Written by Alison Proffitt
February 26, 2026 | When Leroy Hood took to the podium at the Integrative Healthcare Symposium last week to accept the Visionary Award, he told the audience that he had done 200 push-ups before he came downstairs that morning. He is 87 years old.
It was a fitting start to a lecture that was an intellectual autobiography, a scientific tour de force, and a provocation, a challenge to an entire health care system that Hood believes looks at the problem of human health almost entirely backwards.
“This is medicine like no other,” Hood said, describing the vision that consumed the last two decades of his life. “We will practice this in the future.”
A career built on paradigm shifts
Hood, founder of Phenom Health and one of the most iconic figures in modern biology, traces a career that has repeatedly forced the scientific establishment to catch up with him. He said that when he was a young molecular immunologist at Caltech in 1970, he became convinced that the complexity of human biology was the central problem worth solving, and was equally convinced that solving it would require tools and frameworks that did not yet exist.
“It was absolutely untouchable,” he told the audience. “First, we needed technologies that could extract from an individual their unique data. Second, we needed a new way of thinking about how to analyze data.”
What followed was nearly thirty years of tool building. The Hood Laboratory developed six instruments capable of reading and writing DNA, including an automated protein sequencer—which enabled the cloning of rare, previously unsequenceable proteins and opened five new areas of endeavor in human medicine—and an automated DNA sequencer that made the Human Genome Project technically possible. He was one of twelve scientists who met in Santa Cruz in the spring of 1985 to discuss whether sequencing the human genome was possible and desirable. Ninety percent of biologists were opposed, Hood recalls. It took another five years to get the initiative started.
Hence the founding of one of the first interdisciplinary biology departments at the University of Washington – made possible by Bill Gates – and later, after concluding that large institutional bureaucracies were structurally incapable of pursuing truly new ideas, the Institute for Systems Biology.
The Institute for Systems Biology has become an incubator for systems biology as a discipline: mapping the dynamic networks that underlie biological function, following the cascade of network changes that precede and accompany disease, and building the analytical frameworks needed to understand data on a scale that biology has never before encountered.
The four elements – which are not yet resolved
In the early 2000s, Hood said, he began formulating what he called the four Ps of medicine: predictive, preventive, personalized, and engaging. He said the first three are now achievable: the science exists, the tools exist, and the data is beginning to exist. The fourth problem – participatory – remains the central unresolved problem in health care.
“It’s how you convince patients, doctors, health care leaders, health care technology people, and ultimately politicians… to change our entrenched disease-focused systems in line with health and prevention,” he said. Key: Show economy. It demonstrated its success, demonstrated cost savings, and convinced some leading organizations to adopt it first.
The scale of the opportunity, in HUD’s context, is almost impossible to comprehend. The United States spends $5 trillion annually on health care, of which chronic diseases consume 86%. “If we can eliminate a quarter of chronic diseases within five years, we will save trillions of dollars,” he said. He drew a comparison with the Human Genome Project, whose return on investment, calculated a decade after its completion, was estimated at $800 billion on an investment of $3 billion. “I think phenotyping will dwarf the ROI of the genomics project,” he said.
What the data actually shows
The evidentiary basis for Hood’s optimism comes largely from Arivale’s 5,000-person scientific wellness program that his team has run over the past decade — the kind of longitudinal, high-dimensional data collection that he believes represents the future of clinical research (doi: 10.1038/nbt.3870 and many more)
Each participant was analyzed across six main dimensions of data: genomic variants, blood proteins (initially around 500, now upgraded to 5,000), metabolites, clinical chemistry, gut microbiome, and physiological data from wearable devices. The result was not just a collection of health snapshots, but a dynamic picture of individual health trajectories—and a rich source of what Hood calls “actionable possibilities.”
From the initial pool of 100 people, Hood’s team identified 3,500 statistically significant associations between the analyses. Their curation using community detection algorithms has revealed approximately 70 distinct biological communities, which are groups of analytes that vary in ways consistent with normal physiology or specific disease processes. One such community, focusing on LDL cholesterol, includes thyroid hormone as an associated variable; The team hypothesized that it was a potential drug target and later discovered that Eli Lilly was already in phase III trials for a thyroid hormone-based cholesterol drug. The important point, Hood said, is that drug targets can be discovered from this kind of data, without anyone having to guess.
Among the most clinically striking results from the program:
Biological age differs significantly from chronological age. The team developed methods to calculate biological age from blood proteins, metabolites, and other analyses. Wellness program participants lost, on average, approximately one year of biological age for each year recorded. Even more surprising is that individual organs can age at markedly different rates from each other and from the global biological age estimate, creating the potential for organ-specific recommendations. Every disease examined in the study was associated with a significant increase in biological age.
