The Use of AI in Utilization Management
Problem
Regulators have been unable to keep up with the use of AI in health insurance utilization management.
Over the last decade, health insurance plans have started to leverage artificial intelligence (AI) tools in utilization management (UM). UM is the process in which a health care benefit administrator determines whether to approve payment for provider-recommended medical care. State and federal regulators have struggled to keep up with the rapid pace of AI advancements and use of this technology to make important decisions about patient access to health care services.
The Consumer Representatives for Health to the National Association of Insurance Commissioners (NAIC) sought to better understand this uneven regulatory landscape.
Solution
NORC conducted a comprehensive literature review and interviewed key health care sector stakeholders.
NORC’s research team first examined available literature to understand the landscape of AI use in health insurance decision-making processes. We focused on prior authorization as a form of UM and preliminary efforts to regulate it. Next, to create a more holistic understanding of the implications of AI’s use in UM—as well the impact of the current regulatory approach on consumers—we spoke with stakeholders who represented different interests across the health care sector. They included health plan executives, providers, consumer advocates, policy experts, technical experts, and regulators. These interviews identified their perspectives and primary concerns about using AI to make decisions about coverage for patient care.
Result
Greater transparency of AI in UM is needed to prevent harm.
NORC’s research shows that AI is already affecting health care coverage decisions. Our three primary findings were:
- Health insurance plans now regularly use AI for utilization management and that use is wide-ranging and varied. Because health plans have individual approaches, and their AI technologies and practices are often proprietary, greater transparency with both regulators and consumers is needed to safeguard against harm.
- Stakeholders see immense opportunities but warn that proper safeguards are missing and needed to protect consumers. Multiple interviewees were concerned that AI-based decision-making processes will perpetuate and amplify biases that reflect historical patterns of mistreatment of marginalized and minoritized populations.
- Some states have begun to regulate the development and use of AI in health insurance but have not been able to keep up with the proliferation of AI technologies and practices. This needs to be an area of focus and investment for both state and federal regulatory leaders.
These findings helped the Consumer Representatives for Health to the National Association of Insurance Commissioners outline key considerations for regulators as they seek to safeguard consumer health against adverse outcomes that result from the use of AI in utilization management.
