by Emily Oster
The second part of the recommendation regards crib bumpers, which are forbidden by the AAP. In fact, some cities (Chicago, for example) have disallowed the sale of bumpers. The concern is that these can cause suffocation.
This recommendation is slightly more complicated since, in fact, there is a purpose for bumpers in the first place: without them, your child can get their arms and legs stuck between the crib rails. This is unlikely to be life-threatening, but can certainly hurt the baby.
It is useful to think about the magnitude of the bumper risk. A 2016 paper in the Journal of Pediatrics counted all the US deaths attributed to bumpers between 1985 and 2012.25 They found forty-eight. To put this in context, during this period there were about 108 million children born in the US and somewhere in the range of 650,000 total infant deaths. Eliminating bumpers in this period would therefore be expected to lower the risk of death by about 0.007 percent, preventing 1 in 13,500 deaths. By contrast, estimates suggest the “Back to Sleep” campaign reduced death risk by about 8 percent—preventing about 1 in 13 deaths. In other words, eliminating bumpers would have, at most, a very, very small effect on risk.
Does this mean you should have bumpers? No, not necessarily. Among other things, older children can use the bumpers to escape the crib and fall out, which can be dangerous on its own. This is just to say that the overall risk associated with them is small.
MAKING CHOICES
Armed with the data, we are now back to where we started in this chapter: thinking about risks, including the risks of terrible outcomes that we are afraid to contemplate. And yet we do need to think about them, and to think about them in the context of the size of the effects, and of what works for our individual families.
Looking back at the results above, it seems clear, first, that having your child sleep on their back and avoiding blankets and pillows and other soft items in the crib are good ideas. Avoiding sofa sleeping is also strongly recommended. These recommendations have the most compelling evidence, and are also the easiest to implement.
It also seems clear that smoking raises the risk of SIDS, especially if you choose to bed share.
Finally, looking at the data, we have to conclude that in terms of SIDS risk, choices about sleep location—in your bed, in your room—matter much more in the first four months of your baby’s life.
This leaves us with a set of choices in the first few months of life—whether to share the bed, share the room but not the bed, or share neither. And since the data suggests that there is some risk to sharing the bed, and possibly also to having your child sleep in their own room, we may conclude the absolute safest thing is to have your child sleep in your room in their own bed for these first few months.
Yet this setup may not work for your family. Let’s imagine that your preference is to share your bed with your infant—maybe you think it will be easier to breastfeed, or you simply want to have the baby close.
If this is the case, there is a strong temptation to dismiss the evidence on risk. It is easy to find parenting sources that point to one study that doesn’t show significant impacts of bed sharing and say it proves there is no risk. This is not a rational way to make this decision. If you want to do this right, you need to confront the idea of risk, think about how to make it smaller (if you can), and then think about whether the (minimized) risk is one you are willing to take.
If you are going to bed share, start by making sure you are not smoking or drinking and that your bed is not full of covers and pillows. And think about your infant: if your baby was premature or had low birth weight, the baseline risk of SIDS is higher, and the absolute increase in risk from bed sharing will be higher also.
And then, finally, you want to really try to think about the numbers.
If we look at the main graph on this page, and imagine that you have a full-term infant and are a breastfeeding mom who does not smoke or drink (and your partner doesn’t, either), the evidence suggests that bed sharing increases the risk of death by 0.14 per 1,000 births. The death rate from car accidents in the first year of life is around 0.2 per 1,000 live births. The bed-sharing risk is therefore a real one, but it is smaller than some of the risks you are likely taking regularly.
With my own children, bed sharing wasn’t appealing, but neither was room sharing. My daughter was in her own room immediately, and my son after a couple of weeks. We did everything we could to limit the risks to this—the crib was bare, we had a video monitor—but, knowing that sharing a room with an infant was not going to work for our family, we accepted the possibility of some increased risk.
This is not the choice everyone will make, but the bottom line is that it is a choice. If you do want to bed share, or don’t want to room share, you can make this decision by thinking that the benefits for your whole family outweigh the risks, even if you accept there are some risks.
The Bottom Line
There is good evidence that infants who sleep on their back are at lower risk for SIDS.
There is moderate evidence that bed sharing is risky.
These risks are much higher if you or your partner smokes or drinks alcohol.
There is some less-good evidence that room sharing is beneficial.
The benefits to room sharing die out in the first few months.
Infant and child sleep may be better if your child sleeps alone after the first few months.
In the crib:
Wearable blanket: check!
Bumpers: very small risk, although small benefits as well.
Sleeping on a sofa with an infant is extremely dangerous.
7
Organize Your Baby
When you’re pregnant, especially for the first time, people have a lot of advice for you. One thing I recall vividly is another economist earnestly explaining to me that it is very important to get your child on a schedule immediately upon arriving home from the hospital. You should decide when they will eat and sleep, and impose that. Babies love it! (So he said.)
