7Maintaining Infant Health

Publication Details

7.1. Recommendations

  1. Healthy babies should have normal colour for their ethnicity, maintain a stable body temperature, and pass urine and stools at regular intervals. They initiate feeds, suck well on the breast (or bottle) and settle between feeds. They are not excessively irritable, tense, sleepy or floppy. The vital signs of a healthy baby should fall within the following ranges:
    • respiratory rate normally 30 60 breaths per minute
    • heart rate normally between 100 and 160 beats per minute in a newborn
    • temperature in a normal room environment of around 37°C ( if measured).
    • [D(GPP)]
  2. At each postnatal contact parents should be offered information and guidance to enable them to:
    • assess their baby’s general condition
    • identify signs and symptoms of common health problems seen in babies
    • contact a healthcare professional or emergency service if required. [D(GPP)]
  3. Parents, family members and carers should be offered information and reassurance on:
    • their baby’s social capabilities as this can promote parent–baby attachment (in the first 24 hours)
    • the availability, access and aims of all postnatal peer, statutory and voluntary groups and organisations in their local community (within 2–8 weeks) [D(GPP)]
  4. Both parents should be encouraged to be present during any physical examination of their baby to promote participation of both parents in the care of their baby and enable them to learn more about their baby’s needs. [D(GPP)]

Parenting and emotional attachment


Assessment for emotional attachment should be carried out at each postnatal contact. [D(GPP)]


Home visits should be used as an opportunity to promote parent or mother-to -baby emotional attachment. [B]


Women should be encouraged to develop social networks as this promotes positive maternal-baby interaction. [B]


Group based parent-training programmes designed to promote emotional attachment and improve parenting skills should be available to parents who wish to access them. [A]


Health care providers should offer fathers information and support in adjusting to their new role and responsibilities within the family unit. [GPP]

Physical examination and screening


The aims of any physical examination should be fully explained and the results shared with the parents and recorded in the postnatal plan and the personal child health record. [D(GPP)]


A complete examination of the baby should take place within 72 hours of birth. This examination should incorporate a review of parental concerns and the baby’s medical history should also be reviewed including: family, maternal, antenatal and perinatal history; fetal, neonatal and infant history including any previously plotted birth-weight and head circumference; whether the baby has passed meconium and urine (and urine stream in a boy). Appropriate recommendations made by the NHS National Screening Committee should also be carried out. (www​.nsc.nhs.uk/ch_screen/child_ind.htm).

A physical examination should also be carried out. This should include checking the baby’s:

  • appearance including colour, breathing, behaviour, activity and posture
  • head (including fontanelles), face, nose, mouth including palate, ears, neck and general symmetry of head and facial features. Measure and plot head circumference
  • eyes; check opacities and red reflex
  • neck and clavicles, limbs, hands, feet and digits; assessing proportions and symmetry
  • heart; check position, heart rate, rhythm and sounds, murmurs and femoral pulse volume
  • lungs; check effort, rate and lung sounds
  • abdomen; check shape and palpate to identify any organomegaly; also check condition of umbilical cord
  • genitalia and anus; check for completeness and patency and undescended testes in males
  • spine; inspect and palpate bony structures and check integrity of the skin
  • skin; note colour and texture as well as any birthmarks or rashes
  • central nervous system; observe tone, behaviour, movements and posture. Elicit newborn reflexes only if concerned
  • hips; check symmetry of the limbs and skin folds (perform Barlow and Ortolani’s manoeuvres)
  • cry; note sound
  • weight; measure and plot.
  • [D(GPP)]

The newborn blood spot test should be offered to parents when their infants are 5–8 days of age. [D(GPP)]


At 6–8 weeks, an examination, comprising the items listed in item 11, should be carried out. In addition, an assessment of social smiling and visual fixing and following should also be carried out. [D(GPP)]


A hearing screen should be completed before discharge from hospital or by week 4 in the hospital programme or by week 5 in the community programme. [D(GPP)]


Parents should be offered routine immunisations for their baby according to the schedule recommended by the Department of Health. [D(GPP)]

Physical health and well being



Parents should be advised to contact their healthcare professional if their baby is jaundiced, their jaundice is worsening, or their baby is passing pale stools. [D(GPP)]


Babies who develop jaundice within the first 24 hours after birth should be evaluated. (emergency action). [D(GPP)]


If jaundice develops in babies aged 24 hours and older, its intensity should be monitored and systematically recorded along with the baby’s overall well-being with particular regard to hydration and alertness. [D(GPP)]


The mother of a breastfed baby who has signs of jaundice should be actively encouraged to breastfeed frequently, and the baby awakened to feed if necessary. [D(GPP)]


Breastfed babies should not be routinely supplemented with formula, water or dextrose water. [D(GPP)]


If a baby is significantly jaundiced or appears unwell, evaluation of the serum bilirubin level should be carried out. [D(GPP)]


If jaundice first develops after 7 days or a baby remains jaundiced after 14 days in an otherwise healthy baby and a cause has not already been identified, it should be evaluated (urgent action). [D(GPP)]



Parents should be advised that cleansing agents should not be added to a baby’s bath water nor should lotions or medicated wipes be used. The only cleansing agent suggested, where it is needed, is a mild non-perfumed soap. [D(GPP)]


Parents should be advised how to keep the umbilical cord clean and dry and that antiseptics should not be used. [A]



If thrush is identified in the baby, the breastfeeding woman should be offered information and guidance about relevant hygiene practices. [D(GPP)]


Thrush should be treated with an appropriate antifungal medication if the symptoms are causing pain to the woman or the baby or feeding concerns to either. [D(GPP)]


If thrush is non-symptomatic, the woman should be advised that anti-fungal treatment is not required. [D(GPP)]

Nappy rash


For babies with nappy rash the following possible causes should be considered:

  • hygiene and skin care
  • sensitivity to detergents, fabric softeners or external products that have contact to the skin
  • presence of infection. [D(GPP)]

If painful nappy rash persists it is usually caused by thrush and treatment with anti fungal treatment should be considered. [C]


If after a course of treatment the rash does not resolve, it should be evaluated further (non-urgent action). D(GPP)]



If a baby has not passed meconium within 24 hours, the baby should be evaluated to determine the cause which may be related to feeding patterns or underlying pathology (emergency action). [D(GPP)]


If a baby is constipated and is formula fed the following should be evaluated (urgent action) [D(GPP)]

  • feed preparation technique
  • quantity of fluid taken
  • frequency of feeding
  • composition of feed.



