At the beginning of every year, my mother-in-law Martha gets inundated with phone calls, emails, texts and fliers about Medicare Advantage plans. She’s a retired teacher in east Los Angeles and her husband, Barry, is on her plan. Barry has vascular dementia and probably should start getting some kind of in-home help. But the plan Martha is on doesn’t offer that—even though the expansion of Medicare Advantage (MA) originated by the Obama Administration after long-term care help dropped out of the Affordable Care Act was to help pay for people to age in place. Martha chose the wrong plan based off of a commercial she saw on television that made her plan sound amazing.
Martha isn’t alone. Annual expenditures in healthcare digital engagement exceeds $5 billion. Innovations have helped in some areas, such as adherence to appointments and online check-ins, while others remain unfazed.
Medicare Part C health plans (Medicare Advantage or MA) are struggling with 10 percent member disenrollment (churn) each year. In 2023 this is a $43 billion problem and despite increasing investments, churn is forecasted to increase to $59 billion in 2028. The Centers for Medicare and Medicaid Services (CMS) is acutely aware of the growing problem, mandating improved benefit communication standards for 2025.
The accelerating MA churn is a curious exception to technology’s success in improving health outcomes at lower costs. Especially given that the primary cause of MA disenrollment is confusion among beneficiaries. In 2022 over 560,000 MA health plan TV commercials aired. We’ve all seen them. Over-65 beneficiaries like Martha are attracted to advertised benefits, such as vision, gym membership or even free food and monthly checks. Despite the saturation of emails, direct mail and robocalls that detail the same benefits in their current MA plan, the beneficiary switches to what they perceive to be a better health plan. Curiously, increased investments in digital engagement have had no impact on slowing this phenomenon.
Health plans are notoriously complex and difficult to understand. MA plans are the extreme example as a typical plan averages over 100 benefits. At the same time, older MA beneficiaries often don’t understand health plan terminology and they aren’t as likely to seek out digital solutions such as online research. In fact, over 2 percent of these plan members are non-digital and this statistic is higher with underserved populations.
At the high level, MA churn looks like a simple problem, but success requires a broader view of both technology and engagement. Obviously, MA beneficiaries want to get the most out of their health plans, but they want to do it, “their way,” meaning via conversations presented in simple, understandable terms—not through chatbots or jargony calls. And given the more than 85 percent hangup rate that robocalls receive, it’s clear that members expect to have these conversations with another human.
Every MA health plan utilizes call centers, in fact, CMS mandates them. Given the need to explain complex plans, insurance terminology and over 100 benefits—all, ideally, in simple language—call centers rely on scripted engagement. But scripts are problematic in themselves. Yes, scripted calls are human to human, but older beneficiaries are looking for a conversation. Because communication is dry, one way and confusing, members retain less than 20 percent of what was covered by the call center representative.
Enter technology. Generative AI has the ability to analyze and understand every square inch of a health plan, while also listening to beneficiaries and identifying their needs. A completely new approach to health plan engagement creates a symmetry between these technologies and call center reps’ human engagement.
The out of the box approach begins with conversations between the rep and beneficiary. The dialogs are personal, discussing the plan member’s goals, personal life and the types of activities that they enjoy. Acting as a virtual assistant, technology interacts with conversations, prompting reps to point out health plan benefits that are a perfect match to the member’s lifestyle and aspirations. When the member is confused by a complex term, the virtual assistant offers another word, and even a glossary of definitions. At the end of the day, the beneficiary happily utilizes those benefits that fit their needs. They feel heard and known and think back on their MA plan’s introduction as a pleasant surprise. Because of the effective communication, these beneficiaries are not at risk of churning out when the next MA TV ad comes along.
But disenrollment risk still exists. In fact, dozens of factors can aggravate churn. This is where tech-enabled conversational engagement is perhaps most valuable. While conversations seem like friendly chats, technology is intentionally steering discussions toward known churn drivers. Not only are member’s responses analyzed, but so too is the sentiment of those responses thanks to advancements in natural language processing, which often yields greater insights than the answers themselves. Given these analyses, technology can generate a predictive churn risk assessment listing disenrollment risk, including aggravating factors, for each member. For example, a beneficiary may seem very happy, having utilized six plan benefits. But predictive technology may warn that the member was notably disturbed by their $300 per month plan premium, noting a friend whose plan premium is $0. As the health plan diffuses this churn risk with a second call, technology again monitors the beneficiary’s responses and tone as it recalibrates the risk of disenrollment.
AI-driven, tech-enabled conversations are a unique application of technology that deliver impressive results. Supported by their virtual assistant, call center reps are all knowing about the plan they’re introducing. Time to answer is reduced by 40 percent, the need to put clients on hold (to find the answer) is reduced by 24 percent, rep training is 30 percent faster and job satisfaction is 35 percent higher. Instead of reading the same script eight hours per day for five months in a row, call center reps spend their time leading interesting conversations as they help plan members live happier, healthier lives.
And churn. Currently, tech-enabled conversational engagement delivers a double-digit reduction in member disenrollment. As data samples increase, AI is again being utilized to identify the currently unknown, disparate data points that combine to aggravate churn risk. Once discovered, these topics will be emphasized in future conversations, leading to ever increasing client utilization, satisfaction and retention.
All of which to say, this year Martha’s experience will be different. And we hope, one day, everyone struggling like Martha will have a similar experience.
