Does a new study really support leaving your child to cry?

The first issue that strikes me in this piece is in the introduction where it suggests that night wakings and needing comfort to return to sleep is a precursor to “later problems with physiological, emotional, and behavioral self-regulation” with eight citations to support this claim.  Having never read such a claim before, what did I do?  I checked out each and every one of these eight citations.  And here is what they said.

The first, by Ronald Dahl[2] is a conceptual piece, meaning it is hypothesizing a model to explain the development of regulation of sleep and arousal and is not an experimental piece with a study showing this causal pathway, thus it does not actually provide evidence for this claim.

The second by Eric Dearing and colleagues[3] discusses the relationship between poverty and maternal depressive symptoms.  There is absolutely no measure nor discussion of infant arousals or later regulatory problems thus I’m not entirely sure why it’s here at all.

The third is a piece on how sleep affects the development of the visual cortex[4].  This particular study was done with cats and the degree of sleep deprivation given to them would most definitely suggest a problem for an infant and is not at all indicative of the night arousals that many infants experience.  So yes, this study does suggest that there is a possible effect on synaptic plasticity in the brain of cats when severely sleep deprived; how that relates to infant and night arousals, I will leave to you to determine.

The fourth is a large, population based study from the Dunedin Multidisciplinary Health and Development Study and examines the question at hand – long term implications of sleep problems[5].  However, the youngest age at which sleep problems were examined was five years of age.  Nothing from infancy and toddlerhood.  So while the researchers did find that persistent childhood sleep problems were associated with anxiety in adulthood (but not depression), we have no idea if these were problems stemming from infancy or even toddlerhood.  Furthermore, there is no clue as to what these specific problems are as the questions were very vague (“Sleep problems last night?”, “Typically has sleep problems?”, and “Does child have sleep problems?”) and thus may not have anything to do with arousals at night.

The fifth is a review piece by Katherine Karraker[6], in which she reviews the literature on factors that predict night waking in infants beyond the first few months of life.  Importantly, at the very start of her piece, she mentions that the interest and importance of night wakings to clinicians is because of their effects on the parents, not the infant.  She also mentions that at a year of age, approximately 20-30% of infants regularly signal at night to their parents and that many parents who reported infants sleeping through the night see a resurgence in night wakings when infants are between six and 12 months.  What is relevant here is that the research does support that there are a subset of infants who show early signs of behavioural dysregulation who also show frequent night wakings and this seems to follow them into later years.  However, the researcher who has done this work (DeGangi) suggests that the effect is more correlational than causal: Infants with behavioural dysregulation show sleeping problems because of their general heightened arousal and that this heightened arousal is what leads to behavioural problems later.  But that is the only information in this article to suggest night wakings are related to (but not the cause of) later behavioural problems.

The sixth is an article looking at whether sleep problems (defined as night wakings or bedtime struggles) in children six months to four years of age are related to other features of the family[7].  Again, as with the findings in the Karraker piece, the findings are correlational not causational, but they found that sleep problems were associated with any of five factors: an accident or illness in the family, unaccustomed mother absence during the day, maternal depression, co-sleeping, and maternal ambivalence towards the child.  If there were longer term repercussions, I would imagine that, again, we’d be dealing with these factors (minus co-sleeping) as the root cause of both night wakings and later behavioural problems, but this particular piece does not address this.

The seventh is another brief review[8] and the only discussion of a link between sleep problems and behavioural problems is the following sentence “In one survey, 12.5% of 3-year-old children had settling problems, and one-third of them also had a generalized behavioural difficulty”.  That is it.  No reference to causation or any link to earlier night wakings.

And finally, the eighth article is relevant as it examines sleep problems in infancy and later development of ADHD[9].  The study did find that severe sleep problems were predictive of ADHD at age five and half.  However, severe sleep problems were found in only 27 of 2089 children and only seven of the 27 met the criteria for an ADHD diagnosis.  So at the end of the day, only 1.3% of children were diagnosed as having severe sleep problems, a far cry from the 40 or so percent who report night wakings.

Well… that took a while, but hopefully it’s clear that the link between night wakings later self-regulation problems for infants isn’t really established at all, especially outside of any concurrent behavioural problems.  Why is this important?  Because when an article starts off with such a harsh and frightening claim, we have to be certain of its veracity.  Telling parents that their child will have later emotional and behavioural problems because of night waking will lead parents to do whatever they can do avoid that even when it’s not necessarily the truth.

The second issue is that this article presupposes that infants learn to self-soothe by being left alone.  I won’t go into detail here but will suggest you read the following two pieces if you actually believe that this is the case:

1)      From my own site: Educating the Experts, Part Four: Self-Soothing

2)      From Uncommon John (Canadian journalist and now researcher John Hoffman): Self-soothing.  Possibly the biggest lie ever foisted on parents.

