First, a disclaimer: I am not a medical doctor, and despite what I say here, the least risky choice seems to be to place babies on their backs to sleep — every time.  Nor do I intend to downplay the risk of SIDS.  It is a horrifying ordeal for every parent who has lost a child in this way.  I only seek to try to examine some of the facts and unstated assumptions about sleeping position recommendations.

——————————

There are so many things to worry about with an infant.

One of the strongest messages that you get from many hospitals and pediatricians these days is — babies must always sleep on their back, otherwise (so they say) there is a greatly increased risk of SIDS.  Go on any baby forum online dealing with baby sleeping, and if someone suggests tummy sleeping, you’re likely to see at least a few people chime in about how you’re likely to kill your baby.  I’ve seen quotes from pediatricians who have actually had to say things like “Not all babies who sleeping on their stomach will die.”  Really?  No way!

But how great is the increase in risk?  Why does back sleeping prevent SIDS?  A generation ago, the vast majority of babies slept on their stomachs.  Doctors advised parents to do that.  Now we have whole product lines devoted to back sleeping and SIDS prevention — special pads or ramps to prevent children from flipping over in their sleep, not to mention mattress wrappers (to prevent supposed “toxic gases” from leaking from mattresses), specially designed “sleep sacks” which discourage children from flipping and also allow parents to forgo blankets and comforters (another SIDS issue), etc.  I’m not saying such things don’t help.  But how much can they help?  Are there downsides to the back-sleeping recommendation?

My story

As my baby boy got to the 6-week mark, he was still having a lot of trouble getting to sleep.  It was very touch-and-go: you needed to start soothing him at just the right moment, and there’s no way you could put him in bed and have him just fall asleep.  He needed to be carried, rocked, whatever, until he was at the perfect stage of being sound asleep, and then you could sneak him into bed.  If he wasn’t in quite a deep enough sleep, or if he passed that deep sleep window and was stirring a bit because you held him too long, he would generally wake up within a couple minutes and start screaming.  When he was a little younger, sometimes swaddling tightly could help (and then even if he was only slightly asleep), so he would be prevented from startling himself awake.  But now swaddling generally just annoyed him.

At night it was a little easier to get him to stay asleep.  But getting him down for naps during the day was impossible.  Almost every morning, at some point he’d stay awake for a while, and then he wouldn’t want to go to sleep, despite displaying signs of fatigue.  That would rapidly turn into being overtired, which made him cranky, and pretty soon we were walking around a baby for most of the afternoon, and he would start screaming whenever he was put down or you stopped the right motion or he got tired of a particular position.  Finally, after being up way too long (most books say infants can only be awake for 1 to 2 hours at most before getting overtired), he would fall into a deep nap sometime in the afternoon and sleep way too long, which would make him overly hungry and fussy through the evening.  Occasionally, he’d have a good day, but this was roughly the pattern for a couple weeks.  And before that, it had been the pattern for most of the first month, except then there was no predicting when the fussiness would occur — during the day or in the middle of the night.

Some of the sleep books say that you need to start practicing a kind of “extinction” for naps — that is, get the baby used to consistently going down by himself and soothing himself to sleep.  If he cries or screams for 5 or 10 minutes, just ignore it as long as you know nothing else is wrong (hunger, pain, etc.).  Once you get to 6-8 weeks (some books say 12 weeks, some say 16 weeks, some say six months — I couldn’t find any actual scientific studies behind these seemingly arbitrary ages), you can start real “extinction” for night sleep: put the baby down, and if he screams for an hour, that’s fine, because he’ll usually go to sleep for most of the night.  The next night he’ll probably scream for a shorter duration.  Hopefully, within a few days or a week, he’ll barely cry at all before falling asleep.

Is it really necessary for babies to scream themselves to sleep, though?  Most studies demonstrate that there is no long-term harm in a few days or a week of “sleep training,” but is there a better way?

When my wife was visiting her mother, both her mother and grandmother suggested napping him on his stomach.  And then a miracle occurred — he went to sleep.  And he stayed asleep.  He could even be put down in an awake (but tired) state, and he’d fall asleep on his own.  He hadn’t done that at all for weeks, and before then only when swaddled in the middle of the night.  Of course, at first my wife watched him like a hawk when stomach-sleeping.