A baby who is experiencing increased frequency and/or looser stools than usual should be evaluated (urgent action). [D(GPP)]



A baby who is crying excessively and inconsolably, most often during the evening, either drawing its knees up to its abdomen or arching its back, should be assessed for an underlying cause, including infant colic (urgent action). [D(GPP)]


Assessment of excessive and inconsolable crying should include: [D(GPP)]

  • general health of the baby
  • antenatal and perinatal history
  • onset and length of crying
  • nature of the stools
  • feeding assessment
  • woman’s diet if breastfeeding
  • family history of allergy
  • parent’s response to the baby’s crying
  • any factors which lessen or worsen the crying.

Health care professionals should reassure parents of a healthy baby with colic, that the baby is not rejecting them and that colic is usually a phase that will pass. Parents should be advised that holding the baby through the crying episode, and accessing peer support may be helpful. [D(GPP)]


Use of hypoallergenic formula in bottle fed babies should be considered for treating colic, but only under medical guidance. [A]


Dicycloverine (dicyclomine) should not be used in the treatment of colic due to side effects such as breathing difficulties and coma. [A]



The temperature of a baby does not need to be taken, unless there are specific risk factors, for example maternal pyrexia during labour. [D(GPP)]


When a baby is suspected of being unwell, the temperature should be measured using electronic devices which have been properly calibrated and are used appropriately. (Cross refer to Fever in Children guideline) [C]


A temperature of 38°C or more is abnormal and the cause should be evaluated urgently (emergency action). A full assessment, including physical examination should be undertaken. [D(GPP)]

Vitamin K


All parents should be offered vitamin K prophylaxis for their babies to prevent the rare but serious and sometimes fatal disorder of vitamin K deficiency bleeding. [A]


Vitamin K should be administered as a single dose of 1 mg intramuscularly as this is the most clinically and cost effective method of administration. [A]


If parents decline intramuscular Vitamin K for their baby, oral Vitamin K should be offered as second line option. Parents should be advised that oral Vitamin K must be given according to the manufacturers instructions for clinical efficacy and will require multiple doses. [D(GPP)]



All home visits should be used as an opportunity to assess relevant safety issues for all family members in the home and environment and promote safety education. [A]


The healthcare professional should promote the correct use of basic safety equipment, including, for example, infant seats and smoke alarms and facilitate access to local schemes for provision of safety equipment. [A]


Parents should be given information in line with the Department of Health guidance about sudden infant death syndrome (SIDS) and co-sleeping (Reduce the risk of cot death, November 2005) which states that “The safest place for your baby to sleep is in a cot in your room for the first six months. While it’s lovely to have your baby with you for a cuddle or a feed, it’s safest to put your baby back in their cot before you go to sleep. There is also a risk that you might roll over in your sleep and suffocate your baby, or that your baby could get caught between the wall and the bed, or could roll out of an adult bed and be injured.” [B]


Parents should be advised never to sleep on a sofa or armchair with their babies. [B]


If parents choose to share a bed with their infant, they should be advised that there is an increased risk of SIDS, especially when the baby is less than 11 weeks old, if either parent: [B]

  • is a smoker
  • has recently drunk any alcohol
  • has taken medication or drugs that make them sleep more heavily
  • is very tired.

If a baby has become accustomed to using a pacifier (dummy) while sleeping, it should not be stopped suddenly during the first 26 weeks. [B]

Child abuse


Health professionals should be alert to risk factors and signs and symptoms of child abuse. [D(GPP)]


If there is raised concern, the healthcare professional should follow local child protection policies. [D(GPP)]

Table 7-1. Infant health core information and advice.

Table 7-1

Infant health core information and advice.

7.2. Evidence Statements for Infant Health

Note: Each section is linked to the relevant narrative for ease of reference

Emotional Attachment and Parenting

(see narrative)

  1. Providing information to new parents appears to enhance attachment and parenting knowledge and interaction with their infants. [Level 2+]
  2. A variety of strategies may be effective in providing parenting information. These include group based training programmes, home visiting programmes, videos shown in hospital, and one-to-one interventions. [Level 2+]
  3. Close physical contact through use of soft infant carriers or kangaroo techniques may promote maternal infant attachment. [Level 2+]

4. Physical Examinations and Screening

(see narrative)


There is no high level evidence base for the conduct and content of the physical examination of the newborn. Therefore the recommendations for this exam are based upon expert opinion and good practice. They incorporate the screening measures recommended by the National Screening Committee.


There is no robust evidence regarding the optimal frequency of growth monitoring in newborns

7. Physical Health and Well being

(see narrative)


Routine supplementation of nondehydrated breastfed infants with mild jaundice with water or dextrose water is not recommended. [Level 4]


Jaundice before 24 hours of age is always considered pathological and requires further evaluation. [Level 4]


Late breast milk jaundice develops much less commonly at around, 4–7 days after birth and peaks day 7–15. [Level 4]


No evidence was identified for general care of infant skin conditions.


Keeping the cord clean and dry is as effective as treatment with antiseptics and antiseptics appear to prolong time to cord separation. [Level 1+]


Optimal antifungal treatment of thrush remains to be determined.


Candida albicans appears to be associated with severe napkin dermatitis. [Level 2+]


No research studies were identified which addressed constipation in the newborn.


The Paediatric Accident and Emergency Research Group based at the University of Nottingham have defined diarrhoea as, “…a change in bowel habit for the individual child resulting in substantially more frequent and/or looser stools.” [Level 4]


Hypoallergenic formula appears to be an effective treatment for colic in formula fed babies. [Level 1++]


Dicycloverine (dicyclomine) may be effective for colic but can cause breathing difficulties and coma and is not recommended for infants. [Level 1++]


Several studies demonstrate that tympanic thermometry is not as accurate as axillary or rectal temperatures. [Level 2+]


For the purposes of vaccine research fever in infants has been defined as a rectal temperature of > 38° C. [Level 4]

Vitamin K

(see narrative)


In light of available evidence it does not appear that there is a link between childhood cancer and IM Vitamin K prophylaxis. [Level 1+]


A single IM dose of 1mg Vitamin K appears to be effective prophylaxis for both early and late VKDB. [Level 1+]


If oral Vitamin K is given, multiple doses are required for adequate protection of breastfed infants against late VKDB. [Level 1+]


