Argument in favor of sleep training
Hiscock et colleagues conducted a cluster-randomized, controlled trial in 2008, Australia[1]. The aim of the study is to determine long-term effects of behavioural intervention for infant sleep problems on maternal depression and parenting style, as well as on child mental health and sleep, when the children reached 2 years of age.
The study is conducted in Maternal and Child Health Centers (clusters) which were randomly assigned either control or study groups. Mothers who reported a sleep problem composed the study sample. From 174 families allocated to intervention group, at the end of 2 years data of 143 families is available. The sample size is large enough to represent the population. All babies were 8 months old when the intervention began. The intervention group used either graduated extinction or adult fading. At the end of 2 years mother’s depressive scores decreased and parenting styles assessed by ‘Parent Behaviour Checklist’ didn’t change. Child mental health assessed by ‘Child Behaviour Checklist for children aged 18 months to 5 years’ didn’t differ between study and control groups. It is concluded that the intervention didn’t result in any adverse effects in the long term on either mothers’ parenting practices or children’s mental health. The strengths of this study are the large sample size, homogeneous sample regarding to age, randomization and including a control group, long-term follow up, a well-designed randomization flow chart and including data of lost to follow-up group and comparison data of this group according to confounding variables. The main limitation of the study is that inclusion and outcomes were based on parent report measures. Behavioural sleep interventions are considered as treatments but the diagnose of the problem/disease/disturbance is not based on any kind of clear scientific criteria. However this is the most common problem in most of other studies. Also, intervention group couldn’t be blinded so the outcome measures could be biased towards positive. Another limitation which is not unique for this study is that the interventioncouldn’t be observed, it is assumed that all parents respected to recommeded protocol. However use of hospital for this purpose wouldn’t be much better because some other confounding variables would emerge this time. It is very hard to overcome this obstacle.
This setting for Price study is 49 maternal and child health centers and local government areas in Australia, 2003-2009[2]. Price et colleagues aimed to examine long term benefits and harms of behavioural infant sleep interventions. It is a cluster controlled, single blind, randomized trial. Interventions (controlled comforting and camping out) were applied to 8 months old 122 infants. There were 103 infants in the control group. Sample size is large enough to represent population. After 6 years from interventions the study group showed no significant difference with regard to emotion and behavior (measured by 25-item Strengths and Difficulties Questionnaire), sleep, psychosocial health-related quality of life and stress (assessed by salivary cortisol). Large sample size and randomized controlled study design, homogeneity of the sample and very long follow-up period are strengths of the study. It is clearly can be concluded that behavioural sleep techniques have no marked long- lasting effects in children. Limitations are the same as the above article. Inclusion and outcomes are based on parent reports and no direct observation of the interventions.
A recent randomized controlled trial aimed to evaluate the effects of behavioural interventions on the sleep/wakefulness of infants, parent and infant stress, and later child emotional/behavioural problems and parent/child attachment.[3] Parents contacted researchers in response to advertisements at pediatric outpatient clinics and child care centers. 43 infants (6-16 months) were enrolled and 29 were allocated to study group and 14 to control group. Sample size is not large enough to represent population. Interventions were graduated extinction and bedtime fading. Parent reported sleep diaries, salivary cortisol samples, strange situation process (to assess attachment), assessment of children’s emotional and behavioural problems (by questionnaires completed by mothers) were used. At the 12-month follow-up no significant differences were found in emotional and behavioural problems and in attachment styles between groups. The strange situation procedure is the golden method to assess attachment. Therefore this article provides a clear scientific evidence that 1 year after the intervention there were no harmful effect of behavioural sleep interventions. Strengths of the study 1) It is the only article, as far as I know assessing attachment in the long-term. 2)The study design is randomized and controlled. The limitation of the study is the relatively small sample size. The study is very important that it presents no marked difference between infant cortisol samples in the control and intervention group. This finding is in contradiction with one of the main CIO against research[4].
Honaker et colleagues aimed to describe parental practices implementing behavioural sleep intervention outside a clinical setting[5]. Parents (652) were recruited through a facebook group and completed an online survey about their experience using behavioural sleep interventions (BSI). Sample size is large enough to represent population. It is not a controlled study. On average parents implemented the intervention when their infant was 5.6 (more or less 2.77) months. The majority of parents report successfully implementing the intervention with reduced crying by the end of 1 week and success within 2 weeks. The strength of the study is very large sample size. The limitations of the study 1) Data were collected over internet and participants recruited through a facebook group. 2) There is no randomization and control groups. However it is interesting that the age of the infants (5.6 months) were very young with comparison to other studies[1, 2]. There are mixed results about the effectiveness of BSI under 6 months. Therefore the study is important to show effectiveness of BSI under 6 months.