My wife had already done a bit of napping with him on his stomach, which she tried even in the first week in a moment of desperation after he fell asleep while nursing.  (She felt guilty about it.)  But she had always been holding him while he was sleeping on her.  She had also been doing some co-sleeping with him in the bed (another SIDS no-no), just because the proximity seemed to soothe him.  I did this with him sometimes as well.

But now he was sleeping soundly on his own.  And for days in a row, he slept on a regular nap schedule, put himself to sleep, and also slept well through the night (waking up for brief periods to eat, of course).  When he was awake, he was more alert and had less tendencies to cry or get fussy.  (Before, he only tended to have one long period of non-fussy awake time during the day, usually for about an hour in the late morning.)  In sum, he was behaving like a more well-rested baby.

When my wife came home from the weekend with her mother, it was like we had a new baby.  Really.  He had been starting to make “social smiles” for a week or two, but they were rather infrequent, since he was typically either asleep, eating, or fussy.  There was little quiet awake time.  Now we took to calling him “Mister Smiley.”  Instead of a pattern of “sleep, eat, fussy, sleep,” 90% of the time, we had “sleep, eat, play, sleep” 90% of the time.  Yes, the books say that babies start to get less fussy around week six, but this change happened suddenly when the baby started to be able to sleep.  While it might be a coincidence, the extra sleep must have contributed.

All the baby books point out the importance of lots of sleep.  The books that focus on sleep even suggest that there are many links to babies who are not well-rested and various problems and disorders, including delayed development.

My child obviously slept better on his tummy, which seems to be a better thing all-around — both for us and for the baby.  But what are the risks?  Most babies a generation ago slept on their tummies, and they didn’t all die of SIDS.  So how great is the risk, really?

Problems in calculating and evaluating the odds

Note to readers — this section contains a bit of semi-technical discussion on statistical methodology.  If you just care about the problems with the SIDS recommendations, skip down to the next section on “Correlation is not causation.”  If you want to see some examples of how SIDS studies exaggerate some claims, obfuscate data, and neglect various considerations in statistical analysis, read on.

It’s actually quite difficult to find good information on the exact risk of sleeping on the stomach, even though numerous studies have been done.  The lack of information on relative risk is so jarring that I really do wonder if it is deliberately omitted for fear that if parents made an informed choice, they wouldn’t necessarily follow the recommendation.  When statistics are given, they are sometimes exaggerated or given in such a way as to be misleading.

For example, pamphlets rarely tell you that the risk of prone (stomach) sleeping compared to supine (back) sleeping is only about 2 to 3 times as great.  But many pamphlets point out that a baby who has always slept on his back has about 20 times the risk when put down on his stomach only very rarely (by another caregiver or by a mother trying something new, etc.).  Studies have shown that babies who only sleep on their backs don’t develop the skills they need to breathe properly when slept on their stomachs, so a baby consistently slept on his back has a greatly increased risk of dying when he is placed on his stomach to sleep or flips himself (as all infants eventually start to do).  Infants are much more likely to die of SIDS in their first week of daycare than just about any other time.  (Another statistic often quoted.)  Why?  Often because they are put to sleep in a way they aren’t used to.

Why do pamphlets deliberately emphasize the huge risks of certain less-likely behaviors, but they don’t actually provide data for parents to make an informed choice on sleep position in general?

For one thing, I’m more and more convinced that doctors and even medical researchers don’t understand basic statistical concepts.  There have been studies on medical students and doctors demonstrating that most of them lack the basic understanding of probability and statistics to be able to evaluate the numbers in a research study and draw appropriate conclusions about what advice to recommend to patients.  They often make the same mistakes that the general public makes in looking at data with a sort of “gut instinct,” rather than doing a few simple calculations that could give them a better perspective.  This intuitive approach often leads to misleading conclusions.

Okay, you say.  But that’s doctors.  Surely the researchers who are collecting and manipulating data must have better math skills.

Perhaps they do.  But why have I seen an appendix in a recent peer-reviewed SIDS paper that simply “shows the work” in great detail for a simple algebraic calculation involving odds ratios that should be intuitively obvious to anyone who understands the definition of “odds ratio”?  Why do you see odds ratios less than 2 being treated as significant findings?  (Many journals refuse to publish studies unless the odds ratios are at least 3-4, and even then, depending on how control groups are done, there’s a chance that the conclusions are problematic.)  Why do I see consistent confusion in SIDS studies between “odds ratios” versus “relative risk” (or “risk ratios”)?  Those latter concepts are not the same at all.

