94,000 Nights of Apple Watch Data: How Sleep Actually Changes Across Menopause

"I just don't sleep like I used to" is one of the most common things women say about perimenopause, and one of the hardest to pin down. Sleep diaries are unreliable, lab sleep studies capture one artificial night, and most research asks women to remember how they slept months ago.

In May 2026, researchers at the Harvard T.H. Chan School of Public Health published something different: objective, nightly sleep data from Apple Watches worn by women in the Apple Women's Health Study, lined up against each woman's final menstrual period. It is the first look at the menopause transition using this many real nights - 94,118 of them - and it answers a question the "I don't sleep like I used to" complaint has never been able to: how much of it is menopause, and how much is just getting older?

Here is what the data showed, what it did not, and what to do with it.

The study, in plain terms

Who: 338 participants in the Apple Women's Health Study, aged 25 to 59, the large majority 45 to 59, who wore an Apple Watch to bed and consented to share sleep data. They also logged periods in the Health app and answered menopause symptom surveys.

What was measured: The researchers used each woman's last logged period as the anchor and looked at sleep in the 12 months before and the 12 months after it. The main measure was wake after sleep onset (WASO) - the share of the night spent awake after you first fall asleep. It is the number that captures "I fell asleep fine but was up at 3am," which is the menopause complaint, as opposed to "I couldn't fall asleep," which is a different one.

The control: To separate menopause from ageing, they compared the change against participants who simply got two years older without going through the transition.

What they found

After the final period: 4 extra minutes awake

In the year after the last period, participants spent about 0.8% more of the night awake than in the year before. Over an eight-hour night that is roughly four extra minutes.

Participants who aged two years without going through menopause gained about 0.2%, or around one extra minute.

So menopause added roughly three to four minutes of wakefulness per night beyond what ageing alone would have. That is a real, measurable effect. It is also smaller than most women would guess from how it feels - and the gap between those two things is the most useful finding in the study, which we come back to below.

Before the final period: the disruption starts early

The averages hide the more interesting pattern. In the 18 months leading up to the final period, 60% of women with sleep data showed increasing WASO, with an average rise of about 7% over that stretch.

That matches what many women describe: sleep getting steadily worse across late perimenopause, before periods actually stop, rather than something that switches on at menopause. If you are in your mid-to-late forties with cycles that are still coming but getting erratic, this is the window the study is describing. Our guide to how perimenopause is diagnosed covers what else is happening at that point.

Many women showed no change at all

The researchers were explicit that a large share of participants did not follow this pattern. The study's own summary is that "each person experiences perimenopause and menopause differently." Four minutes is the average across women whose sleep barely moved and women whose sleep fell apart.

Which symptoms wreck sleep most

Participants also logged menopause symptoms. Among those tracking sleep, the most common were:

But common and sleep-disrupting were not the same list. When the researchers looked at which symptoms tracked with the most fragmented sleep, the strongest links were with more severe:

• bladder symptoms (urgency, frequency, getting up to urinate)
• joint symptoms (aches and stiffness)
• heart discomfort (palpitations, racing or pounding heart)
• depressive symptoms

Hot flashes were the most reported symptom by a wide margin, but on this data they were not what most separated the women with the worst sleep from the rest. That does not mean night sweats do not wake people - they clearly do, and we cover them in night sweats during menopause. It means that if your sleep is poor and you have been treating it purely as a hot-flash problem, the study points at four other things worth checking.

Finally, 84% of participants with sleep data attributed their sleep changes to the menopause transition, compared with 77% of those without sleep tracking reporting perimenopause symptoms. Women who could see their own data were more likely to connect it to what was happening in their bodies.

Why "four minutes" feels like so much more

This is the finding that deserves the most attention, because it explains a lot of frustrating conversations with doctors.

Four extra minutes awake sounds like nothing. But WASO is an average across the whole night, and it is not evenly spread. Most people do not wake for four extra minutes; they wake for 40 minutes on one bad night and sleep through the next two. On the 40-minute night the experience is a wide-awake 3am, thoughts racing, too hot, needing the bathroom - and the next day is wrecked.

There is a second effect. The transition from deep sleep to waking, and the anxiety of lying awake, are what people remember. Four minutes of extra wakefulness that arrives as one abrupt awakening in the middle of the night is a completely different experience from four minutes added to the time it takes to fall asleep. The watch counts them the same. Your brain does not.

So the study is not saying menopause sleep problems are small. It is saying that the average objective change is modest, that it starts before periods stop, that it varies enormously between women, and that the symptoms most tied to the worst nights are ones that are often not asked about.

