Two People With Identical Sleep Totals Faced Up to 48 Percent Different Odds of Dying Early, and a Doctor Explains the Gap

Almost every conversation about sleep starts with a number. Seven hours, eight if it was a good night, five if the week has been difficult. The tracker on your wrist opens on a total, the guidelines are written around a range, and the question people ask each other in the morning is how long they slept. The answer gets treated as the entire report.

A 2024 analysis published in SLEEP suggests the total may be the less informative half of it.

What the Research Found

Researchers analyzed more than 10 million hours of accelerometer data from 60,977 adults in the UK Biobank, average age 63. Each participant wore a wrist device continuously for a week, which allowed the team to calculate a Sleep Regularity Index, a score from 0 to 100 measuring how closely one day's sleep and wake timing matched the next. Someone falling asleep and waking inside the same window every night scores near the top of that scale, while someone whose timing swings across the week scores near the bottom. Participants were then followed for an average of 6.3 years, during which 1,859 of them died.

Compared with the least regular fifth of sleepers, the four more regular groups showed 20 to 48 percent lower risk of death from any cause. Cancer deaths were 16 to 39 percent lower. Deaths from cardiometabolic causes, meaning heart disease, stroke, and diabetes-related conditions, were 22 to 57 percent lower, the widest gap of the three. The results held after adjustment for age, sex, ethnicity, sociodemographic factors, lifestyle, and existing health conditions. When regularity and duration were then tested directly against each other, regularity was the stronger predictor.

Why Timing Registers Differently Than Hours

Nearly every tissue in the body runs its own clock. Blood pressure, glucose handling, core temperature, and cortisol all rise and fall on daily schedules, coordinated by a master clock in the brain that takes its cues from light and from the timing of sleep and waking themselves. Duration tells that system almost nothing. Eight hours beginning at 10 p.m. and eight hours beginning at 2 a.m. are the same number and two entirely different instructions.

When the timing shifts from one night to the next, the central clock and the peripheral clocks in the liver, pancreas, and blood vessels reset at different speeds, and the body spends part of every day internally out of step. Researchers have proposed this as the most likely explanation for what they observed, and the shape of the data fits it. Cardiometabolic deaths, the outcomes most closely tied to daily rhythm, showed the largest gap of any category measured.

What Separates This Study From Earlier Sleep Research

Most large sleep studies rely on questionnaires, and people are poor witnesses to their own sleep. They round to the nearest half hour and report the night they intended rather than the one they had, and irregularity in particular is difficult to recall across a week. This analysis measured every participant continuously with a wrist device instead. The authors also tested duration and regularity inside the same statistical models rather than examining each separately, which is the comparison that allowed them to say one carried information the other did not.

What This Does and Does Not Mean

This is an observational study, and it cannot establish that irregular sleep causes early death. The most serious alternative explanation is that irregularity is a symptom rather than a cause. Undiagnosed illness, chronic pain, and early neurodegenerative disease all disrupt sleep timing, sometimes years before anything appears in a medical record. The researchers adjusted for known health conditions, but no adjustment reaches the ones nobody has found yet.

The measurement is also a single week standing in for a lifetime of habit, and the participants were older than the general population and, as UK Biobank volunteers, healthier and less diverse than it. Irregular sleep is frequently not a choice either, since shift work and caregiving produce it and both carry risks of their own. None of this makes duration irrelevant, as unusually short and unusually long sleep remain associated with premature death. The finding is narrower than that: regularity carried information duration did not.

The Takeaway

Hours are the easiest part of sleep to measure, which is most of the reason they became the number everyone tracks. If your own timing moves by hours across a week, the more useful lever may be the hour you wake rather than the total you accumulate.

Study reference: SLEEP , 2024. PMID: 37738616