In a study published in the journal Pediatrics this week , an Australian research group found no evidence of harm in kids that were sleep-trained, 5 years after the fact. This is the longest follow-up study of sleep training to ever be published, and it’s a randomized controlled trial no less.
Not surprisingly, the media has jumped all over this study with headlines like this one from the Huffington Post: “Baby Sleep Training Methods Safe For Infants.” (The best part about that article is that, below it, there is a link to another HuffPo article from last December entitled, “Cry It Out: The Method That Kills Baby Brain Cells.” That’s a great demonstration of how much the media loves this story and why parents are probably sick of hearing judgment on it either way.)
Parents – particularly those of us who sleep-trained our kids and enjoyed the benefits – are breathing a sigh of relief. I’m happy if the coverage of this study allows parents to shed an unwarranted layer of stress and guilt. If you’ve been following my blog for a while, then you know how I feel about sleep training. After reviewing the literature on this topic, I have concluded that there is no evidence that sleep training harms kids. In contrast, there is a significant body of literature that sleep deprivation poses very real risks to families and that sleep training is often helpful in these situations. It may not work for all babies and parents, but when it does, it can make all the difference.
So where does this new study fit in with the previous research in the field?
I have read several earlier reports from this group – starting with the initial outcomes of sleep training when these kids were infants  and following up with them at 2 years of age . By looking at these kids at 6 years of age, this study is addressing a very real concern among parents. However, it had some major limitations, which have barely been mentioned by the mainstream media.
First, let me summarize the nuts and bolts of this study and what it tells us. Bear with me…
The study included 328 families with 7-month-old babies. The mothers of these babies had all answered “yes” to the question, “Over the last 2 weeks, has your baby’s sleep generally been a problem for you?”
The study was randomized by maternal and child health centers, where families in Australia take their babies for well-baby check-ups and support. There were 174 families attending the intervention centers and 154 families attending the control centers. At the intervention centers, nurses were trained to offer counseling and advice on babies’ sleep issues. At the control centers, nurses weren’t given any specific training. During the intervention sleep consultation, “nurses elicited the nature of the sleep problem, identified solutions, and wrote an individualized sleep management plan with the mother” . Some mothers were counseled on sleep training methods, including “controlled comforting” and “camping out.” This is actually the first time I’ve heard the term “controlled comforting.” It is more commonly known as “controlled crying,” graduated extinction, or the Ferber method, and I have to say that this new term is more palatable (and intentionally so, no doubt). All of these names mean the same thing: that baby is asked to fall asleep on her own, with parents coming to offer a little reassurance at increasing intervals until she falls asleep. In the “camping out” method, parents initially stay in the room as baby falls asleep and then gradually withdraw their presence over a period of a few weeks.
Nurses at the intervention centers offered their sleep advice at the 8-month checkup. However, of the 174 families in this group, only 100 actually showed up to receive the intervention. (It isn’t clear if they missed the 8-month checkup entirely or if they specifically avoided the sleep advice.) And among the 100 families that did show, only 56 recalled being given advice about controlled comforting and 21 recalled being given advice about camping out. It sounds like the nurses gave truly personalized advice and must have offered sleep training as an option only when they thought it was a good fit. Most mothers reported being satisfied with the advice they received, but just 56% of them said that they actually used the sleep strategies “most” or “almost all of the time.” (Note that these results aren’t reported in the most recent study – you have to go back to the 2007 study to find them .)
Do you see the problem? If I’m reading this right, this means that only about 43 babies were actually sleep trained of the original 174 included in the intervention group. And we have no idea how the control families managed their babies’ sleep problems. Parents turn to many sources of information for advice, and just because these families weren’t offered sleep training advice as part of this study doesn’t mean that they didn’t try sleep training at some point. In fact, an earlier paper from this study noted that the control mothers were significantly more likely than the intervention mothers to seek additional professional help with their baby’s sleep (33% vs. 18%, P=0.02) .
Given this, it is kind of incredible that there were significant improvements, albeit small, in infants’ sleep and maternal mental health at 10 and 12 months . By two years, there was no difference between the two groups in the percentage of kids with sleep problems, but the intervention moms were still better off when it came to mental health .
In the current study , the researchers contacted the original families to see if they would be willing to complete some follow-up assessments now that their kids were 6 years old, including a mailed questionnaire, a home-based assessment, and two salivary cortisol samples. Only 69% of families agreed to the follow-up, 59% participated in the home visits, and 46% gave cortisol samples. That’s not exactly a great turnout.