The exact dosage and timing of oral Vitamin K administration after the delivery dose has not been determined. [Level 2+]


The intramuscular route of Vitamin K administration appears to be more clinically effective and more cost effective. [Level 1+]


(see narrative)


Parent education appears to lead to behaviour change. [Level 1]


Home visiting appears to have a positive effect on prevention of childhood injury. [Level 1]


Most accidents involving babies occur in the home. Fires and falls are the most common causes of accidents in this age group.[Level 2]


Social deprivation is a risk factor for childhood accidents.[Level 2]


Supplemental strategies may be needed in deprived areas to ensure safer homes. [Level 4]


Child car restraint legislation and education appear to change behaviour and reduce injury.[Level 1]


Well designed smoke detector programmes which include implementation appear to reduce injury and promote behaviour change.[Level 1]


General community prevention initiatives may lead to both injury prevention and behaviour change.[Level 1]

Reducing the risk of SIDS

(see narrative)


Prone sleeping position, head covering, and maternal smoking appear to be the most consistent risk factors for SIDS. [Level 2++]


Co-sleeping is a risk for SIDS if infants are less than 11 weeks old [Level 2++]


A change in behaviour from the habitual use of a pacifier may be a risk factor for SIDS. [Level 2+]

Child Abuse

(see narrative)


There do not appear to be any screening tools for child abuse with adequate sensitivity, specificity and positive predictive value to assess for risk of maltreatment with sufficient accuracy. [Level 2++]

Table 7-2. Infant health core information and advice.

Table 7-2

Infant health core information and advice.

7.3. The healthy baby

A healthy baby should have normal colour for his/her ethnicity, maintain a stable body temperature, pass urine and open his/her bowels at regular intervals. A healthy baby initiates feeds, sucks well on the breast (or bottle) and settles between feeds. A healthy baby is not excessively irritable or tense and is not excessively sleepy or floppy. The vital signs of a healthy baby should fall within the following ranges:

  • Respiratory rate normally 30–60 breaths per minute
  • Pulse rate, normally between 100–160 in a newborn
  • Temperature in a normal room environment of around 37 degrees C (if measured)

7.4. Emotional Attachment and Parenting Are there interventions which promote attachment/bonding in the postpartum period?

Attachment theory postulates that the formation of secure attachment depends upon the mother’s sensitive responsiveness to the infant’s signals to provide the context in which the infant’s experiences and feeling of security are organized. The relationship is bi-directional however and the infant’s behaviour is not only influenced by the mother but it may also affect the mother’s behaviour. More distal factors may also affect the quality of the development of attachment, including maternal personality and social support. The quality of parental attachment is believed to be a major predictor of long-term outcomes including childhood psychopathology, behavioural difficulties or developmental delay (Armstrong et al. 2000).

An early meta-analysis (Turley 1985) was undertaken to assess experimental controlled studies which provided information to mothers regarding the sensory and perceptual capabilities of their newborns in an effort to determine whether the intervention influenced maternal-infant interaction. A second purpose was to assess the effect of location where the information was provided. The results showed that providing information to mothers produced positive mean effect sizes for all outcome categories, maternal and infant responsiveness, maternal knowledge, confidence, attitude and sensitivity (no p values provided). The effect of location was significant (p=0.009). Those studies in which the information was presented in the home setting tended to produce the largest mean effect size in terms of maternal-infant interaction. Thus, according to this study, the post discharge period would appear to be the most effective time for health professionals to support mothers in discovering their infant’s feelings, responses and abilities.

A meta-analysis undertaken by a research team from The Netherlands evaluated 70 published studies presenting 88 intervention effects (Bakermans-Kranenburg, Van Ijzendoorn, & Juffer 2003). Studies were not limited to RCTs. As study analyses used various statistical tests, outcomes were recomputed and transformed into Cohen’s d. A core set of 51 randomized control studies was identified, presenting data on 6282 mothers and infants. This core set of studies appeared to be significantly effective in enhancing maternal sensitivity (p<.001). Non randomised studies seemed to run the risk of inflated effects. Interventions focusing on sensitivity only were more effective than other types of interventions (p=0.03). Unexpectedly, the four studies that did not use personal contact as a means of intervening tended to show the largest effect size. These studies relied on the provision of soft baby carriers or introducing the kangaroo method, a workbook on responsiveness or a videotape. The effect of interventions conducted at parents’ homes was not significantly different from the effect of interventions conducted elsewhere. In studies on attachment (n=23) the effect size was small but significant (p<.05). Sensitivity focused interventions significantly affected infant attachment security (<.01). The most effective interventions did not always use a large number of sessions nor did they necessarily start early in life or even before birth.

The Anisfeld et al study (Anisfeld et al. 1990) was included in the Bakermans-Kranenburg et al meta-analysis described above. This small American study (n=49) randomised mothers to soft baby carriers or to infant seats. Follow up was by questionnaire at 2 months and by video tape, Bayley Scales of Infant Development, the Carey Infant Temperament Scale and Product Use Questionnaire. A final assessment at 13 months consisted of the Ainsworth Strange Situation, the Lifestyle Questionnaire II and the Product Evaluation Questionnaire. There were significantly more securely attached infants in the baby carrier group (p=0.019). Fifteen of 16 high users of the carriers had infants who were securely attached. Of the seven moderate/low users, four were securely attached.

An intervention designed to influence mothers’ sensitive responsiveness toward their infant was carried out in Brazil, in collaboration with London Guildhall University (Wendland-Carro, Piccinini, & Millar 1999). Thirty eight mothers and their newborns were randomly assigned either intervention group one or group two. On day 2 or 3 postpartum the intervention group one viewed an “enhancement video” on newborn competence to interact, affectionate handling of the infant and which encouraged mothers to explore and interact with their infants. Group two saw an alternative video on basic care giving skills and infant health issues. One month later, observations were undertaken in the homes of study participants. Free play and bathing were observed and videotaped. The observations were examined for the frequency of synchronous and asynchronous co-occurrences, that is, reciprocal exchanges between mother and infant. Overall synchrony-asynchrony scores were higher for the enhancement group compared to the basic skill control group (p<0.01). For the enhancement group there were more co-occurrences involving infant vocalization with mother’s reciprocal vocalization, smiling, soothing, and stimulation compared to the control group.

