Using Artificial Intelligence to Access Healthcare through Insurance Appeals

Background and Purpose: Enrollment in Medicare Advantage (MA) has rapidly increased, yet recent government reports have underscored how MA insurance plans are increasingly using artificial intelligence (AI) tools to deny care even when requests meet Medicare coverage rules. Despite these denials, less than 1% of patients and clinics submit appeals, largely because the process is costly, time-consuming, and confusing. This study addresses a critical gap in social work science by examining the appeals process as a mechanism for facilitating healthcare access.

Methods: This mixed-methods study uses a concurrent design across nine healthcare clinics in California, Illinois, and North Carolina to compare experiences of MA and insurance denials in three states with varying levels of MA market penetration and state regulatory environments (before and after introduction to an AI tool to assist with insurance appeals). It incorporates surveys and interviews among claims administrators who process MA denials. It also includes data from focus groups among older adults who have experienced MA denials.

Results: Thematic analysis of qualitative data among claims administrators and older adults identified three primary barriers to the appeals process: (1) bureaucratic exhaustion in which denial letters were too vague and complex; (2) time exhaustion in which patients or clinics determined that the time to appeal outweighed potential coverage benefits; and (3) a hybrid of bureaucratic and time exhaustion in which patients and clinics determined that the time and effort needed to appeal during a specified appeal window was unrealistic. Quantitative analysis revealed that appeal rates for MA denials among these clinics increased after using the AI tool. Clinics using the AI tool also saw a higher success rate in overturning denials compared to the manual process, which was even higher for cases involving medical necessity.

Conclusions and Implications: While most research on AI and healthcare focuses on either potential risks or specific healthcare products, this research examines an AI tool to access healthcare but through an adjudicatory mechanism (insurance appeals process). It also innovatively provides an AI-solution to a problem created, in part, by AI. In line with the conference theme (Leading for Transformative Change) this research underscores how innovative AI tools can empower clinics and older adults to challenge systematic barriers to equitable care.