How Much Sleep Do You Need?

I appreciate it when health advice is concrete and clear.

For instance, most experts agree now that adults need 7 to 9 hours of sleep per night on a regular basis.

Simple, right? But how do we know the 7-to-9 hour range is optimal? Can we trust those numbers? What happens when sleep is consistently outside that range? Should naps be counted? Isn't quality of sleep important too?

I want to address these questions here and discuss a study, published last Tuesday, showing that what adults need is the amount of sleep considered normative in their particular culture.

At first glance that's an odd finding. The amount of sleep a person needs should be a matter of individual physiology, not cultural values.

Fortunately, the researchers made a common statistical mistake, and their results point to a simpler and more practical conclusion.

Why is this topic important?

Back in 1898, a Washington Post writer predicted that the need for sufficient sleep would become a "stubborn problem of the new century".

Sure enough, recent CDC and WebMD surveys show that roughly a third of Americans report sleeping less than 7 hours per night on average. More than half of us now say we need more sleep.

I suspect that owing to long work hours, noisy infants, insomnia, illness, etc. people have always complained about insufficient sleep.

What's special about our era is the use of statistics to help quantify the extent of the problem, define the amount of sleep that's optimal, and understand the health impacts of not getting enough.

Why 7 to 9 hours?

In 2010, HHS launched the "Healthy People 2020" initiative that set public health goals for the new decade. Among other things, this initiative supported work on determining how much sleep people need.

For instance, the American Academy of Sleep Medicine (AASM) and the Sleep Research Society (SRS) jointly convened a panel of 15 experts to review the literature and develop evidence-based recommendations.

The National Sleep Foundation (NSF) used similar procedures to obtain recommendations from a separate panel of 18 experts.

AASM-SRS focused on slightly different outcomes and age ranges, but the conclusions were almost identical and published separately by each team in 2015.

• Both teams recommended a minimum of 7 hours of sleep per night on a regular basis for optimal health.

• Both teams expressed uncertainty about the harmfulness of the 6-to-7 hour range, while indicating that it may be less than optimal.

• Both teams addressed the possibility of an optimal upper limit. NSF recommended 9 hours per night as that value.

(AASM-SRS acknowledged that more than 9 hours could be detrimental, but they chose not to call it an upper limit, given that more sleep may benefit some people, and given that scientists aren't sure how excessive sleep undermines health in the first place.)

Experts sometimes recommend at least 7 hours of sleep per night, without mentioning an upper limit. This is partly due to AASM-SRS's uncertainty, and partly because lack of sleep is what most people contend with.

(I don't know about you, but I've never met anyone who complains that their boss lets them sleep in.)

What happens if we sleep too little or too much?

AASM-SRS and NSF each reviewed a number of studies suggesting that sleep outside the optimal range is associated with poorer cardiovascular, metabolic, immunological, and mental health, as well as diminished cognitive functioning, more pain, greater risk of cancer, and higher rates of mortality.

The weeds here are pretty weedy, so I'll just say this: The effects of non-optimal sleep on various health outcomes are neither uniform nor established with equal certainty.

For instance, insufficient sleep is the culprit behind most of the health problems, but the association is clearer for outcomes like cardiovascular disease than for cancer, and premature mortality is more strongly linked to excessive sleep rather than not sleeping enough.

Two caveats

1. The data doesn't show that sleeping more or less than the optimal range automatically hurts us. Rather, it shows that non-optimal sleep increases the risks of various health problems.

(As for how much risk you might incur, that's impossible to say. Studies present group estimates of risks, but individual risk can't be reliably inferred from group data.)

2. The evidence supporting the 7-to-9 hour advice isn't as clear-cut as the advice itself.

For instance, measurements of sleep duration are inherently imprecise.

One approach is to simply ask people how much they sleep. But people may not remember, or their sleep habits may change over time, or they may tell you how many hours they spend in bed (which is not the same as sleep hours), or they may not remember waking up during the night, or....zzzzzz. (That's the sound I imagine you making when my lists get too long.)

Another approach is to give people devices such as actigraphs that record data on movement, breathing, body temperature, etc. The information recorded from these devices can be ambiguous, and researchers can only obtain data for as long as people use them.

The most accurate approach is to monitor people in sleep labs, but for obvious reasons this can't be done for very long stretches.

In short, we might question that 7-to-9-hours-per-night advice. This leads me to a new study that does just that.

Sleep duration and cultural norms

The new study was published this May 6 in a Proceedings of the National Academy of Sciences (PNAS) journal.

In this study, Christine Ou and colleagues at the University of Victoria looked at data on sleep duration and national health across 20 countries.

Roughly 250 participants per country were asked about their previous night's sleep, their typical sleep duration, and their beliefs about the ideal amount of sleep adults should get.

Each participant also received a composite health score based on the number of chronic health conditions they reported, their subjective health, their self-reported mental health, and their levels of depression.

I'd say that Ou and colleagues quantified "health" in a quirky and less-than-desirable way, but their composite measure does seem informative.

Some key findings

1. Amount of sleep per night differed across countries.

French adults slept the most (an average of 7 hours 52 minutes per night), while Japanese adults slept the least (6 hours 18 minutes per night on average).

