For years, health insurers operated under a straightforward fee-for-service model: providers performed services, submitted claims, and insurers paid the bills. Despite a decade of effort to transition to value-based care—where providers are rewarded for keeping patients healthy rather than just treating illness—insurers continue to face significant challenges in making this model work effectively. We collaborated with Humana on this ongoing industry challenge, developing a pilot that helped 100,000 Medicare Advantage members better manage their chronic conditions through an innovative "health concierge" experience.
Humana was facing dual challenges with its Medicare Advantage business—the need to attract and retain profitable group plans while improving their Star ratings—Medicare's key quality measures that directly impact reimbursement. Under their organization-wide Bold Goal initiative, the Stars Group pilot targeted critical gaps in care including medication reconciliation post-discharge, diabetes care, and blood pressure control.
We believed the key to sustainable behavior change lay in tapping into members' intrinsic motivations—their natural desires for autonomy, competence, and relatedness. This led us to develop an unconventional solution: an online game that would reveal these core motivations and allow us to personalize each member's healthcare journey in meaningful ways.
In the healthcare industry, organizations are immersed in a multi-trillion dollar endeavor to meaningfully connect with patients. Even after the peak of the pandemic, health insurance companies continue to struggle with maintaining meaningful connections with their members, as traditional communication channels prove insufficient for understanding their complete health picture. Critical information often remains out of reach, from mental health status to daily lifestyle factors, leaving insurers unable to effectively support their members' health journeys. These blind spots ripple through the healthcare system, affecting everything from preventive care to the management of chronic conditions.
Healthcare outcomes are shaped by far more than medical care alone. Where people live, their access to transportation and healthy food, exposure to environmental hazards, education levels, and economic stability all play crucial roles in their health—often more significant than clinical interventions.
These social determinants of health (SDOH) create a complex web of influence: neighborhoods with poor air quality and few grocery stores tend to have higher rates of respiratory illness and diabetes; areas with limited public transportation often see lower preventive care visits; communities with housing instability frequently experience higher emergency room utilization.
Recognizing the importance of these systemic factors, we undertook the challenge of sourcing and integrating public SDOH data from various government databases, census records, and environmental reports into a single comprehensive data warehouse. This gave us a deeper understanding of the community-level factors affecting our target population before we even began examining individual health journeys.
Source: Institute for Clinical Systems Improvement, Going BeyondClinical Walls: Solving Complex Problems (October 2014)
Missingness map showing spots in our database with missing values (black = missing)
After establishing our SDOH data foundation and integrating Humana's medical records into our warehouse—a significant technical undertaking—we faced the common challenge of incomplete contact information. As expected in healthcare datasets, many member records lacked current email addresses and phone numbers—crucial data points needed to deliver our personalized communications. Locating and verifying this basic contact information became an essential parallel workstream to enable our engagement strategy.
To measure our intervention's impact with precision, we established two parallel member populations: a test group receiving enhanced personalized communications alongside Humana's standard outreach, while a control group maintained usual communications. Statistical validation—using chi-square goodness of fit tests—confirmed these groups were truly comparable across demographics and health status indicators. This rigorous setup would allow us to quantify improvements through two-sample proportion testing, establishing a solid measurement foundation for our next phase.
We established a KPI continuum that would track our progress from surface-level metrics to deeper behavioral change. Starting with basic engagement measurements, we moved to analyzing our combined datasets for population-level insights. As members discovered their intrinsic motivations through our game, we tracked their increasing activation and self-empowerment. This led to measurable improvements in closing specific gaps in care. However, our north star—and greatest challenge—remained achieving sustained behavioral change, the kind that transforms healthcare from reactive to truly preventive care.
We learned about card sorting as a tool for self-discovery from Richard Leider's work in purpose and meaning—an approach that traced its roots from Japanese ikigai to Deci and Ryan's self-determination theory (1985). By inviting seniors to reflect on their gifts, passions, values, and desired impact—questions rarely asked of this population—we transformed academic principles of intrinsic motivation into an emotional journey of self-discovery. This proved more powerful than traditional external motivators, helping members connect their health goals to their deeper sense of purpose.
