Making Predictive AI Actionable: Why Human Behavior and Physician Partnerships are Key
Predictive analytics and artificial intelligence are no longer futuristic concepts in healthcare; they are active components of modern strategy. Yet, many sophisticated AI models fail at the “last mile.” They can predict an adverse event with stunning accuracy but remain powerless to prevent it. Why? Because data alone doesn’t change outcomes. People do.
In a recent Strategy of Health podcast episode, I spoke with Joseph Vattamattam, co-founder and president of Healthmap Solutions, a company that has built its entire model around this truth. They are proving that the true power of predictive AI is not just in its forecasts, but in its ability to be translated into actionable, human-centric interventions that empower both physicians and patients.
Building a Team Around a “Realistic” Mission
To build a program capable of tackling a problem as complex as chronic kidney disease (CKD), you first need to build an exceptional team. Healthmap Solutions has assembled a roster of industry leaders, and according to Joe Vattamattam, the “secret” is a culture built on an actionable and realistic mission.
This mission-driven culture is centered on aligning the company’s financial incentives directly with patient well-being. “We have such a strong culture of mission here,” Vattamattam shares. “…we align the money to the mission. Meaning the only way we get any revenue is by reducing the cost of care.” This isn’t just an altruistic talking point; it’s a hard-wired business model. Healthmap only succeeds if patients stay healthier and out of the hospital. “The way we reduce the cost of care is by reducing, you know, the unnecessary admissions and visits to the emergency department,” he explains.
This alignment creates a powerful sense of purpose that attracts and retains top talent. When a team’s success is defined by measurable improvements in patient health, the work transcends the day-to-day and becomes a “noble cause.” Furthermore, Vattamattam notes that great people want to work with other great people, and the company’s leadership has leveraged past professional relationships to bring in proven, successful teams from prior ventures.
Using Analytics to See the Problem Differently
Before a single predictive model was built, Healthmap Solution’s used its deep analytical expertise to challenge long-held assumptions about kidney care. Vattamattam, who has a background in mathematical finance and investment banking, brought a quantitative rigor to the company’s initial strategy. They didn’t just accept the industry’s conventional wisdom; they let the data define the real problem.
This data-driven approach uncovered three critical, counter-intuitive insights:
-
Myth 1: Focus on the sickest patients. The assumption was to focus on End-Stage Renal Disease (ESRD) patients on dialysis, as they are the most expensive on a per-member basis.
-
Myth 2: Focus only on nephrologists. Kidney disease is a nephrologist’s specialty, so they should be the only physicians to engage.
-
Myth 3: All CKD patients are high-cost.
This last finding, in particular, made the strategic path clear. To manage the cost and improve the outcomes for this population, they had to develop a sophisticated predictive model to identify that high-risk 30% before they became high-cost.
From “Black Box” to Actionable Insights
Healthmap Solution’s predictive model is its “secret sauce,” capable of forecasting adverse events 6 to 12 months in the future. But Vattamattam is quick to point out that a prediction is useless if it’s trapped in a “black box.” The challenge with many advanced AI models is that they can tell you what is likely to happen, but not why. To solve this, Healthmap solution’s built a second, related model that functions as a “driver report.”
Here’s how it works:
-
Prediction: The AI model analyzes a massive dataset (informed by over $30 billion in healthcare expenditures) and assigns a proprietary risk score to a patient.
-
Explanation: The “driver report” model then identifies the top clinical and non-clinical factors for that specific patient’s risk. It translates the complex “why” into simple, actionable terms (e.g., “neurological and musculoskeletal issues,” “running low on anti-convulsant medication,” “recent ER visit”).
-
Action: This simple, clear report is delivered directly into the workflow of a Healthmap solution’s frontline clinician.
This transforms the entire care dynamic. Instead of a generic wellness call, the clinician can have a highly specific, high-value conversation. “They’re able to converse… at a very meaningful way to say, ‘Hey, we noticed… you’re running low on these drugs… We also know you have chronic kidney disease. Let us help you navigate the next 30 days…’,” Vattamattam describes. This connectivity is where the true value is unlocked. “I think the connectivity between the AI ml stuff and the technology platform is where the magic really happens for Healthmap Solution’s”
The Human Element: Analytics Are Useless Without Behavior Change
This brings us to the core thesis of Healthmap’s success and the title of this article. All the data, analytics, and predictive models are ultimately just a means to an end. The real goal is changing human behavior.
“They [analytics] don’t really make any difference unless you can change some human behavior,” Vattamattam states definitively. “At the end of the day, we need a patient to change their behavior, potentially a provider to change their behavior.”
This is why Healthmap Solution’s refused to create a data-only solution. They knew from the beginning that a “pure data model” that bypasses clinicians would fail. These patients are complex, often seeing three to four different providers and taking 15-20 different medications. You cannot effectively manage their care without becoming a trusted partner to their physicians.
This philosophy is what drove Healthmap Solution’s acquisition of Careium, a technology platform designed to operate at the “healthcare delivery edge.” This platform acts as the final-mile conduit, connecting Healthmap’s insights, the patient, and the provider’s care team. It’s the tool that facilitates the behavior change.
This integration also opens the door to new data streams, such as remote patient monitoring (RPM) and wearables. While Vattamattam says it’s still early, the company is excited about the promise of integrating biometric data from weight scales, blood pressure cuffs, and glucometers to provide even earlier leading indicators of risk.
The Future: A Replicable Model for Chronic Care
When asked what’s next, Vattamattam sees a clear path forward: replicating this model for other complex, multi-chronic conditions.
The Healthmap program, while focused on kidney disease as the “ticket of entry,” is already a comprehensive, whole-person care model. “Once you’re in, we’re working on everything with that member,” he says. “We’re working on their cardiac issues, you know, COPD, behavioral health, social determinants of health.”
Because they have already built the clinical programs and predictive models for these comorbidities (like heart failure), expanding to new disease states is a natural evolution. The core asset is the replicable process: a technology platform that combines powerful predictive analytics with a human-driven, clinician-integrated workflow. This model, Vattamattam believes, could be a game-changer for population health across the industry.
The Takeaway
The Healthmap Solution’s story provides a vital blueprint for healthcare leaders navigating the AI revolution. The value of predictive technology is not in its computational power, but in its application. True transformation doesn’t happen when a server flags a risk; it happens when that insight is demystified, translated, and placed in the hands of a clinician as a tool to build trust and guide a patient toward a better action. This isn’t about artificial intelligence replacing human judgment; it’s about augmented intelligence supporting the critical human-to-human relationships that are, and always will be, at the center of healing.