A study to assess the effect of maternal social support on attachment was carried out among low income primiparas in Michigan (Jacobson & Frye 1991). Forty six women were recruited and randomly assigned to intervention or control. The intervention group was assigned a “volunteer coach” who visited them in the home periodically until 12 months postpartum to talk about their pregnancy and delivery and early infant care. At the fourteen month assessment infants of mothers in the experimental group were rated as more securely attached than control on the Attachment Ratings measure (Waters and Deane’s Q sort procedure, 1985), p<0.005.

A meta-analysis of studies that evaluated the relationship between social support and maternal behaviours toward the baby was carried out in 1992 in the U.S. (Andresen & Telleen 1992). They included 66 studies. The study designs were not reported. The majority of the studies sampled white, middle income, married mothers. A wide variety of measures were used to operationalize social support. The types of social support most frequently measured were emotional and material support. Emotional support indicated supportive actions that conveyed empathy, caring, love, and trust. Material support included the physical resources needed to cope with the demands of parenting such as financial assistance and help with child care. Six meta-analyses of these studies showed significant relationships to exist between emotional support and maternal behaviours (CI .03–.60) and material support and maternal behaviours (CI .05–.61). However, some statisticians would argue that the use of meta analysis for any studies except RCTs is inappropriate and that it is not possible to adjust for the bias inherent in less robust study designs. The variability of interventions and outcome measures as well as the homogeneity of the sample make generalisability of this analysis questionable.

Another meta-analysis of studies investigated the relationship between social support and adolescent mothers’ interaction with their infants. This analysis included 13 studies, most of which were cross sectional in design (Clemmens 2001). As noted above, the use of meta-analysis for any studies except RCTs is statistically inappropriate as adjustment for bias in the study design is not possible. This meta-analysis showed a significant relationship between social support of adolescent mothers and their interactions with their infant (r=.30, CI .239–.379). However, these results should be cautiously interpreted.

Are there interventions which promote parenting in the postpartum period?

A “review of reviews” was carried out by the Health Development Agency (HAD) in 2004 (American College of Obstetricians and Gynecologists 1998). In systematic reviews which looked at a number of measures of parenting or mother-child interaction the evidence suggests that home visiting can produce positive effects on various dimensions of parent child interaction. The HAD reviewers recommended further work in evaluating which types of programmes are likely to replicate these impacts.

A recently published study (Puura et al. 2005) evaluated the outcome of the European Early Promotion Project. This study was conducted in five countries: Cyprus, Finland, Greece, Serbia and the UK. In this study a group of primary health care nurses were trained to support mothers with newborn babies and to intervene early if there were problems that might interfere with the babies’ care. A comparison group was seen by a different group of nurses who had no additional training. These mothers received standard care. An initial assessment using the HOME inventory, a detailed interview and observational methods was done at 6–8 weeks. Outcomes were measured at 24 months for each country. In the UK no significant differences in change scores were found on the interview ratings. On the HOME inventory there were significant changes with mothers in the intervention group showing positive gains in verbal and emotion responsiveness (p=0.02), organisation of the child’s environment (p=0.004), provision of appropriate play material and in the total HOME score (p=0.003).

A Cochrane review of group based parent training programmes aimed to establish the efficacy of such programmes and to assess the role of parent education in the primary prevention of emotional and behavioural problems (Barlow & Coren 2003). Five RCTs were included in the review and a meta-analysis was conducted. The meta-analysis of the independent observations of children’s emotional and behavioural adjustment showed a significant result favouring the intervention group (ES −0.54, CI −0.84—0.23). The limited evidence available concerning the extent to which these results are maintained over time is equivocal.

A systematic review to assess the effectiveness of home visiting programmes on parenting and the quality of the home environment was carried out by Kendrick et al (Kendrick et al. 2000) of the University of Nottingham. Thirty four studies met inclusion criteria. Twenty six used participants considered at risk of adverse maternal or child health outcomes; two used preterm or low birth weight infants and two used infants with failure to thrive. Only eight used participants not considered to be at risk of adverse child health outcomes. Parenting and the quality of the home environment were measured in a variety of ways in these studies. The most commonly used measure was the Home Observation for Measurement of the Environment (HOME) used in 17 of the studies. The infant-toddler version of the inventory consists of six subscales measuring aspects of the quality of the home environment in relation to parenting. Twenty seven studies reported other measures of parenting and 10 studies reported both HOME scores and other measures of parenting. Twelve of 17 studies reporting HOME scores were included in a meta analysis. Fourteen effect sizes were extracted from the 12 studies and entered into the meta analysis. A highly significant result was obtained suggesting home visiting was effective in improving the quality of the home environment as measured by the HOME score. Despite the limitations of their review, the authors concluded that home visiting programmes were associated with an increase in the quality of the home environment as measured using the HOME scale and in improving parenting using a range of other measures. Most of the studies reviewed used professional home visitors, most commonly nurses. Eight studies used lay workers and the results of these studies appeared similar to those using professional visitors. Only four of the studies used UK health visitors.

A Cochrane review of parenting programmes for improving psychosocial outcomes for teenage parents and their children was undertaken by Coren and Barlow (Coren & Barlow 2001). Only four randomised controlled trials were included in this analysis. The programmes reviewed were offered either antenatally or postpartum, in a variety of settings and were not combined with a home visiting programme. Despite methodological problems and the heterogeneity of the studies the authors concluded that the results indicated that parenting programmes may be effective in improving a range of outcomes for both teenage parents and their infants including mother-infant interaction, language development, parental attitudes, parental knowledge and maternal mealtime communication.

In 2003 Coren et al published an expanded review of both individual and group based parenting programmes for teenage mothers (Coren, Barlow, & Stewart-Brown 2003). Fourteen studies which used varying study designs including RCTs, controlled studies and one-group design were reviewed. Despite considerable diversity in the parenting programmes and study designs, there was evidence to show that parenting programmes are effective in improving maternal sensitivity, identity, self-confidence and the infants’ responsiveness to their parents. Only one study directly compared a group based programme with a programme delivered on a one-to-one basis. The results of this study showed that the group-based programme produced more changes than the individual programme, especially in the case of high-risk mothers.

An Australian study (Armstrong, Fraser, Dadds, & Morris 2000) aimed to assess the effectiveness of a home-based intervention for at risk families which was focused on promoting the quality of the parent-infant interaction. A total of 181 families were randomised to intervention, a series of home visits from a child health nurse, or control. At a 4 month follow up 160 families were available for assessment. A statistically significant difference was shown on all subscales as well as the total HOME score (p<0.05).

