How AI Can Transform Health Plan Policy Optimization

Pressure continues to build on health plans to control rising healthcare costs while meeting demands for an improved customer experience. Neither goal is optional — and both are harder to achieve by the day.

Employers expect total health benefit costs per employee to rise 5.8% this year, according to a survey Mercer conducted last fall. If that holds, it will be the third consecutive year where increases in health benefit costs exceeded 5%. The current growth is being driven by several factors, including a rebound in elective surgeries, rising drug costs, and an increase in chronic conditions. These trends not only strain health plan margins but also impact affordability for both employers and patients.

Traditional levers for managing costs like utilization management and payment integrity have limitations in how far they can bend the cost curve. They are also under intense public scrutiny, including from policymakers. Health plan executives must be more proactive and innovative in how they tackle these dueling priorities. One answer lies in policy optimization. By harnessing artificial intelligence (AI), health plans can generate an actionable competitive analysis and drive meaningful change in their markets.

Market policies offer clues for improving performance

Health plans continuously update benefits, reimbursement terms, and clinical coverage policies. Gaining visibility across competitors in a market could be a powerful tool. By modeling competitor policies against internal claims data, health plans can go beyond benchmarking to simulate what would happen to their lines of business by making similar adjustments. Imagine being able to answer questions like:

• Are there areas where we are more restrictive than our peers? Could expanding or curtailing policies improve member experience or affordability?
• Are there coverage areas where the market has moved toward stricter controls, but we’re lagging?
• Can we learn from how other plans streamline policy governance, better incorporate evolving clinical evidence, or modify reimbursement policies?

Although answers are publicly available through such documents as prior authorization lists, provider manuals, and clinical policy bulletins, studying them is hard work. The data are unstructured spanning hundreds of complex and lengthy documents. Digging through them is often manual, time consuming, and can be subjective. It also puts added stress on overburdened clinicians. Additionally, governance forums and decision-making processes rarely seek policy benchmarking at scale to make comparisons across many plans.

Deploying generative AI to conduct policy benchmarks

The ability to rapidly ingest, structure, and compare policies across a market is a strategic differentiator. This is where generative AI can be a tremendous asset. The technology analyzes vast amounts of data — structured and unstructured — across disparate sources to structure information from policies and compare clinical coverage criteria and reimbursement terms across plans in real time.

Recent efforts we’ve pursued with our policy analysis toolkit, Oliver Wyman’s proprietary generative AI-enabled capability, are proving successful. A policy optimization toolkit that’s in development can, for instance, flag where a plan’s prior authorization requirements are stricter than local competitors or identify areas where coverage criteria are tightening across the market.

Health plan clinicians can use that analysis for proactive decision-making and to become market leaders instead of followers. That includes the ability to quickly scan the market to understand where their policies might be producing undesired impacts on patients and providers compared to competitors. They can also discover shifts in the market that warrant tighter controls to reduce cost.

Finding the right level of AI and human collaboration

Generative AI is not meant to replace human reasoning. It’s there to supercharge decision-making. To get the most of generative AI, organizations need to adjust workflows and operations. We see four major implications when it comes to policy optimization:

• Operating model redesign: Breaking down organizational silos will be key to fostering deeper collaboration and better decision-making across clinicians, technologists, and business leaders.
• Policy governance: Governance forums must expand their scope to include external market intelligence and real-time feedback. Organizations need to make trade-offs about evolving medicine and utilization trends and consider patient experience more fully in their decisions.
• Technology and tools: Generative AI needs to be integrated with existing platforms, including robotic process automation and traditional analytics tools, to build a truly intelligent policy analysis engine.
• Monitoring process: Capturing cost-of-care trends, patient and provider abrasion metrics, and other real-time indicators will be vital for a dynamic monitoring system that can validate past decisions, flag areas for review, and proactively identify needed improvements to experience.

3 keys to building an AI-driven system

AI-driven policy optimization holds tremendous potential, but it won’t happen overnight. Forward-thinking health plans can start small and scale strategically. Consider these steps:

Start with a focused pilot: Choose a priority clinical area. Use generative AI to benchmark policies against peers and identify where adjustments are necessary.

Benchmark across policies: Build off a successful pilot to expand the analysis across a broader set of clinical domains and incorporate supplemental data to drive prioritization toward hot spots in cost of care and experience.

Reimagine governance: Redesigning and piloting governance, risk assessment, and monitoring capabilities will enable health plans to glean important insights and scale new processes.

Health plan leaders can’t afford to operate under a business-as-usual mindset. Rising costs and shifting member expectations are not easing anytime soon. Generative AI can be a catalyst to spur more informed decision-making and enable health plans to be nimbler in responding to market demands.