Argument against CIO
Middlemiss et colleagues aimed to examine change in the synchrony between mothers’ and infants’ physiology. 25 infants between 4 to 10 months were enrolled.[4] Sample size is not large enough to represent population. It is not a controlled study. It is a 4 days in-residence, hospital-based programme. They measure cortisol in saliva and infants’ signaling and distress behaviours (nurses documented). Results show that by the end of third day while infants’ signaling and distress behaviours decrease significantly, saliva cortisol remained the same. However cortisol level of mothers’ significantly decreased. The study design is problematic beacuse it takes place in hospital. In the article it is written that ‘many of the mothers reported a lack of support in caring for their infants at home’. There is a possibility that cortisol levels of mothers were decreased because they simply get support for caring their infants from nurses. Without a control group (spending 3 nights in hospital with the same amount of contact with healthcare professionals) we can’t exclude this possibility. Also we know many parents can feel some kind of guilt in the process of sleep training. By being in the hospital and having healthcare professionals around would have helped to share responsibility and overcome that feeling. Again it is a possibility that we can’t exclude without a control group (applying the same procedure at home). Saliva cortisol found to be same in infants after 3 days. It is said to be higher but we don’t know according to what it was higher. All nighttime care and naptime care were provided by nurses. It may have resulted in some kind of maternal deprivation for the infants. A stranger puts the infants for sleep every time instead of their mothers’. Also change of environment may have resulted in some kind of stress in the infants. The study has very low scientific evidence with no randomization, control group, small sample size and problems in design which I mentioned above.
Middlemiss and colleagues aimed to examine the effects of a response-based intervention on total sleep duration, maternal/infant saliva cortisol levels and maternal depression, anxiety and stress.[6] The study is not controlled. Participants were provided from parents who attended to a public residential family-care facility. The participants are not randomly assigned. Of the 54 participants, 20 left the study after participation and data of the remaining 34 is used. Study sample size is not large enough to represent population. Babies were between 4 to 11 months, with an average of 7,16 months. The intervention focused on education about sleep signs, maintaining bedtime routines and not allowing the baby to cry by responding with low level support care not exceeding 15 to 30 seconds. However if the time was not enough parents encouraged to sooth their babies until crying ends. Interventions took place in hospital and lasted 3 days (some participants reporting significant sleep problems in their infants stayed 5 days/4 nights). Cortisol saliva samples were taken at home before intervention and during the intervention 3 times (wake, pre sleep and post sleep-after 20 minutes). Data were collected through questionnaires, sleep log and saliva cortisol in the 9 days after intervention. According to the results of the study, total sleep time were increased, post sleep saliva cortisol of infants were higher in day 2 and 3 according to home levels. At home TST of depressive mothers’ infants were lower. No significant difference were found according to anxiety scores. It is concluded that the research provides evidence to support changes in infant sleep patterns based on building parents’ awareness of the importance of responding to infant cues and behaviours. The study lacks randomization, control group and a sample size large enough to present population. Also it is again conducted in hospital and the same confounding factors as the first study[4] couldn’t be excluded.
Maute hypothesized that ignoring children’s bedtime crying (ICBC) is more frequently used by parents with Western-oriented parental beliefs (assessed by an instrument that the researchers developed) and with high levels of stress (measured by Parental Stress Scale) and also aimed to investigate whether the use of ICBC is predicted by low parental sensitivity (measured by a vignette from Situation-Reaction-Questionnaire) and a difficult childhood temperament. (measured by Parenting Stress Index)[7]. The study were conducted in Germany and participants were 586 Swiss and German parents with children between zero and 4 years old, who participated in an online survey. They were paid for participation. The sample size is large enough to represent population. However participants were not randomized, online survey method and payment could have decreased the reliability of data collected. There were no control groups. According to their result parent’s western oriented beliefs predicted ICBC, low sensitivity parents as well as parents of children with a difficult
temperament used ICBC more frequently, parental stress didn’t predict ICBC. It is concluded that ICBC has become part of Western childrearing tradition and shouldn’t be applied because children could suffer from high stress and anxiety in ICBC.