The assessments covered a slew of measures of child sleep issues, stress, and behavior problems, as well as of the child-parent relationship and maternal depression. The cortisol sample was a measure of chronic stress. Among all of these measures, there were no significant differences between the intervention and the control group. This is the result that all of the news stories are reporting: sleep training causes no long-term harm. It also doesn’t provide a long-term benefit to kids, because in none of these measures were the intervention kids better off. As I already mentioned, earlier studies did show a significant benefit at 1 and 2 years of age [2,3].
I see a couple of major problems with this latest round of data. First, a lot of families were lost to follow-up. This is understandable – it’s real life for you. Families move, and the researchers were unable to contact many of them. Others said they were just too busy, and we can all understand that. However, the study doesn’t report how many of the returning intervention kids actually had some documented sleep training. That would be useful information to know, and the authors have it because they reported it in a previous paper. In their data analysis, the authors included all of the families originally assigned to the intervention group, even those that didn’t actually show up to get sleep advice. This type of analysis makes sense if they are trying to determine the overall effects of offering a parenting advice intervention, but it might miss some effects in the subset of kids whose parents actually followed the advice. Given that this study set out to determine if sleep training caused lasting harm, and only a small percentage of the intervention kids were actually sleep-trained, running some specific analyses on just these children would have helped to convince me that sleep training indeed didn’t have any lasting effects.
One final bone to pick: measures of child and maternal well-being were all based on the mothers’ own reports. Mothers certainly know their kids, but they aren’t always the most objective, particularly when asked to judge the quality of their relationships with their own kids.
Even with its limitations, I appreciate that this study was a randomized controlled trial. This is the gold standard for showing causality – rather than just correlation – of an intervention and outcomes. That adds a lot of strength to the findings of benefits in the earlier papers from this study. Contrast this to a study of cortisol concentrations in sleep-trained babies that came out a few months ago . Although this study was interpreted in the media as showing that sleep training was stressful to babies, no such conclusion could actually be drawn, because there was no control group.
The limitations to this study – that many intervention families didn’t show up for their appointments, many didn’t actually need or want sleep training, many weren’t consistent in their approach to their babies’ sleep, and many didn’t participate in the follow-up study – illustrate just how difficult it is to study parenting behaviors, even with the best study design. Ethically, we can’t force any parent to change their behavior in a particular way based on randomized group assignment. The same difficulty arises when we try to tease out the effects of breastfeeding vs. formula-feeding and cosleeping vs. solitary sleep. For these types of parenting behaviors, all we can really do is what this study did: offer advice to one group of parents and not the other and track outcomes over time. Add to this that there are millions of parenting interactions that happen between the time that these parents struggled with their babies’ sleep and the time they were 6 years old. From a researcher’s standpoint, this makes it awfully difficult to tease out the effects of a couple days or weeks of sleep training at 8 months.
So despite the headlines you’ve been seeing, I don’t think this is the last word on sleep training. I wish it were, because I’m tired of all the guilt tripping I see in parenting conversations about this topic. But no, that’s really not the way science works. No study is perfect in design, execution, and interpretation. This is why it takes a body of evidence, built over time, for us to understand anything. So far, the body of evidence on this topic says that sleep training is an effective way to help most kids get a better night’s sleep, and there is no evidence that it is harmful.
I was trained in science, and I value it immensely. However, I have learned a few humble lessons as a parent. For one, parenting is complex. Science can provide some guidance, but no study can account for all of the ways in which we impact our kids. At the end of the day (literally), we parents are left to try to help our babies find their ways to restful sleep. Science can help. Experienced parents and professionals can help. None of them can provide a roadmap for you and your baby. We have to trust ourselves, follow our instincts, and find what works for us. Maybe it’s sleep training; maybe it’s cosleeping; or maybe we just muddle through. Parent with love and respect, and for goodness sake, don’t worry about what HuffPo has to say about it!.
 Price, A. M., M. Wake, et al. (2012). “Five-Year Follow-up of Harms and Benefits of Behavioral Infant Sleep Intervention: Randomized Trial.” Pediatrics. Published online September 10, 2012.
 Hiscock, H., J. Bayer, et al. (2007). “Improving infant sleep and maternal mental health: a cluster randomised trial.” Arch Dis Child 92(11): 952-958.
 Hiscock, H., J. K. Bayer, et al. (2008). “Long-term mother and child mental health effects of a population-based infant sleep intervention: cluster-randomized, controlled trial.” Pediatrics 122(3): e621-627.
 Middlemiss, W., D. A. Granger, et al. (2012). “Asynchrony of mother-infant hypothalamic-pituitary-adrenal axis activity following extinction of infant crying responses induced during the transition to sleep.” Early Hum Dev 88(4): 227-232.