The limits, stated honestly

• Self-selected, Apple-owning sample. 338 women who own an Apple Watch, wear it to bed, and enrolled in a research study are not a random sample of women going through menopause.
• Watch-based sleep staging is an estimate. Consumer wearables infer sleep and wake from movement and heart rate. They are good at detecting long awakenings and less good at brief ones. The direction of the finding is solid; the exact minutes are approximate.
• Menopause was defined by the last logged period. That depends on consistent logging, and "final" can only be known in hindsight - twelve months without a period.
• Association, not cause. The symptoms linked with worse sleep could be causing it, be caused by it, or share a cause (low estrogen does all three).
• One year either side. Sleep may keep changing well beyond the first postmenopausal year; this study cannot say.

None of this undermines the core result. It is the reason to treat the numbers as a well-measured map rather than a personal forecast.

What to do with it

The study's own recommendations are the standard sleep advice, and they are worth restating because they line up with the symptoms it found:

1. Keep the bedroom cool - the researchers suggest 60 to 67°F (16 to 19°C). This is for hot flashes and night sweats, still the most common complaint even if not the strongest sleep predictor.
2. Limit fluids and bladder irritants in the evening. Given that bladder symptoms were among the strongest predictors of broken sleep, this is not filler advice. If you are getting up more than once a night, say so to your doctor; vaginal estrogen is a well-established treatment for the genitourinary symptoms of menopause, and it is underused.
3. Move regularly, which helps joint stiffness, mood, and sleep depth together.
4. Keep a consistent schedule, including on weekends.
5. Use a wind-down routine - mindfulness, breathing, reading - particularly if racing heart or low mood are part of your picture.

Two things the study points to that the standard list does not:

• Screen for depression, not just hot flashes. Depressive symptoms were among the strongest correlates of poor sleep. Menopausal mood changes are treatable, and cognitive behavioural therapy for insomnia (CBT-I) has good evidence in this group. Our menopause sleep guide walks through the options.
• Mention palpitations. Heart discomfort at night is common in perimenopause and usually benign, but it is also the symptom most likely to be dismissed as anxiety. Bring it up.

If you are not sure whether your sleep pattern fits, the menopause sleep quiz scores your symptoms against the same categories the study used.

Where Femora fits

The single thing that made this study possible was that 338 women had logged their periods consistently enough for researchers to find each one's final period and line up two years of sleep around it. Without the period log, 94,000 nights of watch data would have had nothing to anchor to.

That is exactly what a cycle tracker is for in perimenopause. Femora lets you log periods as they become irregular, so that in hindsight you can see when the last one was - which is the only way menopause is ever actually dated. Logging symptoms alongside (sleep quality, hot flashes, mood, joint aches, palpitations) turns "I don't sleep like I used to" into a pattern you can show a clinician: which nights, how often, and what else was going on. If you wear a watch or ring, see what wearables can and cannot tell you for how to pair the two.

Frequently asked questions

Does menopause really affect sleep, or is it just age? Both, but menopause adds more. The Apple Women's Health Study found about 0.8% more of the night spent awake after the final period, versus 0.2% from two years of ageing alone - roughly four extra minutes against one.

When do menopause sleep problems start? Before periods stop. In the 18 months before the final period, 60% of women in the study showed rising wake-after-sleep-onset, with an average increase of about 7%.

Which menopause symptoms are worst for sleep? In this data, more severe bladder symptoms, joint pain, heart discomfort, and depressive symptoms were most linked to fragmented sleep - more than hot flashes, even though hot flashes were reported by 82% of participants.

What is wake after sleep onset (WASO)? The amount of time you spend awake after first falling asleep. It captures 3am awakenings rather than trouble falling asleep, and it is the measure most affected by menopause.

Does everyone's sleep get worse at menopause? No. Many participants showed no change. The four-minute figure is an average across women whose sleep barely moved and women whose sleep changed a lot.

Can an Apple Watch or smart ring track menopause? It can track sleep, heart rate, and temperature trends, which the Harvard study shows are useful when paired with a period log. It cannot diagnose menopause, which is defined by 12 months without a period.

The bigger picture

Women's health has spent decades running on recalled symptoms and one-night lab studies. A study built on 94,000 ordinary nights, in women's own beds, anchored to their own logged periods, is a different kind of evidence - and its most important result is not the headline number but the shape of the data: the change starts early, it is wildly uneven between women, and the symptoms that predict it are the ones nobody thinks to mention.

If your sleep has changed, the study says two things. You are not imagining it. And the useful next step is to write down when and what else was happening, because that is what turned 338 watches into an answer.

Download Femora to log your periods and symptoms through perimenopause: https://femora.app/download/