Website of the Department of Health

Aim/Principal Research Question
The primary aim was to measure the effects of a behavioural intervention on infant crying and sleeping behaviour in the first three months compared with a potentially low cost educational intervention and existing service provision within the NHS. An economic evaluation was designed to estimate the annual cost to the NHS of infant crying and sleeping problems in the first 12 weeks of life and to assess the cost effectiveness of interventions aimed to prevent infant crying and sleeping problems relative to the usual services.

Factors of Interest
The behavioural intervention encouraged mothers to accentuate day/night differences so that babies learned to regulate sleep and waking behaviour according to environmental cues. This involved introducing a late evening focal feed, trying not to rock, hold or feed babies to sleep, to settle them in a darkened environment and to minimise interaction during the night.

The educational intervention comprised an information booklet which incorporated current best practice related to managing/preventing infant crying and promoting good sleeping patterns. This was prepared in collaboration with Health Visitors. Mothers were also provided with a dedicated telephone help-line organised by a voluntary support group -CRYSIS.

The control group mothers and babies received the standard community services offered by Health Visitors and GPs as did all the trial participants.

Methods
A randomised controlled trial in which mother and baby pairs were allocated to one of 3 policies. Randomisation was delayed until the mother was at home in order to allow time for recovery following childbirth and to minimise the risk of mergence between the policies.

Sample Groups
Mothers were eligible for inclusion if they gave birth to their babies at the Royal Berkshire Hospital, Reading, they lived in West Berkshire or South Oxfordshire, were English speaking and literate and owned a telephone: their babies were 37 weeks gestation and had not been admitted to Special Care Baby Unit or shown evidence of congenital abnormality. Mothers were provided with information about the study during pregnancy and again before they left the hospital (usually within 24 hrs of delivery). 7-10 days later they were contacted by telephone to confirm their wish to participate. Formal consent and entry to the study took place during a subsequent home visit between day 10-14. Once entered into the study, mothers were requested to comply with their allocated policy for a minimum of 12 weeks. 610 mothers participated during the period March-November 1997.

Outcome Measures
The primary outcome measures were the number of 'interruption-free nights' per week and measures of babies' crying. An 'interruption-free night' was defined as a night in which parents reported their baby to remain asleep for 5 hours or more between 10pm and 6am. This definition reflects video-recorded evidence that most infants wake in the night but return to sleep without waking their parents. The focus here was on instances where infant night-waking was detected by parents, usually because of crying, since it is parents who seek NHS help for these problems. Secondary measures included mothers' general well-being, approach to baby-care and perception of their babies' temperaments. Maternal well-being was measured at 6 and 12 weeks using the Edinburgh Postnatal Depression Scale. Throughout the first 12 weeks, contact with Health Visitors and/or GPs regarding infants'crying/sleeping problems were recorded; in addition, the babies' weight gain, health and use of medications were monitored. At 9 months mothers' reported whether or not their babies had a regular bedtime routine and whether they had sought professional advice because of their babies' crying or sleeping problems.

Findings
Of the 3403 women who gave birth, 2663 met the eligibility criteria. Most women not entered into the trial were either unable to be approached as they had been discharged home within the first few hours of childbirth before they could be visited on the post natal wards (942 [35%]) or they exercised their right not to participate when contacted within the first 10 days following delivery. Random allocation of the remaining 610 mothers generated groups which were similar in terms of descriptive variables. All three groups showed a rapid increase in the number of 'interruption-free nights particularly between weeks 3 and 9 but at 11 and 12 weeks of age infants and parents in the behavioural group had significantly more interruption-free nights than infants in the educational or control groups (p0.05). The perceived amounts of time the babies spent crying and fussing did not differ between the groups at any age. In the first weeks, the babies spent around 2 hours a day fussing and crying: this gradually reduced to around 1 hour in 24 by 3 months of age. The daily peak of crying time was between 6pm and midnight. Babies' gender, and birth order had little effect on the findings. There was no difference between the groups in the mothers' mean EPDS scores or the number who reported high scores. Those who scored 13 or above were contacted and offered the choice of referral to the health visitor or GP. In all cases mothers had already contacted the support services or accepted the offer of referral. Neither of the interventions had a significant effect on the infants' weight gain over the first 3 months. At 6 and 12 weeks, mothers in the behavioural group were more likely to report using a structured approach to infant care (p=0.02 and 0.001, respectively). At 12 weeks, mothers in the behavioural group rated their approach to care more convenient and easy to implement than mothers in the other groups (p=0.02). There was no effect of the interventions on the mothers' ratings of their enjoyment in looking after their babies or of their babies' temperaments. At 9 months, 10% more mothers in the behavioural than the educational or control groups reported that their baby had a regular bedtime routine. At this time, fewer mothers in the behavioural group reported that they had contacted their health visitors because of infant crying or sleeping problems.

Conclusions
The findings suggest that, if introduced routinely, the behavioural intervention used in this study, could lead to an improvement of around 10% in the number of babies who regularly remain settled for 5 hours a night by 12 weeks of age. Furthermore, parents liked the 'common sense' approach and clear structure of the routine, which they did not find overly rigid or prescriptive but which enhanced their confidence in their ability to cope. One of the limitations of the study was the number of mothers who were eligible but not recruited largely because they were discharged home so quickly following delivery before they could be contacted on the postnatal wards. This illustrates one of the challenges of conducting such a pragmatic trial where clinical circumstances can dramatically influence the day-to-day recruitment to such trials.

Implications for Further Research
The need for further studies which explore whether a behavioural intervention might be more effective with groups considered to be at higher risk of developing infant crying/sleeping problems for example mothers 20 years of age or those who are without the support of a partner.

Publications
A randomised controlled trial to compare alternative strategies for preventing infant crying and sleep problems in the first 12 weeks. (The COSI Study). Accepted by Primary Care R&D for publication in January 2002.

Morris S, St James-Roberts I, Sleep J, Gillham P. Economic evaluation of strategies for managing crying and sleeping problems. Archives of Disease in Childhood; January 2001, Vol 84, No 1, p 15-19.

Use of behavioural program in the first three months to prevent infant crying and sleeping problems. Accepted by Paediatrics 2001.

Contact
Professor Jennifer Sleep
Centre for Evidence-based Practice
Thames Valley University
Royal Berkshire Hospital
Reading RG1 5AN
Tel: 01225 336527
Fax: As above