Using AI to Improve Health Equity and Business Performance

While health insurance trade associations are asking the Centers for Medicare & Medicaid Services (CMS) to revisit the agency’s Health Equity Index (HEI) and its impact on Star rating reward factors, the healthcare industry nonetheless recognizes the importance of identifying social risk factors. Addressing disparities in care related to race, ethnicity and socioeconomic status improves outcomes, reduces overall costs and remains good business practice. What’s more, today’s generative AI and analytics tools make it practical and effective for health plans to realize the business benefits of focusing on health equity. Here’s how:

Identification. Generative AI and analytics tools enable health plans to more effectively identify current and prospective plan members who face social risk factors that can affect their ability to receive equitable care. Electronic health records often don’t capture health equity data, with Z codes still underused. In contrast, generative AI tools and data analysis can identify patterns that indicate social risk factors with data from multiple communications and engagement channels. These factors include housing and food insecurity, transportation challenges, sexual orientation, gender identity (SOGI) issues and race, ethnicity and language (REaL). 

Generative AI can analyze call center conversations to reveal patterns that indicate language barriers, cultural misunderstandings or difficulties in navigating care pathways. Analytics tools use this data to delineate the characteristics of different member segments and their likely reactions to different communication channels and content. Understanding members’ characteristics enables health plans to initiate targeted interventions, such as providing multilingual support, addressing specific health concerns through educational materials or coordinating community services to mitigate social challenges.

Effective engagement through personalization. Most health equity data is highly sensitive. Individuals must trust their health plan and understand how it will use their data before they disclose personal details. Health plans can build trust and display transparency by using generative AI tools and methods that deliver hyper-personalized experiences and engagement that clearly show the plan is addressing an individual’s unique needs.

AI enables health plans to create targeted digital outreach campaigns built on insights from Health Risk Assessment (HRA) surveys and risk adjustment HCC data. Combining this data with AI, plans can reach out to members with personalized follow-up surveys that fit member contexts and are easily completed. Generative AI powered tools can even schedule primary care appointments during which AI agents can help physicians capture additional social determinant of health (SDoH) and SOGI information.

Generative AI agents can greet members when they call or use a chatbot. These agents can understand a member’s context, such as whether they are calling about an insurance card or need to select a physician. Generative AI agents use intent identifiers running on speech-to-text translations and pull up product benefits to accurately answer basic questions. 

Hyper-personal AI-powered communications can also indicate when an at-risk member requires a high-touch experience. For instance, an AI agent may recognize a member calling a helpline does not understand much English. The AI agent can route the member’s call to a service representative who speaks her language. The representative can also help the member navigate a micro website that uses AI to adapt its presentation to the member’s language and content needs.

These techniques will help plans perform better on CMS measures such as “Getting Appointments and Care Quickly,” “Customer Service,” “Call Center,” and “Foreign Language Interpreter and TTY Availability,” which are all weighted at 4x and can highly influence Star ratings.

Better adherence through outreach in preferred channels. Generative AI supports outreach tailored to members’ specific needs and preferences, including targeted emails, paper, text messages, web, microsites and outbound IVR calls. These can gently nudge members toward preventive screenings or medication adherence. Generative AI-powered interactions can provide 24/7 personalized support, answer member questions and even schedule appointments, including those with an impact on quality measures, such as mammograms.

Digital reminders and telehealth consultations for medication management demonstrably increase medication adherence, directly affecting chronic disease management HEDIS measures and related Star ratings. (Source: Journal of Medical Internet Research, 2022)

Steps to Take Now

CMS is scheduled to replace the existing Star reward factor with Health Equity Index (HEI) rewards in 2027. Contracts that meet CMS criteria for serving socially at-risk beneficiaries could receive between 0.1 to 0.4 additional points based on performance data from the 2024 and 2025 plan years. Here are some steps health plans can take to evaluate how they can better tackle health equity: 

• Calculate the potential impact of the HEI reward factor on current ratings and examine how well your organization addresses socially at-risk populations, including underserved members and dual eligibles.

• Assess current technological capabilities, inventorying dependencies on legacy tools and processes and whether the flexibility exists to implement generative AI or AI-like functions.

• Evaluate service providers’ abilities and experience in launching and managing multifaceted AI-first strategies to improve health equity.

By addressing these areas and incorporating AI tools, health plans will be positioned to improve their health equity performance to deliver better service to more members. They’ll also be equipped to comply with regulations and industry trends as healthcare equity standards evolve.