Effective management of chronic conditions such as diabetes and heart disease extends far beyond hospital walls. The key to successful treatment lies in proactive patient care — intervening before complications arise that necessitate returning to clinical settings.
Identifying at-risk patients at precisely the right moment has traditionally relied more on probability than conventional medical evaluations. This is where artificial intelligence (AI) is revolutionizing healthcare delivery. By analyzing vast datasets, healthcare AI systems can pinpoint patients who would most benefit from preventive interventions. Until recently, healthcare organizations faced a dilemma: either hire expensive data scientists or resort to generic solutions not tailored to their specific patient populations.
Enter ClosedLoop.ai, a groundbreaking startup that's democratizing access to healthcare AI through an adaptable analytics platform. This innovative solution enables hospitals to seamlessly integrate their data into sophisticated machine learning models, generating actionable insights almost immediately.
The platform's applications are diverse and impactful. Healthcare providers utilize it to forecast which patients might miss appointments, develop conditions like sepsis, or require regular monitoring. Meanwhile, health insurance companies leverage ClosedLoop's technology to make population-level predictions regarding patient readmissions and the progression of chronic diseases.
"We've developed a healthcare data science platform that can ingest any organization's data, rapidly construct models tailored to their specific patient population, and deploy these models effectively," explains ClosedLoop co-founder and Chief Technology Officer Dave DeCaprio '94. "The ability to transform raw healthcare data into functional models requires deep domain expertise—and that's precisely what we bring to our partners."
Amid the COVID-19 pandemic, ClosedLoop demonstrated remarkable agility by developing a model to help organizations identify vulnerable individuals in their communities and prepare for potential patient surges. This open-source tool, known as the C-19 Index, has facilitated connections between high-risk patients and local resources while enabling healthcare systems to generate risk assessments for tens of millions of people nationwide.
The C-19 Index represents just one innovation in ClosedLoop's broader mission to accelerate AI adoption across the healthcare sector—a mission that has defined DeCaprio's professional journey.
Crafting a Vision
Following several years as a software engineer during the internet boom of the early 2000s, DeCaprio sought a more meaningful career path. This search led him to a genome annotation project at the Broad Institute of MIT and Harvard.
This project marked DeCaprio's first encounter with the transformative potential of artificial intelligence. What began as a professional curiosity evolved into a six-year tenure at the Broad, after which he continued exploring the intersection of big data and healthcare delivery.
"After just one year in healthcare, I knew I couldn't return to my previous field," DeCaprio reflects. "Selling online advertisements or similar services suddenly seemed trivial. Once you've contributed to improving human health, other pursuits pale in comparison."
Throughout his work, DeCaprio observed significant flaws in how machine learning and statistical methods were being implemented in healthcare settings. Particularly concerning was the one-size-fits-all approach to predictive models, which failed to account for the unique characteristics of different hospital patient populations.
"Companies would claim they could predict diabetes or readmissions with a single model," DeCaprio notes. "I recognized this approach was fundamentally flawed. The factors driving readmissions in a low-income New York City community differ dramatically from those in a Florida retirement community. The solution wasn't a universal model but rather a system capable of rapidly processing organization-specific data to create tailored predictive tools."
Armed with this insight, DeCaprio partnered with former colleague and serial entrepreneur Andrew Eye to establish ClosedLoop in 2017. The startup's inaugural project involved developing patient outcome prediction models for the Medical Home Network (MHN), a nonprofit hospital collaboration dedicated to enhancing care for Medicaid recipients in Chicago.
As they built their modeling platform, the founders systematically addressed the most common barriers that have historically slowed AI adoption in healthcare.
The initial challenge for most health-tech startups involves compatibility with diverse healthcare data systems. Hospitals vary significantly in both the types of patient data they collect and how they store this information. Even identical data types may be recorded in completely different formats across institutions.
DeCaprio attributes their success to his team's deep healthcare expertise, which enabled them to develop a solution allowing customers to upload raw datasets directly into ClosedLoop's platform and generate patient risk scores with minimal effort.
Another persistent challenge in healthcare AI has been the "black box" nature of many models, making it difficult to understand how they arrive at specific conclusions. ClosedLoop's platform addresses this by providing transparency into the key factors driving each prediction, thereby increasing user confidence in the system's outputs.
To ensure seamless integration into clinical workflows, the founders designed their platform to deliver clear, actionable insights. The system generates prioritized lists, risk scores, and rankings that care managers can readily use to determine which interventions require immediate attention for specific patients.
"By the time a patient walks into the hospital, it's often too late to avoid costly interventions," DeCaprio explains. "The most significant opportunities to reduce healthcare costs come from preventing hospitalizations in the first place."
Beyond individual patient care, health insurance companies utilize ClosedLoop's platform to forecast broader trends in disease risk, emergency department utilization patterns, and potential fraud indicators.
Responding to the COVID-19 Challenge
In March 2020, ClosedLoop began exploring how their platform could assist hospitals in preparing for and responding to the emerging COVID-19 threat. These efforts culminated in an intensive company hackathon during the weekend of March 16. By Monday morning, ClosedLoop had released an open-source model on GitHub that assigned COVID-19 risk scores to Medicare patients. By that Friday, the model had been used to assess risk for more than 2 million patients.
Today, the model has been expanded to serve all patient populations, not just Medicare beneficiaries. It has been instrumental in evaluating community vulnerability across the country. Healthcare organizations have leveraged the tool to anticipate patient surges and help high-risk individuals understand protective measures they can take to avoid infection.
"Sometimes it's as simple as reaching out to socially isolated individuals to connect them with available resources," DeCaprio shares. "An 85-year-old person living alone might be unaware that local organizations can deliver groceries or provide other essential services."
For DeCaprio, bringing AI's predictive capabilities to healthcare has been both rewarding and humbling.
"The scale of healthcare challenges is so immense that regardless of your impact, it never feels like enough," he acknowledges. "Yet when healthcare providers tell us our platform has become their primary tool for identifying which patients need immediate attention, that's incredibly gratifying."