Caption: card-sorting game used to gather member’s motivations & lifestyle data
ng superficial groupings that miss the psychological drivers of health behaviors. With rich motivational data from our game, combined with clinical and SDOH information, we could identify deeper patterns in how members approached their health. Our clustering analysis revealed unexpected insights—for instance, members who expressed strong family-oriented values and social connections often showed higher rates of missed appointments, suggesting a tendency to prioritize caring for others over self-care. Using these deeper behavioral insights, we employed K-means with silhouette analysis to discover natural groupings of members who shared similar motivations and circumstances, enabling communications that genuinely resonated with each cohort's core values.
With our segmentation insights in hand, we needed expert-backed content that could be personalized at scale. NYU Langone Health's experts provided evidence-based health communications that we transformed into dynamic 'choose your own adventure' pathways. Rather than using one-size-fits-all messaging, we created 108 distinct communication tracks with embedded tone variations—each designed to close specific gaps in care while resonating with different motivational profiles.
Recent studies in Patient Education and Counseling show that entertainment-education ('edutainment') approaches can improve health knowledge retention by up to 27% among older adults (Shi & Salmon, 2022)—making complex medical information more accessible and memorable. Our testing with Medicare members revealed a strong appetite for more conversational, approachable health communications that brought warmth and personality to typically formal medical messages.
While our data clearly showed member readiness for this evolution in healthcare communication, we understood that shifting an organization's established voice requires careful consideration and time. As innovation partners, our role was to illuminate these new possibilities, demonstrating through evidence what the future of healthcare communication could look like. Our approach balanced respect for Humana's current voice guidelines with a clear vision of where member preferences were leading us—laying groundwork for future transformation in how healthcare organizations engage with their members.
Behavioral science demonstrates that 'nudges'—subtle changes in how choices are presented—can guide decision-making when implemented thoughtfully. As nudges have evolved in the digital age, the academic foundation established by the UK's Behavioral Insights Team remains essential: their BASIC toolkit and ABCD framework (Attractive, Brief, Clear, and Doable) provides evidence-based principles for behavioral design. We applied these guidelines to create interactions that members could control, ensuring each communication aligned with their intrinsic motivations and personal health goals. This grounding in behavioral science helped us maintain focus on meaningful engagement rather than notification frequency.
Caption: Beyond tracking gap closures (avg 9.4% improvement vs control), we quantified program impact by mapping closed gaps to potential cost avoidance using published research. For example, early detection through mammography represents ~$11,500 in lifetime savings per case (Health Affairs, 2015)—just one of several preventive measures we tracked. Note: Estimates based on peer-reviewed studies; actual results may vary.
The pilot represented two firsts for Humana: integrating lifestyle data into personalized communications and implementing bilateral SMS engagement. Through this approach, we steered patients towards essential preventive and chronic condition management services, with real-time analytics informing decision-making. We developed a method to quantify impact, translating closed care gaps into estimated savings of $1,598 per engaged life per year.
Beyond our targeted measures, we observed meaningful improvements across other preventive services—mammograms increased by 26.9% and colorectal cancer screenings by 22.1%.
These collateral improvements suggest that when members become more engaged through personalized, two-way communication, the benefits extend beyond specific gaps we initially aimed to close. This broader impact aligns with our KPI continuum, indicating that increased engagement and health activation can create ripple effects across multiple aspects of preventive care.
Health insurers have long recognized the importance of engaging members to improve the value of care
The pandemic created unprecedented distance between health insurers and their members, with critical gaps emerging in areas like behavioral health tracking and overall population health management. This program helped bridge that distance by establishing a personalized 'health concierge' that gave members control over their engagement.
More significantly, it demonstrated how lifestyle data—previously untapped in healthcare communications—could be integrated with clinical and social determinants of health data to create deeper, more meaningful member connections.
The data showed this comprehensive approach's effectiveness through higher engagement rates, improved gap closures, and sustained participation in health activities. This success provided a blueprint for how healthcare organizations can leverage diverse datasets to transform member relationships while driving better health outcomes.
