Thursday, December 23, 2010

Sleep Research Part 2: Effectiveness of Controlled Crying

There is possibly no more emotive topic for new parents than controlled crying. People feel very strongly, and everyone feels they have to defend their position. I will lay my cards on the table and say that I don't personally use a crying method, but this post is not about telling you whether controlled crying is right or wrong. Instead, it summarises the studies on the effectiveness of controlled crying, to help you make your own informed decision about whether it's worth trying.

I was just going to have one post on controlled crying, but I have too much information so it looks like I will need to divide it up. This post is therefore just on the effectiveness of cc methods. The next post will be on the risks of crying methods.

You may also be interested in Sleep Research Part 1: What is a normal sleep?

What is 'controlled crying', exactly?

For some people this means letting your child whinge for a bit in their cot, for others it means full on screaming for a set time or until the child falls asleep. So that we're not confusing apples with oranges, here is the terminology I will use:

  • extinction - leaving a child to cry until they fall asleep. This is the accepted term in paediatric journals. Extinction refers to the extinction of the presence of mum or dad to help the baby fall asleep. This method is sometimes also known as letting a baby 'cry it out'. These days this method is associated with Marc Weissbluth .
  • graduated extinction - leaving a child to cry for a short time (say 5 minutes), then returning for a minute to comfort the child without getting them out of the cot, then leaving them again, and so on until they fall asleep. The time-controlled leaving and returning involved in this method is why it is known as 'controlled crying'. It was devised by Richard Ferber . Tizzie Hall also suggests this method. Tizzie's version involves leaving the child for the same interval of time each time. Ferber's involves leaving the child for progressively longer intervals of time.
  • communication-sensitive extinction - leaving a child to whinge / grizzle / sob a bit to see if they go to sleep but intervening if their sounds escalate into screaming – this method involves trying to decipher different kinds of unhappy noises and ignoring some while responding to others. Jo Ryan and Sheyne Rowley suggest this method. I have made up the name because it doesn't appear to have an agreed upon name. My impression from reading medical research journals is that medical professionals would consider 'controlled crying' to be the graduated extinction method proposed by Ferber, my anecdotal experience is that many parents in Australia practice communication-sensitive extinction and call it 'controlled crying' (and say that those who really let their baby scream are 'doing it wrong') while others are adamant that communication-sensitive extinction is not 'controlled crying'.
  • self-settling - a general term used to describe settling techniques where a parent leaves a child to fall asleep without assistance, with or without crying.
  • active settling - a general term used to describe active, hands on methods of helping a baby go to sleep, such as rocking, patting, and feeding the baby to sleep.
  • settling by movement - using the movement of a car / pram etc to get the baby to sleep - sometimes considered a kind of active settling, but sometimes separated into a different category because it does not involve the child being able to see / touch the parent.
  • social settling - mostly used with slightly older babies and children, involves settling child to sleep through quiet play / reading / talking / singing lullabies etc.

As you can see, a variety of crying and non-crying settling methods are used. The communication-sensitive extinction method hasn't really been studied (that I could find), although it clearly comes generally under the heading of baby sleep strategies I will refer to as self-settling strategies.

How effective are crying methods?

Extinction and graduated extinction are the most studied methods of improving a child's ability to go to sleep and stay asleep independently of parental assistance. Many of the studies have been conducted without control groups, so it is impossible to tell whether the sleep training method made a difference or the children just grew out of the problem. However, in the few controlled studies that have been done, they do appear to have a significant but by no means guaranteed success rate. Here are some examples where the success rate was quantified:

  • A small study of 2 - 4.5 year olds used a combination of a set bedtime, a bedtime routine, and then ignoring the child. If an older child came out of the bedroom, parents were required to smack the child once and return him or her to the bedroom without talking. Parental compliance was recorded with audiotape. 7-15 months later, half the children had improved sleep, but as they were not compared with a control group it is hard to say whether the method made things better, worse, or had no effect at all (Rapoff).
  • A study of 208 9 month – 4 year olds found that a set bedtime, a bedtime routine, and ignoring all crying 'significantly reduced' sleep problems in 78% of children at 3 months after the method was started. Parents recorded what they did with a diary (Seymour et al).
  • In an Australian study of about 150 8-10 month olds, half the mothers were given one-on-one instruction in how to do controlled crying. Two months later, 70% of the mothers who were taught controlled crying said their sleep problems were resolved, as compared to 47% of those not taught controlled crying. Success here meant the mothers responded 'no' when asked whether their baby had a sleep problem, not a more objective measure of sleep improvement.  After four months, the success of the controlled crying group had fallen slightly to 64%, and of those not taught crying 51% indicated their sleep problems had resolved. 12% of the mothers taught controlled crying said they used it 'all the time', 65% said they used it 'most of the time', and 13% used it 'about half the time' (Hiscock and Wake).