A Canadian systematic review (Wade et al. 1999) evaluated the evidence for the effectiveness of peer/paraprofessional one–to-one interventions in promoting positive parental and child health and developmental outcomes. Seventeen of the 21 included studies targeted high risk populations. The evidence showed that peers/paraprofessionals could have a positive impact on child development and parent child interaction whether embedded in a multifaceted programme or when intervening independently. There was also some evidence that high intensity (weekly or bi-weekly visits for at least one year) interventions begun in infancy have both an immediate and long term positive impact on child development.

Attachment and fatherhood:

A meta synthesis of ten published articles focusing on the experiences of fathers of healthy infants was reviewed (Goodman 2005). Four themes were identified in this synthesis, one of them being ‘confronting reality.’ Fathers stated that their expectations for involvement in fatherhood and bonding proved unrealistic due to the immaturity and unresponsiveness of the infant and the men’s lack of time to spend with the infants. Men also felt that their relationship with their infant was slower to develop than the mothers’. Many fathers, especially fathers of breastfed infants expressed a feeling of exclusion from the strong mother infant relationship. Feeling excluded from providing care to the infant was perceived as inhibiting the development of the father-infant relationship.

A further qualitative study was identified which described in depth interviews with 18 new fathers at 6–12 weeks after the birth of their child (St John, Cameron, & McVeigh 2005). Fathers found their new role to be challenging and complex. Although most fathers focused on making time to interact with their newborns, barriers such as the need to work, the wake/sleep/cry patterns of the newborn and the time spent sharing the newborn with family and friends meant that some men deferred developing their relationship with their newborn baby.

Another recent study (Bolzan, Gale, & Dudley 2004) reported the findings of 40 interviews which explored men’s postnatal mental health. A very high proportion of men expressed the desire to father differently from their own fathers (72.5%). Most men in the sample (73%) were in a position of juggling both their desire to be involved fathers and the demands of full-time work and careers. They reported experiencing tension and contradictions between these competing demands on their time. The researchers conclude that increasing workplace flexibility and provisions such as parental leave are important for men’s postnatal mental health.

In 1998 the National Childbirth Trust sent questionnaires to a randomly selected sample of pregnant women throughout the United Kingdom. These women were asked to pass the questionnaire on to their partners to complete. Thirty seven percent (n=817) of the questionnaires were returned. Approximately six months later a follow up questionnaire was sent to the men who responded originally. The response rate was 57%. The sample was broadly representative of British national paternity trends.

Most men felt they had enough time together as a family immediately after the birth. However, 20% of men felt unable to be as involved as they would like because they were not welcomed to stay by the hospital or had to go home to care for other children. Half of the men said that they were ‘completely involved’ in caring for their child but 40% of men said that work commitments reduced their involvement in family life.

Most men whose child was breastfed thought that breastfeeding was convenient and gave the baby special comfort and security. However 20% of men said that breastfeeding made them feel left out and two fifths would have liked to be able to feed the baby.

Specific questions about fathers’ emotional attachment to their babies were not asked in the questionnaires.

7.5. Physical Examinations and Screening

7.5.1. Physical Examination of the Newborn

A basic physical examination of the newborn is routinely performed immediately after birth and prior to admission to the postnatal area whether the baby is born in hospital or at home (ref intrapartum care guideline in development). A more thorough physical examination of the newborn within 72 hours after delivery is accepted as good practice and recommended by the National Screening Committee. The dual purpose of this examination is to confirm normality, thereby reassuring parents and carers, and to identify and act upon any abnormalities. A review of the literature for the conduct of this exam was undertaken. No systematic reviews of the newborn physical exam which met NICE criteria were identified. A review of discreet components of the physical exam was not carried out for the Postpartum Care Guideline. The guidance for the newborn exam is based upon a review of expert opinion, supported by existing scientific evidence, rather than a systematic review. Three documents underpin the recommendations for this exam: The Best Practice Statement for Routine Examination of the Newborn published in April 2004 by NHS Quality Improvement Scotland (NHS Quality Improvement Scotland. 2004); Postpartum care of the mother and newborn: a practical guide published in 1998 by WHO (World Health Organization. 2005); and Textbook of Neonatology (Rennie 2005). National Screening Committee recommendations are incorporated into this examination. The Guidelines Development Group agreed that where a practice is considered valuable by experts and when such a practice is not contraindicated by the evidence and causes no harm to the infant it should be included in the exam.

Any assessment of the infant should be performed with the mother’s consent and, where possible, with the mother present. The setting should ensure comfort and safety for mother and baby, with measures taken to prevent cross infection. The environment should also provide privacy during and after the examination when confidential information may be discussed. Family/maternal concerns should be ascertained and discussed, with particular reference to family history and antenatal history. All aspects of the assessment should be explained. If for any reason the mother cannot be present, other family members should be involved and the mother should be made aware of the findings as soon as possible.

The health care professional performing the exam should be appropriately trained and his/her competency assessed, possibly with the use of simulators, as conditions such as congenital hip dislocation and congenital cataracts are rare. The health care professional should maintain competency through continued practice and review. A consistent approach between the examination and the advice offered to parents should be ensured. If any abnormalities are found on examination, this should be communicated to the parents and appropriate treatment and/or referral should be instituted. The results of the examination should be recorded in the PCHR and any other baby health record.

The initial examination is a screening procedure, which marks the commencement of ongoing child health surveillance. Continued observation of the baby during the first weeks of life by both the health care professional and the infant’s carers is important. In addition, as with all other contacts with health care professionals, it is an opportunity for health promotion and education on a range of areas relevant to maintaining infant health (jaundice, vitamin K, sleeping positions, hearing screening, nutrition, hygiene, breastfeeding, cot death prevention, safe transport in cars and maternal depression). This examination also provides an opportunity to address broader psychosocial issues (mental health, substance abuse, and smoking) as well as spiritual and cultural needs.