This last issue is particularly important when considering SIDS studies.  If you want a number that people understand intuitively, you generally use “relative risk.”  Say the rate of getting a particular cancer when you eat a healthy diet is 1 in 1000.  Say the rate of getting that cancer when you eat non-healthy food X is 3 in 1000.  I might therefore say, “You have three times the chance of dying of cancer by eating X,” which is a relative risk ratio of 3:1, or simply 3.  We intuitively understand what that means.

But SIDS studies rarely report relative risk.  Instead, they use a more abstract measure called “odds ratio,” although some researchers seem to confuse the terms and mistakenly call them risk ratios or something.  It’s a little bit difficult to explain “odds ratios” in a simple way, but basically they always show the same correlation as “relative risk,” except with a different magnitude.  That is to say, if the risks are equal for conditions A and B, the odds ratio and the relative risk will both be 1:1, or simply 1.  If condition A has a greater risk than B, both the odds ratio and relative risk will be greater than 1.  If the reverse is true, they’d be less than 1.

The problem with odds ratios is that they don’t correspond in size to relative risk, and under certain conditions they can greatly exaggerate a particular trend.  For example, in our earlier example of a relative risk of 3 for eating food X and getting a kind of cancer, the odds ratio might be around 3, or it might be 6, or it might be even 9 or greater.  It all depends on a bunch of variables including how prevalent a condition is in a given population, how a control group is chosen, etc.  In simple terms, an odds ratio doesn’t easily allow you to compare the exact risks in an intuitive way between two conditions.

The odds ratios for prone sleeping versus supine sleeping babies in various studies have ranged everywhere from less than 2 up to greater than 12.  But even those latter studies can’t be used to infer that prone-sleeping babies are at a 12-times greater risk.  Most studies done in the U.S. have come up with odds ratios in the 2-4 range, meaning that the relative risk is almost certainly less than 4 times as much for prone-sleepers.

Part of the reason odds ratios are used instead of relative risk is because you need more data to estimate relative risk.  In particular, you need to be able to estimate a rate of death in an overall population under various conditions.  That can be done using certain kinds of studies (like cohort studies) which usually require advance planning, rather than after-the-fact analysis of data (as in case-control studies, which are the most common SIDS studies).  On the other hand, depending on how control groups are selected in case-control studies, the resulting odds ratios can be greatly biased and therefore can’t give a reasonable magnitude for a risk estimate.

How then can we estimate the overall rates of SIDS deaths for prone versus supine sleepers?  Well, the SIDS “Back to Sleep” campaign always touts their successes with the huge reduction of the SIDS rate in the U.S. in the decade following the initial recommendation in 1992.  Many studies and pamphlets contain a graph showing the prevalence of prone sleeping and SIDS between 1992 and 2002, and how they both consistently go down.

Such data can very easily provide a general estimate of an upper bound for the relative risk for prone sleeping.  When you carry out the calculation, you can easily see that even if the change in sleeping position was responsible  for the entire reduction in the rate of SIDS between 1992 and 2002, prone sleeping cannot have a higher relative risk than about 3 to 1 compared to supine sleeping.

Of course, it’s unlikely that the switch of sleeping position was responsible for the entire reduction of SIDS deaths.  During this time, overall smoking rates (another major SIDS risk factor) declined by about a quarter, breast-feeding rates increased (which should have a net benefit on SIDS rates), and there was a general increase in public awareness of SIDS, probably leading to greater care with infant sleep in general that may have decreased other risk factors as well.  Moreover, better diagnosis for certain conditions and other changes in the way infant deaths were classified may also have contributed to the decline.  About the only risk factor that increased in the 1990s was possibly co-sleeping (whose status as a SIDS factor is heavily debated), but that could not make enough of a difference to eliminate all these other benefits.

It’s also important to note that SIDS was also decreasing rapidly before the back-sleeping recommendation, mostly due to general improvements in infant health-care.  In the U.S., SIDS rates decreased by about 25% in the decade from 1982 to 1992, while 90% of infants still slept on their stomachs.  Overall infant mortality (due to all causes) was decreasing similarly.  (It’s embarrassing to note this, but it’s clear that some study authors don’t even bother to look up the statistics for the previous decade.  When one study started at 1990 instead of 1992 for example, they noted that SIDS was already decreasing, so they tried to come up with some way that people could have known about sleeping recommendations or something before the campaign started.  It’s truly mind-boggling that such research gets published.)

