One Clock, Seven Nights

Open your phone and look at your alarms.

If there is one for weekdays and a different one for Saturday, you are running two schedules a week. Your body has to renegotiate the handover every Friday night and again every Sunday. Researchers named that gap social jetlag, because functionally you are flying two time zones each weekend without leaving the house.

The video made the case that how steady your sleep is predicts more than how much of it you get. That claim is stronger than it sounds, and it has one soft spot that most articles on this subject skip. I am going to walk you through both.

The headline study tracked 60,977 UK adults with wrist accelerometers, not sleep diaries, for close to eight years. The most consistent sleepers had up to a 48% lower risk of dying from any cause than the least consistent, and the effect held after accounting for how many hours people actually slept (Windred et al., 2023, Sleep).

It is not one study. A systematic review published in 2025 pulled together 59 primary studies and found that across five separate low-bias cohorts, the least regular sleepers carried between 20% and 88% higher all-cause mortality, independent of both duration and quality (Kalkanis et al., 2025, Sleep Medicine Reviews). That range is five different cohorts reporting five different numbers, not one pooled estimate. The direction is what replicates.

The pattern shows up on hard outcomes other than death too. In a US multi-ethnic cohort of 1,992 adults free of cardiovascular disease at baseline, those whose bedtime varied by more than 90 minutes night to night had roughly double the rate of cardiovascular events over five years compared with those varying by half an hour or less (Huang et al., 2020, Journal of the American College of Cardiology). Only 111 events, so the individual numbers are loose. The dose-response trend across categories is the finding.

And a study of 88,094 people found irregular sleep predicted incident dementia over a median seven years (Yiallourou et al., 2024, Neurology). One honest wrinkle in that one: the curve was U-shaped. The most extremely regular 5% also trended slightly upward, though that estimate crossed the line into no-effect. Whatever is going on, it is not a simple more-is-always-better dial.

This is the finding that changed how I talk about sleep, and it is new.

Researchers took 72,269 UK adults, measured a week of sleep with wrist accelerometers, and followed them for eight years for major cardiovascular events. Irregular sleepers carried a 26% higher risk than regular sleepers (hazard ratio 1.26, 95% CI 1.16 to 1.37).

Then they asked the question everyone actually wants answered: what if the irregular sleepers still hit the recommended hours? For the moderately irregular, hitting the duration target did offset the risk. For the genuinely irregular, it did not. The authors put it flatly: adequate sleep duration was not sufficient to offset these adverse effects among irregular sleepers (Chaput et al., 2025, Journal of Epidemiology and Community Health).

That is the argument for the seven-night test in one sentence. You cannot bank hours on the weekend and settle up. The schedule is doing something the total is not.

Association studies tell you a pattern exists. They do not tell you why. For that there is one small, very controlled experiment that does something the big cohorts cannot.

Ten healthy adults lived in a lab for ten days on a deliberately misaligned schedule, running on 28-hour days in dim light so their behaviour and their body clocks came apart. Within days: leptin, the hormone that signals fullness, dropped 17%. Glucose rose 6%. Insulin rose 22%. Mean arterial pressure rose 3 mmHg. Their daily cortisol rhythm inverted completely. Three of the eight participants whose glucose could be assessed produced post-meal readings in the prediabetic range (Scheer et al., 2009, PNAS).

That is a causal result, and it is worth holding onto. Misalign the clock and metabolism degrades quickly, in a lab, under controlled conditions.

It is also ten young healthy people pushed to a twelve-hour misalignment. That is shift-work magnitude. It is not the ninety minutes of weekend drift most readers are dealing with. The study proves circadian misalignment can cause metabolic harm. It does not prove that your Saturday lie-in is doing it. The bridge from that lab to your alarm clock is an inference, and I would rather name it as one than let it pass as a finding.

Here is the soft spot.

Everything above says irregular sleepers have worse outcomes. Almost none of it says that becoming regular improves yours. Those are different claims, and the second one is where the evidence gets thin enough to see through.

There is no randomised controlled trial of sleep regularisation with a hard health outcome. Not one. The closest thing is a proof-of-concept study in which 11 adults with high blood pressure were asked to go to bed at the same time each night for two weeks. Their bedtime variability dropped from about 32 minutes to 7. Their 24-hour systolic blood pressure fell by 4 mmHg and diastolic by 3 mmHg, and, importantly, their sleep duration did not change at all, which points at the regularity rather than at extra sleep (Thosar et al., 2025, SLEEP Advances).

Eleven people. No control group. No randomisation. Two weeks. The authors themselves say it ought to be tested in a larger randomised trial. The 2025 systematic review reaches the same verdict from above, naming schedule stabilisation as a priority target for randomised intervention trials, which is the polite way of saying the trials have not happened.

So the honest state of play is three-tiered. That regularity predicts outcomes independently of duration is well supported, device-measured, and replicated across cohorts and endpoints. That circadian misalignment causes metabolic harm is experimentally established, at magnitudes larger than yours. That fixing your schedule fixes your outcomes is an eleven-person pilot and a reasonable hope.

I still think the seven-night test is worth running. Almost everything worth doing for your health sits on evidence roughly this shaped, and unlike most interventions this one is free, reversible, and finished by Sunday. But you deserve to know which of those three tiers you are standing on.

Set one alarm time, seven days a week, and set it before you run the test rather than after.

Most phones still hold a separate weekend alarm from some job you no longer have. Delete that one first. The test reads your real week more honestly with the anchor already in place than it does measuring a schedule you are about to change underneath it.

The video gave you the test. This gave you the evidence behind it, and the honest edge of that evidence.

If you want to run the full seven nights properly, the complete instrument, with the scoring bands and the anchor method worked through for a schedule that survives a real week, lives in Publication 03: One Clock, Seven Nights. This write-up was the evidence. That is the tool.