The physical examination should include the following:

  • A review of family, maternal and perinatal history
  • A review of previously plotted birthweight and head circumference
  • A check of whether the baby has passed meconium and urine (enquiring about urine stream in a boy) and documentation of failure to do so for follow up by midwifery staff
  • Observation of the baby’s general condition including colour, breathing, behaviour, activity and posture
  • Ascertainment of whether parents or carers have anxieties and an opportunity to observe mother-infant interaction
  • Enquiring about method of infant feeding, and if mother or staff have any concerns about feeding to date. If necessary breastfeeding should be observed, and mother assisted with this, provided the healthcare professional is competent to do so.
  • Examination of the exposed parts of the baby first: scalp, head (including fontanelles), face, nose, mouth including palate, ears, neck and general symmetry of head and facial features
  • Examination of the baby’s eyes (size, position, absence of discharge) including with an ophthalmoscope and test for the ‘red reflex’
  • If exposed, examination of the baby’s neck and clavicles, limbs, hands, feet and digits, assessing proportions and symmetry. Undress baby to complete the exam
  • Assessment of the baby’s cardiovascular system – colour, heart rate, rhythm and femoral pulse volume as well as listening to the heart for a murmur, and checking laterality disturbance of heart
  • Respiratory effort and rate can be assessed simultaneously with the cardiovascular assessment and listening to air entry
  • Observation of the baby’s abdomen – colour, shape and palpate to identify any organomegaly, and examination of the condition of the umbilical cord
  • Observation of the baby’s genitalia and anus, to check completeness and patency. Check for undescended testes in male infants.
  • Inspection of the bony structures and skin of the baby’s spine, with the baby prone
  • Noting the colour and texture of the skin as well as any birthmarks or rashes
  • Observation of the tone, behaviour, movements, and posture to complete the assessment of the central nervous system (CNS)
  • If concerned, undertake more detailed neurological examination e.g. eliciting newborn reflexes
  • Regarding hips, check symmetry of the limbs and skin folds. Perform Barlow and Ortolani’s manoeuvres on a firm flat surface. The baby should be calm when the examination is performed.
  • Noting sound of baby’s cry
  • Consideration of any specific known risks in the baby’s home, and alerting appropriate professionals to parents who may have problems in caring for their baby
  • Ensuring that parents know how to assess their baby’s general condition and to contact a midwife, health visitor, doctor or emergency services if required

7.5.2. National Screening Committee (NSC)

The National Screening Committee recommends that all pre-screening information for parents should be in acceptable formats and provided appropriately to support decision making.

The NSC (www.nsc.nhs.uk) recommendations are noted below:

Examination: (first 72 hours)

  • Eyes – opacities, including cataracts (red reflex)
  • Hips – dysplasia (Barlow & Ortolani)
  • Heart – congenital disease (Cyanosis, tachypnoea, murmur, femoral pulses)
  • Testes - cryptorchidism


  • A hearing screen should be completed prior to discharge or by week 4 in the hospital programme or by week 5 in the community programme..
  • Newborn blood spot; 5–8 days. ( Phenylketonuria [PKU], Congenital Hypothyroidism [CHT] and Sickle Cell Disease [SCD], with Cystic Fibrosis [CF] and Medium Chain Acyl CoA Dehydrogenase Deficiency [MCADD] in some areas)
  • Examination: (6–8 weeks)
  • Eyes – opacities, including cataracts (red reflex)
  • Hips – dysplasia (Barlow & Ortolani)
  • Heart – congenital disease (Cyanosis, tachypnoea, murmur, femoral pulses)
  • Testes - cryptorchidism
  • Immunisation, 8 weeks:
  • DTaP/IPV/Hib and Men C

7.5.3. Physical Examination at 6–8 weeks

The second postpartum examination at 6–8 weeks might be performed concurrently with administration of the first set of immunisations. The Department of Health recommendations for immunisations should be followed, with parental consent. The 6–8 examination should repeat the assessments made at the physical examination of the newborn, and also include an assessment of whether the baby has a social smile and is fixing and following visually. Any parental concerns should be addressed. Documentation of the examination should be made in the PCHR and the baby’s medical record.

The competencies for this examination are therefore unchanged from those of the newborn examination.

7.5.4. Growth monitoring

Routine growth monitoring, including height (length) and weight measurements are widely accepted and standard components of child health surveillance. Growth monitoring has been used as a ‘screening tool’ for the identification of disorders and diseases which may affect growth. As there were no nationally set norms for growth monitoring in the U.K., the Child Growth Foundation organized a meeting of paediatricians, endocrinologists, public health doctors, GPs and nurses from various disciplines in July, 1998 to consider whether growth monitoring actually fulfilled the criteria required of a screening programme (Elliman & Hall 2003). After examining the evidence regarding the potential for detection of children with growth abnormalities and looking carefully at the issues of measurement error, guidelines for clinical practice were developed. Future debate and research was also encouraged. The participants in this meeting, known as the ‘Coventry Consensus,’ recommended the following policies for normal newborns:

  1. Birth weight, correctly taken and recorded, and related to gestational age, is an essential first step in growth monitoring
  2. Measuring the length of the normal neonate is not part of a national programme of growth monitoring
  3. No justification has been found for the routine monitoring of length in the first two years of life.

The group also considered further weight monitoring. They recognized that weight gain in the first few weeks of life is occasionally a cause for concern but found little evidence on optimal frequency of weighing in the neonate. During the first year of life frequent weighing was seen to be undesirable as short term fluctuations in weight may simply increase parental anxiety and stress. Routine attendance at the GP surgery at ages 2, 3, 4, 8 and 12 months (immunization and surveillance contacts) was seen to be an opportunistic time to weigh a child if indicated. However, there was no evidence that regular weighing of a baby who is healthy and thriving is of any benefit. Weight monitoring could not be considered as a screening procedure due to variations in rate of weight gain in individual babies. The Coventry Consensus recommendations on routine weighing included the following:

  1. Babies should be weighed (nude) at immunisation and surveillance contacts.
  2. Normally growing babies should not be weighed more than once per fortnight under the age of six months and no more than monthly thereafter, as this may simply increase anxiety.

A Cochrane systematic review of growth monitoring in children (Panpanich & Garner 1999) found only two RCTs, both conducted in developing countries. They concluded that there was insufficient reliable information to be confident whether routine growth monitoring is of benefit to child health in any setting.

With regard to growth charts the Joint Working party on Child Health Surveillance recommended the 1990 9-centile charts. They also cautioned that:

“Growth charts show the measurements for babies and children at different ages, taken from large numbers of subjects at a range of ages. They do NOT mean that normal babies always grow along the lines shown on the chart.”