In any case, the 50% reduction seen between 1992 and 2002 undoubtedly had much to do with back-sleeping, but some of that reduction certainly must have to do with other continuing decreasing trends in SIDS and overall infant mortality.  The important message is that despite the huge odds ratios found in most studies, the actual relative risk for prone-sleeping must be less than 3 times as great as supine-sleeping.  I’ve tried various methods to estimate it based on various statistics, and I’d guess it’s close to 2, but it is certainly in the 1.5-3 range.

Now, reconsider that number in light of the many warnings that prone-sleeping “greatly increases the risk of death.”  Not only is “greatly” perhaps too strong a word, but it should be noted that a relative risk less than 3 is generally considered insignificant enough that many major medical journals wouldn’t generally publish such research without a proven causal mechanism explaining the reason for the risk.

Enough studies have been done to prove that this correlation is significant, but what does that tell us about actual recommendations?  How can we evaluate the risk?  Should we change our behavior?  Will it save my baby?

Correlation is not causation: the actual statistics on SIDS

Basically, the statistics in the United States are roughly thus:

– about 1 in 1000-1500 prone-sleeping babies die of SIDS
– about 1 in 2000-3000 supine-sleeping babies die of SIDS

(Black and Native American populations have 2-3 times the rates as whites, but the cause is unknown.  There are also other seemingly random variations that occur in different regions of the US, different seasons of the year, etc.)

No matter what position, 999 of 1000 babies will be just fine.  If we randomly chose a sleeping position for infants, about 1/3 of SIDS deaths would still occur in babies on their backs.

Now, looking at these statistics, you still might want to recommend back-sleeping in general.

But consider one flaw in this hypothesis — you don’t know what the cause of SIDS is.  SIDS, by definition, is UNEXPLAINED.  If there is another cause of death, that cause is listed instead.  A death is only ruled to be SIDS if it is otherwise unexplained.

A recommendation to sleep all babies on their back makes an additional assumption — namely, you would have to assume that prone-sleeping is actually causing SIDS.  But that hasn’t been proven.  There’s a correlation, but no proven causation.  At most, prone-sleeping may be a contributing factor in some cases.

Suppose there are 2000 babies.  If they all sleep prone, about 2 babies will die.  For example, babies 1 through 1998 will be fine, while babies 1999 and 2000 die.  Pediatricians often make the assumption that if all these babies sleep on their backs, then one of the babies will be saved (perhaps 1999), but maybe another will still die in a back-sleeping position (2000).  But we have no evidence that this is the case, since SIDS is unexplained.  It could be that babies 1999 and 2000 both survive on their backs, but now baby 1998 dies… or baby 1993… or baby 852… because they were on their backs, rather than tummies.

Without a mechanism to explain the deaths (and there are dozens of proposed ones, almost all of which are also argued against by some other study), we don’t know that back-sleeping decreases the risk for any given babyAll we know is that back-sleeping reduces the overall death rate. For all we know, sleeping some babies on their backs could be killing them.

Does this seem unlikely?  Why?  We already know that babies who are usually slept on their backs are at a disadvantage when they roll over or are placed on their stomachs, presumably because they don’t develop certain breathing skills as well.  In effect, we’ve made stomach-sleeping more dangerous for babies by only sleeping them on their backs.  Moreover, there are studies showing increased sleep apnea in babies who sleep on their backs.  Sleep apnea has been linked to increased SIDS risk, though back-sleeping is supposed to make it less likely that apnea would cause problems.  And yet, what if the increased apnea rate causes some back-sleeping babies to die?

Or, what about the studies that show that infants who are overtired tend to fall into a deeper sleep and tend to be harder to arouse?  Deep sleep is also implicated in SIDS.  If a baby has a lot of trouble sleeping on his back, he eventually gets overtired and then may fall into a deeper sleep, perhaps resulting in SIDS.  Might that same baby have been okay if he was sleeping more regularly and more comfortably because he preferred to sleep on his tummy?