Microbiome health predicts survival in older adults In a study of nearly 9,000 individuals ages 60 and older, Hood’s team found a striking difference in microbiome trajectories between healthy and unhealthy people. Healthy individuals have gradually lost their core microbiome from midlife and have developed a unique and individually differentiated microbiome composition. Unhealthy individuals did not show such differentiation. Over four years of observation, those with unhealthy microbiome traits were four times more likely to die. (doi: 10.1038/s42255-021-00348-0)
Signs of cancer appear in the blood years before diagnosis. Perhaps the most important finding: Examining blood samples collected one to four years before participants were diagnosed with cancer, Hood’s team identified proteins that were “strikingly elevated,” which sometimes map onto known disease networks characteristic of particular types of cancer. They documented 167 transitions from health to disease in the data set, including 35 transitions to cancer. “The only solution for most diseases is prevention or early diagnosis and response,” Hood said. “We will never be able to cure the disease with just one drug.”
He described the development of the disease as a long period before the clinical stage, during which the biological networks associated with the disease expand dramatically, quietly, and invisibly, before any symptoms appear. By the time a clinical diagnosis is made, “you have a really complex set of disease networks,” he said, and targeted therapies are working to catch up to a problem that has already become complex. The window to intervene is during that early stage, when the networks are simpler and reversible, he said.
Blood is a window to every organ
One of the central regulatory concepts in HUD’s framework is that blood, which flows through all organs on a continuous basis, carries organ-specific molecular signatures, which are proteins secreted at detectable levels by each of the body’s approximately 25 major organs. By tracking these proteins longitudinally, Hood argues that doctors can monitor the functional status of individual organs in ways that medicine today cannot.
He pointed out that the brain is a particularly interesting case: under normal circumstances, it produces approximately 600 organ-specific proteins, but less than 10 of the corresponding proteins appear in the blood. When the blood-brain barrier begins to lose its integrity, more proteins leak into the circulation, and the specific proteins that appear indicate where the barrier has been compromised. In Alzheimer’s disease, metabolic signals can appear in the blood up to 15 years before a clinical diagnosis, he said. In contrast, the placenta produces 37 organ-specific proteins, providing an unprecedented window into the dynamics of fetal growth.
The genomics revolution continues
Hood also identified the next frontiers for genetic analysis, enabled by long-read sequencing technologies. While most genomic research to date has relied on short-read sequencing, which struggles with regions of high GC content and cannot fully resolve structural variants, long-read sequencing allows complete assembly of individual genomes, comprehensive methylation analysis across all cytosines, maternal and paternal chromosome stages, and characterization of the entire landscape of insertions, deletions, and rearrangements that are associated with disease.
“Epigenetics is not just a methylation process,” Hood said. “There will be many other things.” He suggested that long-read sequencing would open up new dimensions of epigenetic analysis that current methods cannot reach. His team is also incorporating viral history reconstruction, biomarkers of brain health, sound-based disease stratification (drawing on the work of collaborators from Luxembourg who have conducted large-scale studies in diabetics), and multispectral saliva analysis capable of distinguishing between healthy and sick individuals in three seconds.
Hood believes that bringing together all these data streams will enable a level of biological integrity and a density of actionable health insights that is qualitatively different from anything medicine has tried before.
Towards the phenomenon project
Hood concluded by outlining an ambitious programmatic agenda under the slogan “The Billion People Genome Project,” with current pilot projects underway in South Korea, South Carolina (1,000-individual diabetes-focused study), Tampa General (60 heart failure patients), and very early-stage rural health initiatives in Alabama and Montana linked to federal funding announced by the Trump administration.
The clinical trial model he envisions differs fundamentally from the traditional randomized controlled trial. Rather than enrolling a homogeneous population, he described a “broad patient spectrum” design that included high-risk individuals with normal biochemistry, those with pre-disease conditions, and patients in the early, intermediate, and late stages of a given disease and followed them all longitudinally for five years.
“This is the way all clinical trials should be done in the future,” Hood said. “You’ll get the life history of all the different subtypes of the condition being studied.” With the density of data now achievable, convincing results are possible with as few as 100 to 200 participants, he said.
The message has been consistent throughout: the future of medicine lies not in better treatments for existing diseases, but in understanding each individual well enough – and early enough – to prevent the disease from establishing itself in the first place.