If, after reading these pieces, you still believe that infants learn to self-soothe on their own (as opposed to learning through guidance from parents) or that this “self-soothing” in infancy is an active and intentional act, please email me.  We need to chat.

Now, to the study at hand.  The data was collected in 1991 – over 20 years ago – but I would imagine that not much has changed since that time.  Children were assessed at 6, 15, 24, and 36 months of age and children had to have at least two times points to be included.  As with other studies, a large proportion of infants were night-waking at least 3 nights per week at the various time points, including the later time points, which suggests to me that this remains normal child behaviour into the second and third years of life.  The results suggest two groups of sleepers, one made up of approximately 66% of the sample who from six months on experienced night wakings between one and two nights per week and this remained steady until 36 months.  The second, made up of the remaining third, were deemed “transitional sleepers” and they began with night wakings seven nights per week, but by approximately 18-21 months of age, averaged less than one night a week of wakings.

What’s amazing to me is that this natural course, as found by these researchers, shows that by 3 years of age, almost all children are sleeping through the night with few problems to do with night wakings or sleep disturbances.  A third of children took longer to get there, but they got there.  The authors point out that certain variables distinguish between the two groups and make it more likely that the child is a transitional sleeper: child sex (being male), difficult temperament at six months, breastfeeding at six months, breastfeeding at 15 months, maternal depression at six months, health of mother’s partner at 24 months, and health of mother’s partner at 36 months.  However, none of the effect sizes for these are practically significant and seem to be due to the very large sample in the current study (over 1,000 families).  Because they are practically meaningless differences (and the only effect size that was above .01 was breastfeeding) it is difficult to conclude that any of these factors truly influence sleep patterns.

So how do we get to the idea that we should leave children to cry?  After all – the articles in the media have titles like “Don’t get up if your baby cries at night: Mothers SHOULD leave their babies to ‘self-soothe’ says leading expert” so doesn’t that imply crying-it-out is being supported?  I honestly don’t know where this came from given the actual article. In the discussion, the authors state that “there appear to be two distinct patterns over the first 3 years of life that characterize the sleep awakenings of healthy, community-living children”.  They go on to acknowledge that “by 18 months old, not only are the Transitional Sleepers able to sleep through the night without awakening their parents, but they are indistinguishable from the children who, from the beginning, had frequently slept through the night”.  As for the argument that breastfeeding to sleep might cause problems, the authors state “the finding that infant awakenings decline steeply with declining incidence of nursing suggests that it may not be, as suggested, that breastfeeding contributes to diminished capacity to self-soothe, or we would have found that breastfeeding early in life continued to contribute to sleep awakenings even after the incidence of breastfeeding diminished.”  One of the other findings that is notable is that mothers who were more sensitive had children who showed greater night waking.  The authors take this finding to suggest that maternal sensitivity may be reinforcing these negative behaviours (akin to what Karraker suggests, see [6]), but yet, the failure to see this continue with age suggests that’s not the whole story.  (Personally, I see it as a case of maternal sensitivity allowing the child to express his or her biological needs more readily – including, for some babies (but not all), the need to rouse at night and receive comfort.)  But NOWHERE is there the idea that parents should leave their infants to cry at night.  Nowhere.

The final implications, as stated by the researchers, include that for infants older than 18 months of age who are still waking at night, “interventions may be necessary”, BUT the reason is that the parents reported the behaviour as problematic for them and the family and the first suggestion is to consider a more nuanced bedtime routine.  The authors also acknowledge that their “findings are indeterminate in regards to supporting specific expert recommendations for optimal infant sleep patterns concerning greater bedtime ritualization, discouragement of parental intrusion, and encouragement of infant self-soothing to promote sleep that is more continuous”.  Again, a far cry from leaving your child to cry.

All in all, though I don’t like the initial framing of the work, the findings themselves are actually quite helpful to those of us who are against sleep training.  Nowhere do the authors suggest that leaving a child to cry is best (though comments from Dr. Weinraub to journalists could suggest otherwise, though I believe her comments were taken out of context in that saying not to respond immediately can mean waiting a single minute to see if a child settles without response).  In fact, the research supports what many of us have said time and again: Children will sleep through the night at their own pace.  Even at three years, 6% of children were still waking regularly, and there was no evidence that night wakings were problematic for children at the time or later.  Not all children are the same, and some need more time to adjust to sleeping through the night.  Why we need to be looking for problems where there are none is beyond me.  My advice?  Ignore the popular press and know that the article does not suggest leaving children to cry is best, in fact, it does suggest that children will end up sleeping on their own in their own time.  And if that’s 6 months or 3 years, they are all healthy and normal.