My fellow economist was not alone in this belief. A whole army of books and philosophies—Babywise being perhaps the most well-known—suggest getting your baby on a schedule right away. These recommend that, even very early on, when it really is very hard to predict when your baby will sleep, you attempt to impose structure, the idea being that the baby will adapt to and adopt the structure. This can be quite appealing to the new parent struggling to figure out how to understand their baby. Not to mention the promise that such a schedule would let parents better predict when they themselves can sleep.
We didn’t listen to our fellow economist, and with Penelope, there was no schedule. When I was first pregnant with Finn, Jesse sent me the following transcript of a Messenger exchange from when Penelope was four weeks old.
oster.emily(23:41:00 (UTC)): do you want to do something?
oster.emily(23:41:02 (UTC)): I dont know what
oster.emily(23:41:06 (UTC)): also, maybe we should have dinner sometime?
oster.emily(23:42:08 (UTC)): hello?
Note that these messages were sent at midnight. Not only was Penelope not on a schedule, but neither, it seems, were we.
Eventually, of course, she did end up on a schedule, one that looked very much like all the other kids’: sleep at night, three naps during the day at first, then two, then one, then finally none. But each of these transitions was a struggle—to implement, yes, but even just to figure out the timing of. How do you know when your child is ready to drop one of the naps? At some point when we were dropping the morning nap, our nanny went into the other room during lunch, and returned three minutes later to find Penelope asleep in her food.
And this isn’t just about convenience or planning your day. Sleep is important! It’s important for baby development, and for parents.
Your child will be in a better mood if they get the right amount of sleep. For a toddler, napping too much may make it harder to get to sleep at night. This means no sleep for parents. If they nap too little, they may be too overtired to get to sleep at night. This also means no sleep for parents.
How much sleep is enough, and when should it happen? It seems like a simple question, but answers differ widely. Take, for example, the two category-killer sleep books: Ferber (Solve Your Child’s Sleep Problems) and Weissbluth (Healthy Sleep Habits, Happy Child). Both provide some guidance on the amount you should expect your child to sleep.
The trouble is, they do not agree.
Ferber, for instance, says that at six months, a baby should sleep a total of about 13 hours: 9.25 hours at night, and two 1- to 2-hour naps. Weissbluth suggests this same six-month-old should sleep a total of about 14 hours, but with more of those hours falling at night: 12 hours at night, and two 1-hour naps. This is a 3-hour difference in the suggested nighttime sleep.
Weissbluth goes further, suggesting that if your child does not sleep much—for example, if they sleep only nine hours at night—this is a serious problem. And I quote: “Children who slept less not only tended to be more socially demanding, bratty, and fussy but they also behaved somewhat like hyperactive children. Later, I will explain how these fatigued, fussy brats are also more likely to become fat kids.”1 So, no pressure!
But note that nine hours of sleep at night is what Ferber recommends. So is this optimal sleep, or the path to obesity?
In addition, the age ranges for the various important transitions are wide and can be vague. The books generally note that around six weeks, infants start to sleep longer at night; at three to four months, naps start to consolidate; at around nine months, the third nap disappears; at a year to twenty-one months, the second nap disappears; and at three to four years, the final nap disappears. On these latter two transitions in particular, these ranges are wide. A year to twenty-one months is a long time!
Roughly speaking, these claims are based on averages across the population. To see this, consider a meta-analysis of studies of sleep duration.2 The two graphs that follow show, based on this analysis, the expected length of the longest sleep period (which is almost always at night) and the number of naps, both graphed against age.
You can see general patterns emerging here. Around two months, there is a big jump up in the average longest sleep period—this is the consolidation of nighttime sleep. This then increases more slowly as the child ages.
The nap graph contains even more information. Nine to ten months is the point at which the average number of naps is two; at eighteen to twenty-three months, it moves all the way to one.
This paper also summarizes total sleep duration; newborns sleep an average of sixteen hours a day, which falls to thirteen or fourteen hours around one year.
This gives you a sense of what to expect if your child is the average child. Of course, your child is probably not exactly average, and these graphs fail to summarize variation across children.
One of the biggest innovations in data collection over the past few years is the ability to collect data through apps. The era of smartphone parenting has put data collection into overdrive for many of us, and sleep data is no exception. It is not surprising, therefore, to find researchers mining this data trove. One of the advantages of having so much data is that you can look at variations across people.
In 2016, five authors published a paper in the Journal of Sleep Research that used data from a Johnson & Johnson–sponsored app that allows parents to record infant sleep patterns.3 They focused on the set of people who seemed to record reliably, and were able to isolate data from 841 children over a period of 156,989 sleep sessions. (This means the average parent in the study using the app is recording almost two hundred sleep sessions. That is devotion to data.) The granularity of the data allows for interesting analyses and, most important, for us to see how sleep varies across kids.