I don’t know the answers to any of these questions, because they haven’t been studied well.  In fact, in 1992 when the SIDS “Back to Sleep” campaign was started, there had been few studies that linked sleep position to SIDS in any statistically significant way.  It all got started because of lower incidence in SIDS in some Asian countries that had a cultural tendency to place babies on their backs.  Then a couple studies were done that showed some benefit, and now most babies sleep solely on their backs.  While some studies claim that prone sleeping has a 5-10 times greater risk, most studies demonstrate the risk is only about 2-3 times as much than back sleeping.

And we don’t know the cause.  So the “Back to Sleep” campaign amounts to a massive experiment, not based on any causal evidence.

Perhaps that claim seems too strong.  But consider what happened in the couple years following the initial recommendation by the AAP (in 1992).  At first, side (lateral) sleeping was also recommended, rather than prone sleeping.  But it turned out that babies who slept on their sides had an increased risk of rolling onto their stomachs.  And — guess what — babies who were unaccustomed to sleeping on their stomachs tended to have a higher SIDS rate.  In other words, some side-sleeping babies had an increased risk of SIDS compared to those who slept on their tummy consistently, and the pediatricians initial recommendation may have resulted in an increased death rate for some side-sleepers (who might otherwise have slept on the stomachs, according to previous recommendations).  By 2005, they dropped the side-sleeping business and recommended back sleeping as the only suggested position.  How many infants may have died as a result of this uninformed initial recommendation?

It’s very likely that some babies have been saved by back-sleeping recommendation, but it’s also possible that some babies have died as a result.  Without a clear cause, we can’t know how many.  And very few (if any) studies have attempted to determine if there is any difference in risk factors for SIDS deaths in prone position versus deaths in supine position.  We only know that the other commonly identified risk factors (smoking, etc.) seem to occur in both positions.  But dozens of babies in the U.S. die every year of SIDS on their backs with no risk factors, and hundreds die on their backs with some other risk factors.  Why?

Previous medical studies have shown with other conditions that making assumptions about risk factors for rare conditions (and SIDS, with an incidence rate less than 1 in 1000, is a rare condition) without a clear causal link can be problematic or even dangerous.  This is particularly problematic when so many studies are based on odds ratios from retrospective data analysis, rather than randomized cohort studies.  Statistically, what appear to be risk factors may turn out just to be elements of some greater problem — perhaps there is a third factor that is worsened for some babies on their backs and other babies on their stomachs, for example.  Or perhaps there are two different causes for back versus stomach sleepers, even though since the “Back to Sleep” campaign started, research has focused almost entirely on breathing problems that could be worsened in the prone position.  Until other causes are ruled out, all you do is reduce the death rate — you can’t claim to save or reduce the risk for any baby in particular, since you don’t know why they are dying.

The flip side of back-sleeping

But, if we don’t know one way or another, shouldn’t we still err on the side of reducing the death rate?

Well, consider that there actually are proven negative consequences to back sleeping.  Namely, multiple studies have shown developmental delays in back sleepers versus stomach sleepers.  They are still within the normal range, but pediatricians have even proposed pushing back developmental milestones by a month or so because of the increased number of back sleepers.  Shouldn’t this be a consideration?  One study did a follow-up at 18 months and said the difference was no longer significant between the two groups, but that was only one limited study, and no long-term studies have been done.

Then of course there are other disorders — besides sleep apnea, there’s an increased amount of gastroesophageal reflux that occurs with back-sleeping babies.  And even more disconcerting, there’s a much, much greater number of babies who end up with plagiocephaly, or “flat head,” in the first year, resulting from babies sleeping on their backs on the recommended harder surfaces.  Most of the time it resolves itself (and most pediatricians claim it is only a cosmetic issue anyway), but the massive increase shouldn’t be ignored.  Some pediatricians say that they only had a few cases each year in the early 90s, but now they deal with a dozen cases per week.  Could this condition have any risks?

But perhaps most concerning is the sleep issue.  Babies who sleep on their backs experience less slow-wave sleep, which in adults has been proven to be critical in memory formation, as well as a general feeling of restfulness.  Overall sleep is decreased by 8-9% in the first six months, and back-sleeping babies lose about 120 hours of sleep in the first month alone compared to prone sleepers.  Is it any wonder that these babies then end up developmentally delayed?