    Two key difficulties with interpreting these results is: a) we don't know whether the mothers using controlled crying all or most of the time were the only ones whose children's sleep improved, and b) controlled crying was not the only strategy taught (although it was the main one). If we assume that the mothers using controlled crying all or most of the time where the only ones whose children's sleep improved (the most favourable scenario for controlled crying), then controlled crying had an 87% success rate for those that tried it.  Interestingly, 2-3 years later, 69% of mothers from this study reported that their infants' sleep problems had resolved – and there was no difference in the amount of parents with a problem for the controlled crying group as compared to the the non-controlled crying group (Lam, Hiscock and Wake).
  • 33 children aged between 6 months and 4 years with night waking problems were divided into groups, and one group tried extinction. The average number of night wakings in this group was 2 per week after extinction, as opposed to the group who did nothing, which had an average of 8 night wakings per week. I do not have information on how many children in each group saw an improvement (Rickert and Johnson cited in Ramchandani et al).

In some cases I was able to obtain the original studies mentioned above, but in others I relied on two major academic overviews of studies on sleep training methods – which show the above examples are typical and some of the strongest available in support of controlled crying methods (Mindell; Ramchandani et al).

As you can see, the studies do not suggest crying is an 100% effective solution. What seems to happen is the authors report that the method was 'effective' in their summary or abstract because it had a statistically significant effect, and this is simply repeated by those giving sleep advice to parents without looking at the size of the effect. This leads to parents being given the impression that crying methods always work, and they rationalise that if they didn't work for someone that parent must not have 'done them right'.

But the truth is that if there is a 'right' method that works for all babies, no one has yet established what it is – which makes it pretty difficult for you to implement it with your baby. Crying methods in general (when repeated as necessary and even when combined with a bedtime routine) appear to be effective in the short-term for only 50-80% of babies / children over 6 months of age. This is not to be sneezed at, but it's not as compelling as you may have been led to believe. In the long term, it is doubtful whether crying methods are any more effective as non-crying methods for babies, as sleep problems during the first year do not seem to have much bearing on sleep problems down the track.

Persistence of sleep problems

One of the reasons parents try crying methods is that they are warned that if they do not, they will be setting themselves up for years of trouble. But how true is this?

Numerous studies now suggest that only a very small percentage of children have persistent sleep problems:

  • An Australian study of 483 infants found that most sleep problems in the first 2 years of life are transient. Only 6.4% of children had persistent sleep problems for those 2 years (Wake et al).
  • In a study of 308 mothers and their babies at 8 months, 10 months, and 3 years, only about 8% reported a sleep problem at both 8 months and 3 years of age. 89 of the 308 children had sleep problems at 3 years of age, but only 23 of those of these had been babies with sleep problems at 10 months (Zuckermen et al).
  • A longitudinal study of 83 children followed from birth to 4 years of age found that sleep problems in early childhood bore no relationship to which children were self-soothers before 12 months of age (Gaylor et al).
  • An academic overview of various other studies suggest that most children naturally grow out of waking by five years of age at the latest, and that along the way they grow into and grow out of sleep problems (Middlemiss).
  • A large study of Japanese children found that 80% developed a habit of frequent night-time waking and crying between 18-21 months, but by preschool age, less than 20% of these children still had sleep-related night time crying (Fukumizu).

One of the studies that is repeatedly cited in academic sleep articles as indicating a sleep problem will be persistent unless addressed by methods like controlled crying studied a group of 60 children once and then once again three years later (Sudesh et al). This study found that 87% of the children with sleep problems at the start of the study had them three years later. However, in interpreting these results it is important to recognise that this was a small sample, and the children were all toddlers or preschoolers at the start of the study – the youngest was 15 months and the oldest was 4 years old. Of the 23 children with persistent sleep problems, over half had environmental stresses at the time of follow up, such as the unexpected absence of the mother, or illness or accident. The definition of a 'sleep disturbance' was 3 or more bedtime struggles in a week where it took over an hour to settle the child. The study did not consider whether parents had sought to address sleep problems with any kind of sleep training method.