The study was published in 2018 and its link was published on the Web sites of parenting magazines. I think the results have magazinish value instead of scientific value. First of all the measure used to assess Western –Oriented – Beliefs haven’t been examined in terms of reliability and validity before. We don’t know if it really measures what it claims to measure. Therefore the measurement has no scientific value. Second they used one vignette to assess parental sensitivity. Again we don’t have a clue whether it really examines parental sensitivity at all. The vignette is ‘You are very busy with something that needs to be done urgently. Just a few seconds ago you looked after your child, but now your child iS loudly drawing attention to itself.’ Then the parents were labelled as sensitive or insensitive according to their response to this vignette. How does that measures parental sensitivity? It just measures the behaviour of a parent incase her child cry loudly to draw attention. If a mother is using a behavioural sleep intervention, surely she won’t respond. I think they have asked something measuring ICBC but labeled it as measuring parental sensitivity. After that they draw a totally wrong conclusion that low sensitivity parents used ICBC more frequently. There are other many problems and limitations with regard to the study design. The way data collected (online survey) and participants being paid and not randomized harm the reliability of the data. Also we don’t know anything about the age the intervention applied, the details of the procedure. Another conclusion is that parents who used ICBC had rated their children being more difficult in terms of adaptation and mood. Is it the reason they applied ICBC or the result? We don’t have any data to clarify this but the writer interpreted as it is the result of ICBC procedure.
I focused on studies examining long-term effects of sleep training on babies because there is some evidence that cry it out method is effective in improving infants’ sleep and even people who are against the method admits that it is effective[1, 2, 8-10]. I encountered some randomized controlled trials. In a large sample sized and 5 year follow-up trial Price and colleagues didn’t find any adverse effects regarding with emotion, behaviour and stress. Gradisar showed that the babies who had sleep training showed no significant differences from the control group regarding with attachment styles, emotional and behavioural problems. Although the sample size is limited, this study has successfully assessed attachment by strange situation procedure. It is important because people who are against claim that CIO method leads to insecure attachment. However there is no clinical trial in favor of this hypothesis. There is clearly more scientific evidence in favor of CIO method.
One of the most well-known article against CIO is Middlemiss study in 2011 but it has some methodological limitations. First of all it is not a controlled study and participants were not randomly assigned. Second, it took place in hospital which means cortisol levels of infants may have stayed same because of the stress of novel environment or mother levels of cortisol may have decreased because they got help from nurses. Third there is no data of the mothers who left the study to compare with the ones who completed. Also cortisol levels were just measured for 3 days and no data in the long-term. Middlemiss has conducted a new research in 2017, presenting response-based intervention. However, it is not controlled and participants aren’t randomly selected. Except for the total sleep duration of infants, other findings are not very consistent. I found out there were some articles about supposed harmful effects of sleep training but no direct scientific evidence exists. I run into a meta-analysis concluding that behavioural sleep interventions in the first six months of life don’t improve outcomes for mothers or infants.[11] However again the heterogeneity of studies regarding age of infant and the procedure applied is confusing.
To sum up, argument in favor of sleep training has the stronger argument because there are more studies with larger sample size, randomized controlled study design and long-term follow up data. There are no controlled, randomized and studies with long-term follow up data for CIO against argument.
REFERENCES
Hiscock, H., et al., Long-term mother and child mental health effects of a population- based infant sleep intervention: cluster-randomized, controlled trial. Pediatrics, 2008. 122(3): p. e621-e627.
Price, A.M., et al., Five-year follow-up of harms and benefits of behavioral infant sleep intervention: randomized trial. Pediatrics, 2012: p. peds. 2011-3467.
Gradisar, M., et al., Behavioral interventions for infant sleep problems: a randomized controlled trial. Pediatrics, 2016: p. e20151486.
Middlemiss, W., et al., Asynchrony of mother–infant hypothalamic–pituitary–adrenal axis activity following extinction of infant crying responses induced during the transition to sleep. Early human development, 2012. 88(4): p. 227-232.
Honaker, S.M., et al., Real-World Implementation of Infant Behavioral Sleep Interventions: Results of a Parental Survey. The Journal of pediatrics, 2018.
Middlemiss, W., et al., Response-based sleep intervention: Helping infants sleep without making them cry. Early human development, 2017. 108: p. 49-57.
Maute, M. and S. Perren, IGNORING CHILDREN’S BEDTIME CRYING: THE POWER OF
WESTERN‐ORIENTED BELIEFS. Infant mental health journal, 2018. 39(2): p. 220-230.
Hiscock, H. and M. Wake, Randomised controlled trial of behavioural infant sleep intervention to improve infant sleep and maternal mood. Bmj, 2002. 324(7345): p. 1062.
Lam, P., H. Hiscock, and M. Wake, Outcomes of infant sleep problems: a longitudinal study of sleep, behavior, and maternal well-being. Pediatrics, 2003. 111(3): p. e203- e207.
Behavioral treatment of bedtime problems and night wakings in infants and young children. Sleep, 2006. 29(10): p. 1263-1276.
Douglas, P.S. and P.S. Hill, Behavioral sleep interventions in the first six months of life do not improve outcomes for mothers or infants: a systematic review. Journal of Developmental & Behavioral Pediatrics, 2013. 34(7): p. 497-507.