In a review of growth charts in the U.K., Sachs et al (Sachs, Dykes, & Carter 2005) described the database for the 1990 U.K. chart. Essentially this chart combines cross sectional data from the British Standards Institute for 163 children ages 0–3 months and 90 children ages 3–6 months and longitudinal data from the Cambridge Infant Growth study which included 252 children recruited between 1984 and 1988. There was some selection bias of babies included. Both samples are more than 98% white and of a relatively high socioeconomic status and in the case of the Cambridge study, from the same geographic area. One hundred and twenty of the babies in the Cambridge study population were breastfed for at least 24 weeks with solids introduced at a mean of 15 weeks.

In 2002 new U.K. growth charts based exclusively upon breastfed babies were produced (Cole, Paul & Whitehead 2002). They were developed using the data for the 120 Cambridge infants who breastfed to a minimum of 24 weeks in the Cambridge Infant Growth Study. It has been argued that these charts should be adopted as the norm for use in the U.K. for all babies. However, these charts have not been tested to see whether they improve the accuracy of referrals, affect breastfeeding duration or parental confidence and whether they are generalisable to babies outside of Cambridge. At the time of writing, new growth charts have just been launched by WHO (http://www.who.int/childgrowth/en/index.html). Evaluation of these growth charts for use in the UK has yet to be undertaken.

7.6. Physical Health and Well being

A review of eight common health problems in infancy was undertaken. Evidence in this area is not robust. When evidence was lacking, good practice recommendations were made. The Department of Health publication Birth to Five also addresses all of these areas in its guidance for new parents.

7.6.1. How should common health problems in the infant be identified and managed? Jaundice

  • During the first week of life approximately 50% of term infants have visible signs of jaundice (British Columbia Reproductive Care Program. 2002). In the narrative background of their guideline on jaundice in the healthy newborn, the British Columbia Reproductive Care Program identified the following potential causes of jaundice:
  • Physiologic jaundice due to increased red blood cell volume and immaturity of the liver function in the newborn.
  • Increased breakdown of red blood cells due to blood group and Rhesus incompatibility
  • Decreased conjugation of bilirubin due to prematurity
  • Increased reabsorption of bilirubin from the gastointestinal tract due to asphyxia, delayed feedings, bowel obstruction, delayed passage of meconium
  • Impairment of bile excretion due to sepsis, hepatitis, biliary atresia, cholestatic syndromes
  • Breast milk jaundice, cause unknown

    Early breastfeeding jaundice develops 2–4 days after birth and may be related to infrequent feeding and/or increased reabsorption of bilirubin from the bowel


    Late breast milk jaundice develops much less commonly, 4–7 days after birth and peaks day 7–15. Cause unknown

Most jaundice is benign, but because of the potential toxicity of bilirubin, all newborn infants must be monitored to identify those who might develop severe hyperbilirubinemia. Jaundice before 24 hours of age is always considered pathological and requires further evaluation (British Columbia Reproductive Care Program. 2002).

Jaundice is usually seen first in the face and progresses in a cephalocaudal fashion (head to toe) to the trunk and extremities. The accuracy of clinical judgment in neonatal jaundice was studied among paediatric residents (junior doctors), paediatric nurse practitioners and paediatric attending physicians (consultants) (Moyer, Ahn, & Sneed 2000). A total of 122 healthy infants underwent serum bilirubin concentration measurements and examinations by 2 observers. Infants were observed under bright indoor fluorescent lighting augmented by natural daylight from large windows in the nursery. Assessments were recorded for pre-specified parts of the body and additional specific sites that were suggested by experienced paediatric faculty such as the conjunctiva, the tip of the nose and the palate. Infant skin tone was also classified as “light” or “dark.” Finally, each observer made a prediction of the total serum bilirubin concentration based on the infant’s clinical appearance. Although agreement was good for infant skin tone, the weighted K statistic for agreement between observers for jaundice at each level was at best only marginally greater than chance alone (ranging from 2–23% greater than would have been expected by chance). A serum bilirubin value of 12 mg/dl has been used to distinguish physiological from nonphysiologic jaundice. In this study, the only consistent finding was that infants with no jaundice below the middle of the chest (nipple line) had bilirubin values less than <12 mg/dl. The researchers concluded that bilirubin testing should be based on risk factors for severe hyperbilirubinemia rather than on clinical observation alone.

No studies for risk of elevated serum bilirubin levels were identified. The British Columbia Reproductive Care Program lists risk factors in the introductory narrative review of their guideline on jaundice. These include:

  • Family history of newborn jaundice (especially sibling), anaemia, liver disease, or inborn errors of metabolism
  • Plethora, polycythemia, bruising, cephalhaematoma
  • Poor feeding, vomiting, delayed passage of meconium
  • Excessive weight loss
  • Sepsis
  • Asphyxia
  • Relative prematurity or small for gestational age
  • Hypothyroidism, hypopituitarism
  • Certain ethnic groups i.e. East Asian, Native American
  • Infant of a diabetic mother
  • Maternal ingestion of sulfonamides or antimalarial drugs.

The effect of breast feeding frequency on serum bilirubin levels in the first 3 days after birth was studied in a randomised controlled trial of 275 full term breastfeeding mothers and infants (Maisels et al. 1994). All babies were fed within 1 hour of birth. Mothers and babies were assigned to the frequent feed group or to the demand group. Frequent feeders were asked to nurse their infants at least every 2 hours during the day and no less than every 3 hours at night. Demand feeders were asked to feed whenever the infant cried or appeared to be hungry. Records were kept of the number of feedings, time of the first stool, frequency of stools and administration of supplemental dextrose water. Serum bilirubin levels were measured in cord blood and again between 48–72 hours after birth. Infants in the frequent group nursed a median of 9 times per day and the demand group 6.5 times per day. There was no correlation between frequency of breast feeding and serum bilirubin level attained between 48 and 80 hours in either group. However, it is possible that those infants who were measured early had not yet reached maximal bilirubin levels. Measurements of all infants at 4 days would have been more clinically important because this is often when interventions for hyperbilirubinaemia are initiated.

Maisels and Gifford investigated the relationship between breastfeeding, weight loss and jaundice (1983). It has been theorized that poor intake among breastfed babies leads to excessive weight loss and subsequent jaundice. One hundred babies were grouped according to birth weight: 2500–3000 g, 3001–3500 g, 3501–4000 g, and >4000 g. There were no significant differences in the mean serum bilirubin concentrations between groups. When all serum bilirubin values were plotted against the cumulative weight loss for each infant, no relationship was found.