There are even people proposing possible links to the so-called “autism epidemic” of the past couple decades and this “Back to Sleep” campaign.  For various reasons dealing with changing diagnostic criteria, it’s difficult to estimate how much autism has actually increased versus how much it was simply not diagnosed before.  But most people seem to think it has increased at least somewhat, and it definitely follows the trend of the back-sleeping.  Both autism and SIDS have been linked to problems regulating seratonin levels.  Moreover, there’s a 4:1 male-female ratio in autism, which also happens to be the same 4:1 male-female ratio in babies who develop sleep apnea on their backs.  Coincidence?  Perhaps.  I think this link is much more tenuous, but until there are actual studies done comparing the two, it’s difficult to say.

(Please don’t think I’m actually claiming a link here to autism; I’m merely including one hypothesis that is more on the fringe to point out that such a major change in the way we interact with our babies will undoubtedly have broad effects, some of which may seem minor — like a little reflux — some of which may be quite severe.  But we don’t know, because the studies haven’t been done.)

The actual risk for prone sleeping put in perspective

Now, consider one more thing — the vast majority of SIDS cases have 2 or more risk factors (prone sleeping, soft bedding, blankets or other materials in bed, bed sharing, smoking in the household, young or single parents, African-American or Native American race, no prenatal care, low socioeconomic class, previous respiratory infections or problems, premature delivery, low birth-weight, etc.).  Suppose you have parents who have no significant risk factors (like us).  What are the real odds they should be worrying about?

I find it very weird and significant that I couldn’t find a published study that could begin to answer that question until last year (2009).  This sleep campaign has been going on for almost two decades, and no one bothered to do a statistical analysis that actually separated out the supine-prone position issue from other risk factors, as well as considered the various combinations of risk factors in a comprehensive way.  I didn’t do an exhaustive search, but I think this is really the first study that could answer the question.

And the answer is actually really interesting.  Only about 10% of the prone-sleeping infants who died from SIDS had no other risk factors.  Meanwhile, about 3% of infants who died from SIDS had no known risk factors (including no prone sleeping).

This is very important to realize, since it has critical ramifications for parents trying to make an informed choice.  Most proposed mechanisms for the cause of SIDS deal with particular breathing abnormalities.  Factors like smoking, suffocation hazards in cribs, etc. logically seem like they could be related to respiratory problems.  But roughly 10% of cases have no such risk factors, and about a third of them still die on their backs.  How do we know that the cause for this very small percentage of SIDS cases is the same?  Remember — the cause of SIDS overall is, by definition, unknown.  It’s certainly possible that there are multiple causes.  If research focuses on the majority of cases that might have some sort of respiratory risk factor, it could easily miss out on a secondary cause that occurs in the smaller group without risk factors.  Even if prone sleeping were more dangerous with certain risk factors, we still would have no reason to think that it’s safer for any given baby without those risk factors.

Taking this into account, for people who have no other risk factors, here are the new odds (roughly):

– Chances of a prone-sleeping infant succumbing to SIDS with no other risk factors: about 1 in 20,000 (range 10,000–25,000)
– Chances of a supine-sleeping infant succumbing to SIDS with no other risk factors: about 1 in 50,000 (range 25,000–60,000)

(Note — the numbers don’t quite match up with the other statistics quoted above, because the estimates were generated from a few different studies and calculated in a couple different ways.)

Overall, again, we’re talking about a factor of 2-3 between the two positions.  But the odds are much lower.  If you have other risk factors, perhaps sleep position should be something you should worry about.  But what if you don’t?

Let’s put those odds in perspective:

– Chances of the average person dying within the next year in a car accident: about 1 in 20,000
– Chances of the average person dying within the next year as a pedestrian involved in an accident: about 1 in 50,000

How often do doctors ask you if you walk versus drive to work?  If you had a choice, would they offer a strong recommendation that you walk, on the sole basis that you’d be less likely to die in an accident in the next year?

Most doctors would think such a recommendation on that basis alone would be ridiculous, since the odds are so low and there are so many factors involved that it would be difficult to prevent any given accident — and they only differ by a factor of 2-3.  Moreover, comparing SIDS to an accident is appropriate in this case, since without a proven cause (particularly in cases with no risk factors), statistically it is equivalent to an “accident.”

So why do they issue such a blanket recommendation to parents who have no other risk factors?  Why not tell them the odds, and let parents decide?

