Guy Leitersdorf, Founder & CEO of Longevity AI.
Years before my father was diagnosed with cancer, he was doing everything right. He had access to the best medical care in the world. He went to regular checkups. He followed medical advice. On paper, he was the kind of patient our healthcare system is designed to serve well. And yet, by the time his cancer was detected, it was already too late to change the outcome.
What stays with me is not just what happened when my father got sick, but what didn’t happen before. The signals were there—not in a single test result or a moment that clearly demanded action, but in patterns unfolding over time. Subtle changes and early shifts can reveal a progression you can only see when looking at the full arc of a person’s health. But no one could see the progression, because our healthcare system isn’t built to detect it.
Why The Warning Signs Are Often Missed
Medicine today is good at documenting events. It captures diagnoses, lab results and interventions, but it can struggle to see the bigger picture. Health unfolds over years, shaped by patterns—not moments—and by trajectories, not thresholds. A patient’s numbers can still be “normal,” even as their path toward chronic disease is already taking shape.
We are trained to act when something crosses a line, but by then, the story is already well underway. This is not a failure of clinicians. It is a structural limitation of how we have built healthcare. We organize care around visits. We define disease by diagnosis. We intervene once thresholds are crossed. But chronic diseases develop slowly, often silently, across systems, years before detection. By the time we label a chronic disease, we are reacting to a trajectory that has long been in motion. And today, I believe we are stepping in far too late.
The cost of that delay is staggering. In 2023, 51.4% of U.S. adults—approximately 130 million people—reported having two or more chronic conditions. Among Americans aged 65 and older, that figure rises to nearly 80%.
According to the CDC, 90% of the nation’s $5.3 trillion in annual healthcare expenditures are for people with chronic and mental health conditions. The burden spans some of the country’s most prevalent and costly diseases: heart disease and stroke cost the U.S. healthcare system $223.2 billion annually. The cost of cancer care is projected to exceed $240 billion by 2030. Diagnosed diabetes was associated with an estimated $640 billion in medical costs and lost productivity in 2022.
The scale of this burden underscores the need to better use the health data we already have. Many healthcare systems still lack the kind of longitudinal, continuous data needed to fully understand how health evolves.
Israel’s health maintenance organizations (HMOs) are a notable exception, having built decades-long, population-scale health records that make this level of insight possible. Research using long-term population data shows how health factors measured early in life can be associated with disease risk decades later. A population-based Israeli cohort study of 2.3 million adolescents, for example, found that higher BMI in adolescence was associated with increased cancer risk in adulthood.
Turning such patterns into actionable insight for individual patients is what I see as the next step. Imagine if clinicians could see how risk is trending (not just where it is today), early deterioration while values are still “normal,” interactions across systems that compound over time and where small interventions could have the greatest impact. What if every patient had a continuously updated “book” of their health that shows where they’ve been, where they’re going and how that story could still change?
From Medical Records To Health Trajectories
This is where AI begins to matter, but not for the reasons we often hear. The focus is often on diagnosing disease faster. The bigger opportunity is earlier: detecting the first shifts toward disease, when there may still be more opportunities to intervene and change the course. AI should not replace clinicians. It should give them visibility, allowing them to see the trajectory beneath the snapshot.
I believe the future of healthcare is one where long-term health trajectories—not episodic disease events—guide everyday clinical decisions. In that future, prevention is continuous, clinicians are supported with timely, future-facing insight, patients remain engaged between visits, and care is delivered early enough to meaningfully improve healthspan at population scale.
That future is beginning to take shape. Researchers are increasingly using longitudinal health data to model how health evolves over time and identify patterns associated with future disease risk. A 2025 study published in Nature Medicine, drawing on more than 24 million longitudinal clinical visits from 9.6 million people, developed a full-life-cycle biological clock that modeled development and aging across the lifespan and showed that biological-age patterns were associated with current and future disease risk.
Research using U.K. Biobank data developed a machine-learning model to estimate the risk of accelerated onset across dozens of age-related diseases. The model achieved predictive performance above the study’s prespecified threshold for 38 of 47 conditions, including dementia, rheumatoid arthritis and chronic kidney disease, illustrating the potential to identify individuals at elevated risk earlier in the disease trajectory.
Research using longitudinal electronic health records from Clalit, Israel’s largest healthcare provider, developed and validated a machine-learning model that uses routine blood tests and demographic data to identify people at elevated risk of advanced liver fibrosis and prioritize them for earlier clinical follow-up.
The Next Chapter Of Preventive Medicine
Beyond research, hospitals and health systems are beginning to embed AI into preventive care, chronic-disease management and virtual outreach, moving some applications from isolated pilots into clinical practice. But adoption alone is not the goal. What matters is whether these capabilities help clinicians recognize risk sooner and understand how a patient’s health is changing.
My interest in this problem stems from what I saw—and what was missed—with my father. We had data and access to care. What we lacked was the ability to understand what it meant in time to act. Closing that gap means giving clinicians a clearer view of where a patient’s health is headed while there is still time to intervene. Every patient deserves that chance.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
