The Power of Large Behavior Models in Healthcare Consumer Engagement

AI solutions are on the rise throughout healthcare – and both payers and health systems are aiming to deploy these technologies as a way to help increase meaningful engagement with members and patients.

Large language models (LLMs) have contributed greatly to many of the recent advances in generative AI. And while LLMs can support patient engagement efforts, they fall short in helping organizations effectively move people to take specific actions that lead to better health outcomes.

The reason is simple: when it comes to patient and member engagement, language is only part of the equation. Understanding human behavior is the rest.

Uniting AI & Behavioral Science to Empower Healthier Populations

Healthcare organizations don’t need to simply develop automated messages for people they seek to engage. Each person’s health journey must be orchestrated in a way that accounts for individual differences and optimizes for health outcomes specific to their unique challenges and barriers to better health.

The fusion of Behavioral Science (BeSci) and AI makes this possible, ushering in a new era of hyper-personalized health consumer experiences.

Lirio has developed the world’s first Large Behavior Model (LBM) for healthcare, pioneering the bridge between BeSci and AI to significantly enhance health outcomes through hyper-personalized recommendations and communications.

This model closely observes individuals’ health journeys, understands their behaviors within their unique contexts, and draws from past experiences to accurately predict likely future actions and potential health impacts.

This behavioral insight makes it possible to discern the right steps in a person’s unique health journey to meaningfully change those behaviors and consequently, the outcomes.

How the Lirio Large Behavior Model Works

Large language models are known as foundation models because they are not trained for a single purpose, but rather, trained on large amounts of data in a specific area in order to capture patterns useful for solving various problems in that domain.

However, while LLMs are the most recognizable type of foundation model, they are just a subset of these technologies. And just like any other application of machine learning, the algorithmic structures, data, patterns, and objectives must be aligned with the intended use of the model in order to achieve success.

For that reason, foundation models are being built in many different fields in which the data and expertise is available. In all cases, these models utilize scale to deal with the intricacies of complex problems by distilling fundamental patterns upon which more targeted, application-specific models can be built.

LLMs were not built with the right structures, and trained with the right data and objectives, to effect behavior change in healthcare, and thus, they will never be as good at achieving such a goal as a model that is built specifically for that purpose. In other words, LLMs were built to understand linguistic contexts and facilitate the generation of language that fits naturally within said contexts.

Lirio’s Large Behavior Model was built to understand the real-life context of human health journeys to facilitate understanding of how to influence those journeys to achieve better outcomes. An example of this is understanding in which contexts various behavioral science-backed strategies can achieve their intended behavior change objectives. Lirio has sent millions of communications encoding various behavioral science methodologies and tied these to the data and contexts in which they are each most likely to be effective. What’s more, Lirio’s LBM is constantly evolving and getting better as we engage with new populations and with existing populations using additional behavioral interventions.

The Lirio LBM thus informs the creation of a highly individualized representation that captures a person’s healthcare behaviors, motivations, and decision-making processes – which our Personalization Engine can then use to deliver tailored health interventions for each individual in your population.

We call this Precision Nudging^®.

Lirio’s LBM can support the creation of health interventions covering many key areas, including:

• Personalized Health Recommendations: Tailor behavioral recommendations for individuals to support health goals;
• Chronic Disease Management: Provide individualized support for managing conditions like diabetes or heart disease through lifestyle changes, therapy adherence, and at-home care management;
• Behavioral Health Interventions: Customize strategies for improved mental health and emotional well-being;
• Preventive Health Strategies: Identify and proactively intervene with at-risk individuals;
• Patient Engagement & Education: Utilize targeted educational materials to increase healthcare understanding, involvement, and usage of support tools;
• Medication Adherence: Encourage consistent and correct use of medications through monitoring and reminders;
• Postoperative Care & Recovery: Personalize follow-up care and rehabilitation to enhance recovery, support at-home care, and reduce the risk of hospital readmission;
• Nutritional Guidance: Offer diet recommendations tailored to individual health needs, taking account of health conditions, medication use, and other factors;
• Physical Activity Promotion: Design exercise programs that align with specific health conditions and goals;
• Health Risk Assessments: Use patient data to anticipate and address potential health risks;
• Healthcare Navigation: Help people connect with coverage, maximize their benefits, plan for healthcare costs, and support your revenue cycle management.

These interventions are based on proprietary algorithms and data and cannot be replicated by publicly available solutions.

The Lirio platform automatically coordinates and optimizes your engagement strategies so all interactions with your healthcare organization become part of the essential context that the LBM uses to better understand individuals and guide population health initiatives.

Drive High-Impact Behavior Change & Improved Health Outcomes

How Lirio’s LBM helps move people on their journey to better health:

What Makes the Lirio Model Different

There are several key differentiators that separate Lirio’s LBM from any other machine learning (ML) model. Here’s what you need to know about this transformative technology that automates personalization at the enterprise scale.

1. Focused on Human Behavior

We understand that people are different from each other; they are affected by the context in which they live, work, and play, and they change over time. Only Lirio’s LBM can account for these dynamics.

Our LBM has applied many behavioral science-backed interventions at scale in various contexts, enabling it to uncover intricate patterns of actions, decisions, and responses.

Lirio’s LBM is essentially an always-on virtual behavioral scientist able to deftly choose the right approaches to support each individual in your population. This is possible because our team of human expert behavioral scientists has completed extensive research on each specific healthcare behavior we nudge. Moreover, Lirio’s AI models have been designed to encode this behavioral expertise in innovative ways.

Lirio doesn’t just understand what people say, we understand what they do and why they do it.

2. Created for Healthcare

When it comes to behavior change, what works in other industries doesn’t work in healthcare. Health is too important and too individual to resort to anything less than precision.

That’s why we created a specialized model to understand and address the nuances of how individuals interact with their healthcare journeys in a person-centric manner.

Our LBM generates insights deeply rooted in a person’s unique medical history and health status. This knowledge allows us to drive effective, hyper-personalized communications around preventative measures, treatment adherence, care utilization, condition management, and more.

We work with your clinical experts and recognized scientific guidelines to ensure people are encouraged to engage with the recommended behaviors for their health needs. And we flexibly integrate with your care pathways to reduce provider burden, support operational efficiency, and drive your targeted health outcomes.

3. Built on Authentic Data and Personalized Health Information

Unlike publicly available information and data – which may lack the depth and authenticity of real-world healthcare experiences – personal health information, clinical history, and social determinants of health (SDOH) fuel our model.

Moreover, our LBM continues to learn how behavioral science-backed communications affect actual health outcomes via your organization’s patient or member interaction history. This means it is informed by genuine behaviors, decisions, and engagements within the healthcare ecosystem.

To protect the information that drives an LBM, we are committed to security and appropriate use of data in order to maintain the privacy of the people we serve. Our technology is HIPAA and SOC Type 2-compliant and HITRUST certified. Lirio’s team and processes are guided by the tenets of human subjects protection and respect for autonomy.

Unlock the Future of Healthcare Communication