Or, why not at least mention that infants less than a month old could significantly benefit from tummy sleeping?  They could get over 4 hours/day more sleep in that first month, and perhaps they’d even be better prepared for tummy sleep if they are later placed in such a position or flip themselves, rather than the current situation where back-sleeping babies are at very high risk in such situations.  SIDS cases are very rare for infants less than a month old — less than 5% of SIDS cases occur then (except for premature babies, who could have different recommendations).  In fact, SIDS is much more likely to occur in such young infants when they are asleep in a seated position (like a car seat) instead of in a prone position.  Why isn’t this recommendation mentioned, since most new parents may think of a car seat as being an “upright” sleep position?

Wouldn’t an extra 120 hours of sleep in the first month be worthwhile for both babies and parents?  I assume the reason this is not recommended is because babies would then have to be switched to back-sleeping to fit in with SIDS recommendations when 4-6 weeks old, and parents would often be faced with fussy screaming infants not used to that position.  They would be less inclined to continue with back-sleeping, and thus the “Back to Sleep” campaign would fail.  Sleep is so important to development in infants — are pediatricians really sacrificing almost 25% of infants’ sleep in the first month in order to promote their campaign to protect less than 0.1% of babies?

Moreover, what about all the negative consequences?  There are few other proven negative consequences to stomach sleeping, but there are quite a few distressing trends for back sleeping (as I pointed out).

What to do?

About five years ago, the New York Times ran a story about how parents were quietly going against SIDS recommendations and letting their babies sleep on their stomachs.  Online forums where someone has admitted to the practice often then have follow-ups from many other parents who say their babies sleep better, though they wouldn’t generally admit to anyone that they were doing the tummy-sleeping.  Some informal online polls have suggested that the percentage of parents using back sleeping is about the same as stomach sleeping (about 40% for each, though these obviously aren’t good statistical samples), despite surveys by pediatricians that say the breakdown is more like 85-90% sleeping on their backs since the “Back to Sleep” campaign.

The pediatricians keep issuing warnings about people who are supposedly uneducated or unaware of the risks — they need to reach grandparents, some minorities, etc.  Studies that interview parents, though, demonstrate that few are unaware of the “Back to Sleep” recommendations.  You can’t miss them; they’re everywhere.  Why do parents do it?  Studies show that the most common reasons are:

  1. concerns about choking
  2. advice from elders
  3. perception that the medical recommendations are unstable, since they used to recommend prone sleep
  4. perceived comfort of the baby
  5. babies sleep better

Concerns (1), (2), and (3) aren’t necessarily good ones — the risk of choking on the back is minimal (perhaps less than on the stomach), elders don’t always know better than doctors, and doctors’ advice sometimes changes.  Whether the baby is “comfortable” is difficult to determine.  However, we can say without a doubt that many babies sleep better on their stomachs.  All the SIDS studies admit that: it’s the reason why babies are placed on their backs — so they can’t go into a sound, deep, restful sleep.

Of course, they say that the sound, deep, restful sleep can be dangerous sometimes for infants who are still learning breathing skills.  Yes — but without other risk factors, we’re talking about maybe 1 in 20,000 of babies that this is a concern for — 19,999 of those 20,000 babies could perhaps use better sleep.  And yet what about the statistic that says babies who aren’t used to sleeping on their tummies (and don’t learn proper breathing skills) are 20 times more likely to die when placed that way, perhaps by a well-meaning grandparent or an exhausted mother desperately trying to get the baby to sleep?  Who is culpable for these deaths?  The caregivers who are supposedly unaware of the risks, or the pediatricians that recommended keeping some babies who don’t like back sleeping so sleep-deprived for the first months of life that they fell into such a deep sleep on their tummies that they died?

I don’t know.  This is all rather difficult to sort through.

But if you’re a parent who is feeling guilty about tummy-sleeping — don’t.  If you’re at high risk due to smoking or other factors, perhaps back sleeping is worth it, even if your baby doesn’t like it.  (For example, SIDS rates for low-income smokers may be as high as 1 in 200 or higher, rather than the 1 in 20,000 for no risk factors.)  And, of course, many babies are perfectly happy on their backs; if so, don’t rock the boat.  But if you’re a parent considering a switch, be very cautious at first, since the highest risk is when babies who are unaccustomed to sleeping on their stomachs are switched.  On the other hand, you might be rewarded with a baby who is healthier and happier because he or she is better rested.

Again, I’m not a doctor, so this is just my opinion after reading a lot of the literature out there.  Consider the risks in context, and make a decision based on all factors.