Does a regular sleep schedule matter more than sleeping a lot?

The right question

"How many hours did you sleep?" is the question asked everywhere, and it has an answer: for most adults, 7 to 9 a night. But two people who sleep seven hours can have very different sleep. One goes to bed and wakes at almost the same time every day; the other sleeps five hours on Tuesday, nine on Saturday, and changes bedtime three times a week. For a long time there was no simple way to measure that difference, so it stayed out of the studies.

What sleep regularity is

In 2017, Andrew Phillips and colleagues proposed a measure: the sleep regularity index. The idea is simple. Pick any moment and ask whether the person is in the same state, asleep or awake, exactly 24 hours later. Repeat that for every moment over several weeks. Someone who sleeps and wakes at the same time every day has an index close to 100; someone who changes every day has a low index. Duration does not enter the sum: a person can sleep little and be regular, or sleep a lot and be irregular.

In the first study with the measure, in 61 university students, the most irregular had a body clock delayed by about two and a half hours compared with the most regular, and melatonin started rising close to midnight instead of half past nine in the evening. The authors' mathematical model explained most of the difference by one thing alone: light exposure, which in the irregular sleepers was less intense by day and more spread into the night.

The study that changed the conversation

In 2024, Daniel Windred and colleagues applied the index to more than 60,000 adults in the UK Biobank who had worn an accelerometer for a week, and followed them for about eight years. Those with the most regular sleep had a clearly lower risk of death, from any cause, than those with the most irregular. And, set beside duration, regularity was the stronger predictor: of the two, it was the one that best distinguished those who died from those who did not.

It was not the first signal. In 2018, Jessica Lunsford-Avery and colleagues had measured the same index in almost two thousand older adults and seen that irregularity was associated with more obesity, hypertension, blood glucose and diabetes, and with a higher ten-year cardiovascular risk, regardless of how many hours they slept. The systematic review by Jean-Philippe Chaput and colleagues, in 2020, pooled 41 studies and more than 92,000 people: late bedtimes and large variation in schedule were associated with worse outcomes in almost every dimension of health they looked at.

And the hours?

They still count. The meta-analysis by Francesco Cappuccio and colleagues, with 16 studies and more than 1.3 million people, showed the U-shaped curve on the pillar page: both persistently short and long sleep are associated with higher mortality. The recommendations for adults stayed at 7 to 9 hours a night. What Windred's study adds is that, to predict what happens to a person, knowing at what times they sleep is worth more than knowing how many.

The answer to the title's question is therefore "yes, and": regularity matters more than was thought, and the hours matter as they always did. A person who sleeps five hours at set times is not protected by regularity.

Why regularity weighs

The body has a clock of about 24 hours that organises temperature, hormones, digestion and sleep itself. That clock is set every day by light, above all morning light. When wake time changes a lot from day to day, light arrives at different hours, the clock receives contradictory signals and the systems it coordinates fall out of step with one another. It is what happens, in miniature, to anyone crossing time zones, which is why the gap between the weekday schedule and the weekend one is called "social jet lag".

In Chaput's review, social jet lag was associated with worse outcomes; sleeping more at the weekend, in itself, was associated with better ones. The difference between the two is the hour: recovering sleep by going to bed earlier does not shift the clock; recovering it by waking three hours later does.

In practice

1. Fix the wake time. If you have to choose one anchor, this is it: it decides when light arrives. A margin of half an hour is regularity; three hours at the weekend is not.
2. Light in the morning, little light at night. That is what Phillips's model points to as the main lever. Ten minutes outdoors after waking do more for the clock than anything at night.
3. Recover sleep by going to bed earlier, not by waking later.
4. Two weeks of observing before changing anything: the time you wake, the time you go to bed, how you wake. Regularity shows in the pattern, not in one night. On the platform, admission asks about hours, awakenings and how you wake; the reading of the whole belongs to the appointment.
5. The hours are still there. Regular and short is not the goal. Regular and enough is.

As with all population studies, these say what happens on average to many people; they do not say what is happening to you. If sleep worries you, it is a conversation for the appointment, with the rest of the context.