By COLIN LAWLOR
A patient comes in for an ordinary primary care appointment. The nurse runs through the usual checklist: temperature, blood pressure, pulse, weight, sometimes pulse oximetry. Sleep probably won’t come up. If it does, it will be a side note, and if the patient says, “not great,” what often follows is a brief look of sympathy and the familiar advice to relax a bit before bed.
That is, more or less, what sleep looks like in the most common diagnostic interaction in American medicine. Don’t worry, it is not much, if any better in any other country. The other vitals get numbers, while sleep gets small talk. Calling this a minor gap misses the point.
What the Evidence Says
Sleep sits among the strongest behavioral and physiological predictors we have for chronic illness, cognitive decline, mental health outcomes, and burnout.
Work out of Stanford recently showed that just one night of sleep data (admittedly from a hospital sleep lab), processed by a foundation model called SleepFM, could flag elevated risk across 130 disease categories with high accuracy. The outcomes on that list are not trivial and include all-cause mortality, dementia, myocardial infarction, and heart failure.
A 2025 umbrella review that pooled 29 systematic reviews found two-way, physiologically mediated links between sleep and depression, anxiety, plus a long catalog of cardiometabolic conditions.
And researchers at Washington State University published what is, so far, the longest objective description of sleep in chronic insomnia. Eight weeks of continuous, in-home measurement pointed to something clinicians have struggled to capture for years: night-to-night swings in sleep efficiency, sleep latency, and intermittent wakefulness are central to the condition. Sleep diaries and one-night lab studies kept missing that pattern.
The clinical rationale for measuring sleep is settled, but what remains unclear is whether medicine intends to behave as if it believes its own evidence.
Look at the present setup. Obstructive sleep apnea affects an estimated 960 million people worldwide, and as much as 80 percent of moderate-to-severe cases are still undiagnosed. Chronic insomnia hits more than 800 million people worldwide. Both disorders feed into downstream consequences that are costly and common, like cardiovascular disease, depression, motor vehicle crashes, workplace injuries, dementia, and more. Both can be treated. Yet routine primary care generally does not screen for either.
The American College of Physicians has recommended cognitive behavioral therapy for insomnia as first-line treatment since 2016. Still, most people with chronic insomnia never receive CBT-I, partly because they are never identified in the first place. Clinicians cannot treat what they do not uncover, and they often do not even ask the questions that would surface it.
The Vacuum that Consumer Tech Filled
Talk to working professionals, parents of young kids, perimenopausal women, older adults, teenagers, almost anyone, and sleep comes up fast. People know it matters. They have read about it, they monitor it on a watch, they bring it to their doctor. And more and more, when the clinical system has nowhere to put that concern, they go looking elsewhere.
After more than 16 years in sleep science and health technology, the biggest shift I have watched is the change in what patients do when medicine leaves a gap.
Consumer tech moved into the space that healthcare left open. People measure their sleep, sometimes well, sometimes poorly, through wearables, phone apps, and bedside devices. Apple, Google, and the broader consumer market have helped make sleep feel “countable,” something worth paying attention to. That is genuine progress.
But the next step is where things break. If a patient sees a steady decline in deep sleep reported by their watch across six months, there is typically no clinical pathway for that signal. Most primary care practices are not designed to receive it. Physicians often have little training in interpreting it. Insurers are rarely arranged to pay for the time and work needed to investigate it.
The data is available, but what is missing is the machinery that turns data into insights and care.
So, patients end up doing the interpretation themselves, usually with mixed results, and often while surrounded by wellness content that ranges from thoughtful to careless. That gap is not a consumer problem. It is a medical one.
What Medicine Needs To Do
This case is more practical than it might sound. Medicine does not need to swallow the entire consumer wearable world to take sleep seriously. It needs to do four concrete things.
First, bring validated sleep measurement into routine primary care, right alongside the other vitals. At population scale, the tools already exist for smartphone-based measurement, clinical-grade bedside sensors, and standardized aggregation of wearable data that has been benchmarked against polysomnography. The science is not the bottleneck. Reimbursement, workflow, and training are.
Second, screen consistently for the three most common, most underdiagnosed sleep disorders, obstructive sleep apnea, chronic insomnia and Restless Legs Syndrome, especially in groups where prevalence is high. Primary care is an obvious home for this, but so are obesity medicine, cardiology, endocrinology, mental health, and women’s health. None of these areas do it reliably today.
Third, build a referral and treatment path that functions. When sleep measurement points toward a clinical problem, there has to be somewhere for a patient to go. That means more sleep medicine capacity, broader access to CBT-I, and tighter collaboration between sleep specialists and the rest of the care team. Right now, the route often runs through too few sleep labs and even fewer sleep doctors or behavioral sleep clinicians, which leaves patients waiting or never getting seen. Capacity needs to expand.
Fourth, treat the sleep data people already collect as a legitimate input. Tens of millions of Americans track sleep every night. The data quality varies, and the interpretation is often uncertain, yes. Still, the signal gets much clearer when you add validated measurement and clinical context. That is made easier through high-quality harmonization tools. When a patient walks into an appointment carrying months of self-collected data, they are doing work the system has not formally asked anyone to do. Medicine should take that seriously.
The science is sufficiently advanced. What is left is the operational work of sorting the wheat from the chaff, creating workflows, defending reimbursement, training clinicians, expanding capacity, and treating sleep with the same gravity we have given other vital signs for a century.
Sleep is also an obvious entry point to a larger question. How should medicine leverage the power of continuous physiological signals in everyday care? Sleep is becoming easier to measure, deeply consequential, felt personally, and it has one of the widest gaps between what we know and what we do.
If the healthcare system cannot figure out how to measure and respond to sleep—something universal, intuitive to patients, and supported by uncontested evidence—then the larger promise of preventive medicine driven by physiological data looks shaky. We are past arguing about whether sleep matters. We are past proving the technology can measure it. The question that remains is simpler, and harder: is medicine willing to treat sleep like the vital sign it is?
This one has been ‘slept on’ long enough.
Colin Lawlor is the founder and CEO of Sleep.ai, where he has spent more than a decade developing validated sleep measurement and intelligence technologies.