Those 'bad habits' of rocking / feeding / patting your baby to sleep

A very comprehensive British study investigated how much of a difference settling techniques made to how well children sleep by following 259 children for a year (Morrell and Cortina-Borja). These children were between 12-19 months at the start of the study. What they found was that using more active settling did increase the probability of your child having a sleep problem, but not by very much. They found that most parents of babies without sleep problems used a variety of settling methods, including active settling on average about 30% of the time. When they combined data on the child's age, the settling technique of the parent, the parents attitude and level of emotion about sleep, and the baby's temperament, these factors ALTOGETHER only correctly predicted sleep problems for 2 out of 5 babies. Settling technique alone predicted even less.

In a large American survey of over 5000 parents of children between 0-36 months, babies who were breastfed or bottlefed back to sleep, slept in the same room as their parents, or had an irregular bedtime routine did have more night wakings. But again these factors IN TOTAL only explained 20% of the statistical variance between children with less night wakings and more night wakings. In other words, for 4 out of 5 children, these 'bad habits' appeared to have no effect on whether they were night wakers or not (Sadeh et al).

The worst case of the 'bad habits' causing problems is made by a very large Canadian study of 1741 children who were followed longitudinally at 5 months, 17 months, and 29 months, it was found that at 5 months, about 1 in 4 children did not sleep 6 consecutive hours. At 17 months only 7% of children did not sleep 6 consecutive hours, and at 29 months, 10% of children did not sleep 6 consecutive hours . At 6 months, infants were more likely to sleep for less than 6 consecutive hours if they were breastfed. At 17 months of age, children who were actively rocked/patted by their parents were 4.5 times more likely not to sleep 6 hours in a row, but this was only about 2 times as likely in children at 5 months or at 29 months. At 5 months, only 43% of the children who slept 6 hours in a row were left to self-settle (with or without crying), but at 17 months and 29 months, about 75% of the children who slept 6 hours in a row were left to self-settle (Touchette et al). It should be noted that this study was conducted using only a parental questionnaire, and this method has been shown to have significant discrepancies when compared to sleep diaries and direct observation.

Judging whether a method 'works'

In the majority of studies, the effectiveness of controlled crying (or any other method) is judged by parent's reports. But parents only report the night wakings they are aware of. If your aim is to reduce disruptions to your sleep, then this is an effective way of measuring what works. But if you are concerned to train your baby to sleep well (for example, because you think this has health benefits or will improve their IQ), then you need to know whether the method actually improves the baby's sleep.

After sleep training, parents may believe their child is sleeping continuously at night when the child is actually awake for a substantial amount of time. In a study of 59 kindergarten children, where sleep was monitored by use of an actigraph on 4-5 consecutive nights, parents reported the children woke on average 0.5 times a night whereas the actigraphs revealed the children woke an average of 2.7 times a night. 29% of the children slept for less than 90% of the time the parents believed they were sleeping, and 41% had significantly fragmented sleep. Nearly all children woke up at least once a night (Tikotzky and Sadeh). These children had successfully learned not to wake their parents at night for the most part, but that did not mean they were getting a good night's sleep.

What age is controlled crying effective?

There is scant evidence on the effectiveness of crying methods before 6 months. The studies on the effectiveness of sleep training tend to be on babies 6 months and older.

One would expect them not to be effective for the majority of babies before 6 months because the child has limited capacity to appreciate or learn from what is happening.

I found one study of trying sleep training (exact method unspecified) in very young babies - namely 3 to 12 week old babies. It found that only 10% more of those using the method slept 5 hours a night by 12 weeks of age when compared to those who hadn't (St James-Roberts and Gillham). This is a fairly small difference for the discomfort involved. In babies under 12 weeks of age, babies do not learn not to cry by being ignored – except for babies with colic ('persistent criers'), prompt parental attention has been found to signficiantly reduce the amount of crying - for babies with significant persistent crying, nothing helped (St James-Roberts et al).

After 6 months there appears to be a limited window of time before the toddler stage, when the child is capable of both getting out of bed and a more complex appreciation of their world, factors such as family discord or psychological insecurities are major factors in sleep disruption – it is not just a matter of teaching children to break habits of parent-assisted settling (Smaldone).

Controlled crying is most effective when it is used through the night

For sleep training to be effective at improving settling and night waking, it is important that parents consistently use the method during the night, and not just at bedtime. A small intensive New Zealand study using video recording and parental diaries to ensure parents adhered strictly to the program looked at the effect of controlled crying (graduated extinction) for seven children. When the controlled crying was done only at bedtime one child rapidly learned to self-settle except when ill, two children showed increases in time to settle, one child occasionally settled but mostly did not, and one showed a steady reduction in the time to settle. However, when the parents chose to use controlled crying every time the child woke up at night, as well as during bedtime, there was a noticeable effect on all seven children. (Healey et al).

