The literature published since 1970 (Mindell et al 2006; Basis (unpublished)) provides considerable evidence that sleep training in general produces desired outcomes on one or more parameter of interest such as parental sleep, infant sleep, night-waking, infant signalling, maternal mental health -- and/or other outcomes including child behaviour and parental well-being). Mindell and colleagues found that 94% of the studies they reviewed reported "clinically significant reductions in bedtime resistance and night wakings" (2006:1269). In our own review 83% of papers presented a positive result on one or more measures.
The aim of the sleep management programmes that have been empirically tested can be broadly divided into two groups - one aiming to prevent the development of (parentally perceived) sleep problems, and another that aims to reduce or treat existing infant sleep problems (as defined by parents).
Approaches for young infants
The former target very young infants -- in our review all studies (n=3) focused on infants under 6 months of age with participants recruited in the immediate postnatal period. With one exception (Symon et al 2005) recent preventative interventions have not utilised extinction (CC or CIO) as a primary technique, focussing instead on emphasising day/night distinction and encouraging settling to sleep and self-soothing in a cot without parental assistance. Such techniques appear to encourage early settling and prolonged sleep periods, but may increase SIDS risk and have detrimental effects on breastfeeding initiation and maintenance. These concerns about potential risks associated with some sleep training techniques are echoed by a larger systematic review of this literature conducted by Douglas and Hill (2013). Of the 43 articles Douglas and Hill reviewed, behavioral sleep interventions in the first 6 months of life did not improve outcomes for mothers or infants and risked unintended outcomes. These include increased amounts of problem crying, premature cessation of breastfeeding, worsened maternal anxiety, and, if the infant is required to sleep either day or night in a room separate from the caregiver, an increased risk of SIDS. More recently, Stremler and colleagues' (2013) randomised trial found that the effects of a behavioural-educational sleep promotion intervention were ineffective in improving maternal and infant sleep or other health outcomes in the first months postpartum. Hiscock's (2014) randomised trial found that providing parents with education about sleep in the early postnatal period, including information about normal sleep/crying behaviour again had little effect, with the exception of infants who were classified as frequent feeders (more than 11 feeds/24 hours). The mothers of infants in this group showed reduced postnatal depression symptoms, and reported their infants had fewer sleep and crying problems. This suggests that mothers who find their infant's night-time waking the most difficult to cope would benefit from information that helps them to realistically calibrate their expectations.
Sleep training techniques
Of the specific methods for sleep training that aim to reduce 'sleep problems' (usually an improvement involves increased infant and/or maternal sleep duration, promotion of self settling/self-soothing, reduction of night wakings and/or reduced 'signalling' (crying) during night waking), the majority of work has examined extinction methods (encompassing unmodified extinction, graduated extinction and extinction with parental presence). The research evidence reviewed by Mindell et al (2006) provides strong support for the efficacy of these methods in terms of the above aims: 19, 14 and 4 studies respectively found the above forms of extinction to be effective, but stressful for parents.
The studies reviewed by Mindell et al also lend support for positive routines (2 studies found this method to be rapid and effective) and scheduled awakenings (4 studies indicated this method is probably effective, but takes longer and is more complicated to implement). Bed-time routines and positive reinforcement may also be effective in reducing sleep problems (14 and 15 studies), however as they tend to be employed in conjunction with other behavioural management techniques, it is hard to disentangle the effective elements of such interventions. The same problem arises with parental education/ prevention interventions (5 studies - results were "statistically significant [and] clinically meaningful" (Mindell et al 2006:1268)), elements of which are often combined with other behavioural strategies for infant sleep management.
In our own review, 9 unique interventions that utilised some variety of extinction were examined -- graduated extinction predominated (8 studies) but unmodified extinction (1 study) and extinction with parental presence (4 studies) were also represented. Infants were aged between 0 and 45 months at the time of the intervention. All reported an initial improvement in recorded outcomes (these vary among studies but include objective or subjective report of infant sleep/wake duration, sleep latency, frequency and duration of waking, and/or parental perception of the existence of a sleep problem). Of interventions not utilising an extinction element (n=3), two (both 'preventative' in aim, and involving infants <12 weeks of age) reported 'positive' outcomes including increased maternal and infant sleep duration, reduced infant wakings, and greater odds of sleeping for more than 5 hours in a single bout.
More recent research by Gradisar (2016) and a review by Crichton and Symon (2016) claims that graduated extinction and bedtime fading (involving gradually altering bedtimes) provide significant sleep benefits, with no adverse stress responses or long-term effects on parent-child attachment or child emotions and behavior. However, these findings are controversial. Crichton and Symon’s review has been criticised for displaying unexamined assumptions and biased interpretation of data. Gradisar’s study has mainly come under fire for the very small participant sample size and high dropout rate, meaning that the results should be considered as preliminary and should not be relied on for recommendations.In both Mindell's review and our own, several studies found that the intervention tested was more effective when used with mothers who were depressed at the start of the intervention; some also reported that maternal depression improved alongside improvements in infant sleep outcomes.
Hiscock & Wake (2002) conducted a large randomised trial and found that the overall improvement observed in maternal report of infant sleep problems at 2 months post-intervention, in response to controlled crying/camping out when infants were 8-10 months old, was due to significant improvement in the sub-group of participants that included depressed mothers. There was no significant improvement in the 'non-depressed' group. Follow-up also indicated that at 4 months and 2-3 years post-intervention (Lam et al 2003) there were no differences in perceived sleep problems between intervention and control groups, regardless of depression status, and no differences in family functioning or coping, based on perception of sleep problems, at the latter time-point.
Mothers exhibiting depression therefore seem to benefit most from the implementation of sleep training techniques. The picture is complicated, however, because mothers experiencing depression are more likely to perceive that their child has a sleep problem than are non-depressed mothers. More research is needed on the actual sleep patterns of the infants of depressed mothers.