The U.S. sample averaged exactly 7 hours per night. (There we are – right on the cusp of not sleeping enough.)

2. Across countries, differences in average sleep duration didn't predict differences in health.

No surprise here. You wouldn't expect country-level associations between sleep and health, given that health is influenced by so many variables.

Drum roll please...

3. People whose sleep duration was closer to their cultural norm had better health.

"Cultural norm" was simply the number of hours and minutes of sleep each participant considered ideal for an adult.

By subtracting how much each participant actually slept from how much sleep they considered ideal, Ou and colleagues generated a discrepancy score.

The smaller that score, the better the participant's health. To put it the other way around, the greater their discrepancy score, the poorer their health.

(Generally, people were sleeping less rather than more than what they considered ideal.)

Why would it be healthier to sleep in a more culturally normative way?

Ou and colleagues floated several possibilities. For instance, they speculated that people might feel stress if they notice they're out of synch with others. In the end, the researchers just shrugged their shoulders and said, in effect, who knows?.

No matter. If we trust their data, the implication is that what's optimal is not 7 to 9 hours per night, but rather the specific number of hours that's normative in a person's particular culture.

An alternative interpretation

A common misunderstanding about data is the assumption that if you can trust the statistics, you know what those statistics tell you.

Let's assume we can trust Ou and colleagues' statistics. Do they really tell us that sleeping in culturally normative ways is more healthy?

No, not really.

Cultural norms were measured by asking each participant "what do you think is the ideal amount of sleep an adult should get per night?" Answers were recorded in hours and minutes.

I don't see this as a measure of cultural norms. This question simply elicits opinions about ideal sleep duration.

Thus, the discrepancy between a person's actual sleep and how much they consider ideal doesn't tell us anything about how closely they hew to cultural norms. For instance, if you feel sleep-deprived, you're almost surely going to say that what's ideal is more sleep than what you ordinarily get.

Here's how I would interpret Ou and colleagues' data: We would expect greater discrepancies between actual and ideal sleep to predict poorer health for one or more the following reasons.

1. People who don't sleep enough may be in poorer health simply because of inadequate sleep, not because they fall short of the perceived ideal.

If you only sleep 4 hours per night, your health may suffer, but that's not because you consider 8 hours ideal. It wouldn't matter what your ideal is.

2. People whose actual sleep deviates from what they consider ideal may be in poorer health because of stress about not sleeping as much as they think they should.

This doesn't reflect cultural norms but rather cultural messaging. In recent years we've seen so many news reports about the consequences of insufficient sleep – more accidents, poorer school performance, greater risk of depression, etc. Not sleeping enough is both fatiguing and scary.

3. It may be that health problems lead to sub-optimal sleep rather than the other way around. (Ou and colleagues acknowledge this possibility.)

Some practical advice

1. If you feel alert and energetic most of the day, you're probably getting enough sleep.

7 to 9 hours per night is a reasonable goal, but if you fall outside that range and still feel good, you might just differ from most people in how much sleep you need.

If you sleep 7 to 9 hours per night but still feel poorly during the day, you should consult with your doctor. You may have a health issue (or the quantity and/or quality of your sleep may be lacking).

2. Quality of sleep is important.

The NSF states that someone between the ages of 18 and 64 probably has good sleep quality if the can answer yes to "most of these questions":

• Did you fall asleep in 30 minutes or less?

• Did you wake up 1 time or less during the night?

• In total, were you awake for 20 minutes or less after falling asleep?

• Were you mostly asleep while in bed (i.e., 7 out of 8 hours)?

I find this a little squishy. People often don't remember how many times they wake at night, and even if you're looking at the clock, it's hard to take not of the moment you fall asleep, because…you're falling asleep

All the same, I do think this list gives us an idea of what to work on if we're not feeling alert and energetic throughout the day.

Sleep apps can help, but given the ambiguity of the data they record, you're probably better off just looking at NSF's list and asking, informally, whether it's taking you "a long time" to fall asleep, or you're waking up "a lot", etc.

For tips on how to improve the quality of sleep, check out the AASM, SRS, or NSF websites. Each of them offers good advice.

3. Naps and catch-up strategies can sometimes make up for insufficient sleep.

That "sometimes" is critical. Naps are healthy (not to mention delightful), but they can't compensate for extreme sleep deprivation, and if they're too lengthy they can disrupt the quantity and quality of sleep during the night.

Studies suggest that catching up (e.g., by sleeping in on the weekends) can offset minor deprivation during the week, but it's unclear whether the benefits are sustainable over time.

4. Don't stress about sleep.

Easier said than done, I know, but worrying about sleep tends to undermine it. Meanwhile, as I've noted, the studies don't justify highly precise or one-size-fits-all recommendations, and sub-optimal sleep merely increases the risk of health problems rather than guaranteeing that you'll experience them.

Again, I think that how alert and energetic you feel during the day may be the best indicator of whether you're sleeping well and otherwise living healthy.

Besides, you're already good at sleeping. I sense this. I sense you're so good at it, you can do it with your eyes closed.

Thanks for reading!