Coming up next...

Potential risks and side-effects of controlled crying. After that I will look at the effectiveness of non-crying methods.


Fukumizu et al, 'Sleep-Related Nighttime Crying (Yonaki) in Japan: A Community-Based Study' (2005) Pediatrics Vol 115(1), p217.

Gaylor et al, 'A Longitudinal Follow-Up Study of Young Children's Sleep Patterns Using a Developmental Classification System' Behavioural Sleep Medicine Vol 3, p44.

Healey et al, 'Treating sleep disturbance in infants: What generalizes?' (2009) Behavioural Interventions Vol 24, p23.

Hiscock and Wake, 'Randomised controlled trial of behavioural infant sleep intervention to improve infant sleep and maternal mood' (2002) British Medical Journal Vol 324, p1062.

Lam, Hiscock and Wake, 'Outcomes of Infant Sleep Problems: A Longitudinal Study of Sleep, Behavior, and Maternal Wellbeing' (2003) Pediatrics, Vol 111, p203.

Middlemiss, 'Infant sleep: a review of normative and problematic sleep and interventions' (2004) Early Child Development and Care Vol 174, p99.

Mindell, 'Empirically Supported Treatments in Pediatric Psychology: Bedtime Refusal and Night Wakings in Young Children' (1999) Journal of Pediatric Psychology Vol 24 (6), p465.

Morrell and Cortina-Borja, 'The Developmental Change in Strategies Parents Employ to Settle Young Children to Sleep, and their Relationship to Infant Sleeping Problems, as Assessed by a New Questionnaire: the Parental Interactive Bedtime Behaviour Scale' (2002) Infant and Child Development Vol 11, p17.

Ramchandani et al, 'A systematic review of treatments for settling problems and night waking in young children' (2000) British Medical Journal Vol 320, p209.

Rapoff et al, 'The management of common childhood bedtime problems by pediatric nurse practitioners' (1982) Journal of Pediatric Psychology Vol 7, p179.

Sadeh et al, 'Sleep and sleep ecology in the first 3 years: a web-based study' (2009) Journal of Sleep Research Vol 18, p60.

Seymour et al, 'Management of night-waking in young children' (1983) Australian Journal of Family Therapy Vol 4., p217.

Smaldone et al, 'Sleepless in America: Inadequate Sleep and Relationships to Health and Well-being of Our Nation's Children' (2007) Pediatrics Vol 119, pS29.

St James-Roberts and Gillham, 'Use of a behavioural programme in the first 3 months to prevent infant crying and sleeping problems' (2001) Journal of Paediatrics and Child Health, Vol 37(3) 289.

St James-Roberts et al, 'Objective confirmation of crying durations in infants referred for excessive crying' (1993) Arch Dis Child Vol 68, p82.

Sudesh et al, 'Persistence of sleep disturbances in preschool children' (1987) The Journal of Pediatrics Vol 110(4), p642.

Tikotzky and Sadeh, 'Sleep Patterns and Sleep Disruptions in Kindergarten Children' (2001) Journal of Clinical Child & Adolescent Psychology Vol 30(4), p581.

Touchette et al, 'Factors Associated With Fragmented Sleep at Night Across Early Childhood' (2005) Archives of Pediatric Adolescent Medicine, Vol 159, p242.

Wake et al, 'Prevalence, Stability, and Outcomes of Cry-Fuss and Sleep Problems in the First 2 Years of Life: Prospective Community-Based Study' (2006) Pediatrics Vol 117, p836.

Zuckerman et al, 'Sleep Problems in Early Childhood: Continuities, Predictive Factors, and Behavioral Correlates' (1987) 80(5) Pediatrics 664.

1 comment:

  1. After a few months, I tried controlled crying for a whole week (with my first baby), after my health centre nurse and a lot of helpful mothers told me that the trouble I was having with a waking baby, was because I had not tried this method for long enough. The studies showed xyz if you follow xyz procedures !
    I was instructed to put my baby in their cot, close the door to the room, and wait a minimum of 20 mins - and to do this for a week. I persisted as I was desperate to stop the frequent night waking. It did not work. It was a sad and painful experience for me, and given that my baby never did cry themselves to sleep more than a couple of times, (but endured 20 mins of misery/fear every time) ....I decided that it was not kind or helpful in our case and stopped. But my second baby slept beautifully right from birth, (oh heaven), but at about 3 years of age...developed sleep problems which went on for many years !!