It varies a lot.
Take, as an example, the question of nighttime sleep length. In this data, the average six-month-old baby sleeps ten hours a night. Great—that’s about what we saw in the studies I mentioned earlier. What about the baby at the 25th percentile (this would be a baby who didn’t sleep much)? Nine hours. What about the 75th percentile? Eleven hours.
Now, what about the whole range of the data for six-month-olds? It turns out, in the data they see babies who sleep as few as six hours at night, and babies who sleep as many as fifteen hours.
This makes things a bit clearer: at least part of the reason the books are vague is that there is not really one answer to the question of how much children sleep at night.
Data on daytime sleep shows a similar amount of variation. The longest sleep session during the day on average increases from an hour to about two hours over the first two years of life, but there is a huge range in this, with some children napping not at all at most ages, and some for up to three hours at a stretch.
And similarly, the timing of the move from two naps to one also shows a lot of variation. Around eleven months, most children have two distinct naps, and by nineteen to twenty months most have one, but there is a long period of transition in the data, showing that the age at which children switch to a single nap varies quite widely.
In conclusion, many aspects of scheduling will be kid specific, and attempts to organize your baby are likely to meet with some of these variations. But not everything varies. In particular, one thing that doesn’t show as much variation is wake-up times. Even at around five or six months, the majority of children wake between six and eight a.m. By the time they get to age two, the range is smaller—six thirty to seven thirty a.m.
Putting together the variation in total nighttime sleep and the lack of variation in wake-up time, you can naturally conclude that bedtimes vary a lot. They do. If you think your child needs a lot of sleep, you probably have to put them to bed earlier, since you cannot really get them to wake up later. If you try to schedule your child to go to bed late and sleep late into the morning, you will probably not succeed.
Some things about a second child are harder, the main one being the presence of the first child. But some things are easier, and at least in my experience, schedule is one of them. Before you have any children, you’re on an adult schedule—wake up for work, eat dinner late, maybe stay up to watch some TV. Catch up on sleep on the weekends. Sometimes, maybe, you go to bed earlier, sometimes later.
Once you have even one child, you’re on their schedule. Wake up between six thirty and seven thirty a.m., breakfast, nap, lunch, nap, dinner, bedtime around seven thirty p.m. (ideally). When the second child arrives, they are not on this schedule immediately, of course, but you know where you are going. The Messenger chat Jesse sent was intended as a warning about where we were headed, but we didn’t get there at all. Yes, Finn was up during the night, but I was in bed with him—or, rather, with him in the cot next to me—from day one. We stuck to the schedule we’d used with Penelope, and he actually got there much faster than she did.
The other thing you realize with your second child is that the unscheduled mess of the first year does end. Your baby will, eventually, arrive at a more predictable sleep schedule. Maybe not right away, maybe not exactly the one you envisioned, but they will get there. And this is perhaps the most reassuring thing of all.
The Bottom Line
There are some broad guidelines for sleep schedule.
Longer nighttime sleep develops around two months.
Move to three regular naps around four months.
Move to two regular naps around nine months.
Move to one regular nap around fifteen to eighteen months.
Drop napping around age three.
There is tremendous variability across children, which you
mostly cannot control.
The most consistent schedule feature is wake-up time between six and eight a.m.
Earlier bedtime = longer sleep.
8
Vaccination: Yes, Please
In the 1950s, about five hundred people—mostly children—died of measles each year in the US; 3 to 4 million were sickened. In 2016, zero children in the US died of measles, and there were an estimated eighty-six cases.1
There is a very simple reason for this decline: the development of a measles vaccine.
Vaccinations are among the most significant public health triumphs of the past hundred years (public sanitation is another good one, although less controversial). Simply put, millions of lives worldwide have been saved by the introduction of vaccines for diseases like whooping cough, measles, smallpox, and polio. A tremendous amount of discomfort and itching, and also some deaths, have been prevented by the chicken pox vaccine. The vaccine for hepatitis B has reduced liver cancer. Newer vaccines also matter: the HPV vaccine has the potential to significantly lower rates of cervical cancer.
Despite this, vaccinations remain one of the most central focal points in the Mommy Wars. Some parents do not want their children vaccinated, fearing injury, autism, or some other unspecified downside. Some parents want to delay vaccines, feeling that risks will be mitigated by spacing out vaccinations.
These concerns—which have grown over time—have visible impacts on disease outbreaks. In May 2017, for example, there was a measles outbreak in Minnesota, with at least fifty cases. The outbreak was concentrated in the Somali immigrant community, where antivaccination activists had made efforts to convince the population that vaccines were linked to autism. Many families did not plan to vaccinate, or were waiting until their children were older. In the meantime, their children got measles.