt for a different treatment covered by their insurer or pay for the recommended treatment themselves, which is often financially unfeasible.

Concerns About Withholding Care

Insurers feed patient health records into coverage algorithms, comparing the data against medical standards to decide on claims. However, insurers have declined to disclose how these algorithms work, making it difficult to understand their practical operation.

Using AI to review coverage saves insurers time and resources by reducing the need for medical professionals to review each case. Additionally, insurers might save money if an AI system denies a valid claim and the appeal process extends for years, especially if the patient has a terminal illness.

This raises concerns that insurers might use algorithms to withhold care for expensive, long-term, or terminal health problems. As one reporter noted, “Many older adults who spent their lives paying into Medicare now face amputation or cancer and are forced to either pay for care themselves or go without.”

Research indicates that patients with chronic illnesses are more likely to be denied coverage and suffer as a result. Moreover, racial and ethnic minorities, as well as individuals identifying as LGBTQ+, are disproportionately affected by claim denials.

Moves Toward Regulation

Unlike medical algorithms, insurance AI tools are largely unregulated and do not undergo FDA review. Insurance companies often claim that their algorithms are trade secrets.

The Centers for Medicare and Medicaid Services (CMS) has announced that insurers in Medicare Advantage plans must base decisions on individual patient needs, not just generic criteria. However, these rules allow insurers to create their own standards without external validation.

Several states, including Colorado, Georgia, Florida, Maine, and Texas, have proposed laws to regulate insurance AI. However, many state laws suffer from similar weaknesses, leaving too much control to insurers and not requiring neutral expert reviews.

Even strong state laws may be insufficient because states generally cannot regulate Medicare or insurers operating across state lines.

The FDA’s Potential Role

Many health law experts believe that regulating health care coverage algorithms is essential. As **Jennifer Oliva**, a legal scholar, argues, the FDA is well-positioned to do so.

The FDA has medical experts capable of evaluating insurance algorithms before they affect coverage decisions. The agency already reviews numerous medical AI tools for safety and effectiveness. FDA oversight would provide a uniform national regulatory scheme, replacing a patchwork of state rules.

While some argue that the FDA’s power is limited, Congress could amend the definition of a medical device to include health insurance algorithms. In the interim, CMS and state governments could mandate independent testing of these algorithms for safety, accuracy, and fairness.

The movement to regulate the use of AI by health insurers in coverage determinations is underway, yet requires further momentum. Patients’ lives depend on it.