The treatment of jaundiced breastfeed infants was evaluated in a Swiss study (Amato 1985) which randomised 50 breast fed full terms babies with hyperbilirubinemia to receive either phototherapy or to interrupt breast feeding, substituting infant formula. There was no significant difference in the maximum bilirubin levels between the two groups (p<0.05). Bilirubin levels reached the maximum by a median age of 5.2 days in the first group and by 5.6 days in the second group. Ten infants had increased rebound levels after stopping phototherapy but only 3 of 25 babies treated with interruption of breast feeding had a further rise. It is noteworthy that there was no control group of babies who continued breastfeeding in this study.

A study of 27 full term formula fed infants with clinically detectable jaundice was undertaken at the University Hospital of South Manchester (Alexander & Roberts 1988) to evaluate sucking behaviours in these infants. The researchers postulated that as an infant’s sucking ability gradually improves to reach optimum sucking by the fourth to seventh day of life, and as serum bilirubin values reach a peak between the third to fifth day of life then rapidly decline, that hyperbilirubinaemia would be correlated with poor feeding behaviours. A pressure transducer was attached to a measured volume of milk formula and recordings made of the length of time spent sucking, pause times, sucking rates and mean sucking pressure. Clinical observations of feeding were also recorded on four visual analogue scales. Bilirubin values from days 2–6 were measured. As expected, milk consumption, duration and pressure of sucking all showed a significant increase during the period of study. Sucking rate did not vary with age. Among infants who appeared disinterested in feeds by clinical observation, milk intake was not impaired by measurement.

The Skin Cancer Research Group at the James Cook University of North Queensland, Australia undertook an investigation of professional advice advocating therapeutic sun exposure for a number of conditions, including neonatal jaundice (Harrison, Hutton, & Nowak 2002). As the development of both melanoma and moles are related to early childhood sun exposure there is concern about early protection of infants against sunlight. This cross sectional study of hospital based staff was designed to determine the prevalence of inappropriate and/or inexplicit advice about the therapeutic uses of sun exposure given to postpartum women. Self administered postal questionnaires were utilised. Among nurses, 42.1% said they would recommend exposing the infant to sunlight. Midwives were significantly more likely to advise exposure than other nursing professionals (45% vs. 21.4%; p=0.004). Among medics 54.2% of obstetricians, 23.8% of paediatricians advised sunlight for neonatal jaundice.

A guideline for the management of hyperbilirubinemia was published by the American Academy of Pediatrics in 2004 (Subcommittee on Hyperbilirubinemia. 2004). This is an evidence based guideline which only searched Medline and Pre-Medline. A quality assessment of the literature was conducted. Recommendations for primary prevention include:

Clinicians should advise mothers to nurse their infants at least 8–12 times per day for the first several days. (evidence quality C: benefits exceed harms).

The AAP recommends against routine supplementation of nondehydrated breastfed infants with water or dextrose water (evidence quality B and C: harms exceed benefits).

Recommendations for ongoing assessment include:

  • All hospitals should provide written and verbal information for parents at the time of discharge, which should include an explanation of jaundice, the need to monitor infants for jaundice and advice on how monitoring should be done (evidence quality D: benefits versus harms exceptional).
  • All infants should be examined by a qualified health care professional in the first few days after birth to assess infant well-being and the presence or absence of jaundice. The timing of this assessment may be determined by the place of birth and/or the length of stay in hospital and presence of risk factors for hyperbilirubinemia (evidence quality C: benefits exceed harms): Skin

No research studies were identified which addressed general care of the skin of the newborn infant. Nappy rash is addressed as a distinct issue. Advice about bathing and treatment of eczema is found in Birth to Five (Department of Health. 2005).

Twenty one studies were evaluated in a Cochrane Review of umbilical cord care (Zupan, Garner, & Omari 2004). Most of the studies were from high income countries. No difference was demonstrated between cords treated with antiseptics compared with dry cord care that is keeping the cord clean and dry, or placebo. Antibiotics and antiseptics reduced skin bacterial colony counts but the clinical significance of this reduction is unknown. No systemic infections or deaths were observed in any of the studies reviewed. Antiseptics also appeared to prolong the time to cord separation. Thrush

Oral candidiasis (thrush) is usually caused by Candida albicans which may be transferred from mother to child at birth. Newborn babies are immunologically immature and more susceptible to infection. Treatment usually consists of oral application of nystatin. One RCT was identified which compared treatments of oral thrush in 35 newborn infants (Boon et al. 1989). In this study newborns with thrush were randomised to treatment with 1 ml (20 mg) oral ketoconazole suspension three times daily or with 1 ml nystatin topical suspension 4 times daily. Blinding and allocation concealment were not described. Two days after clinical cure, cultures were taken from the mouth for mycological confirmation. After 1 week, symptoms in all 20 infants on ketoconazole were resolved but only 8 (53.3%) of babies on nystatin (p<0.001) had resolution of symptoms.

A German study compared two commercial oral gels of nystatin with miconazole oral gel (Hoppe & Hahn 1996). Ninety five infants (body weight < 10 kg) with clinical oral thrush and positive fungal cultures were randomised to one of three treatments. Blinding and allocation concealment were not described. Drugs were administered four times daily after feeding. After 14 days the clinical cure rate was significantly higher with miconazole than with either of the nystatin preparations (p=0.0032 and 0.00068). The duration of treatment varied according to the amount of drug contained in one tube; it was 8 days with miconazole and 10 or 14 days with the nystatin. However, even after treatment with miconazole many oral cultures were still positive. The clinical significance of this finding is unclear, therefore, the optimal duration of treatment remains to be determined. Napkin (Nappy) rash

Napkin dermatitis is an inflammatory reaction of the skin in the area covered by a nappy and has a range of causative factors. Local irritants including urine, faeces and chemicals such as deodorants, preservatives, creams and oils have all been implicated in nappy rash. Candida albicans appears to be a causative organism in severe rash.

A U.S. based cross-sectional study (Benjamin 1987) of 1089 infants was carried out to correlate the incidence of nappy rash with possible causative factors. Infants were examined by two registered nurses trained as graders. An analysis of faeces and the presence of C. albicans was also done. Results showed a strong association between severe rash and level of Candida in the faeces. No such association was observed with mild or moderate rash. The rash incidence/severity was significantly lower when the mean number of reported changes of nappy per day was above average. (p<0.03). Infants wearing disposable nappies exclusively had a significantly lower mean rash grade (p<0.02) and a significantly lower incidence of moderate and severe rash (p<0.001). The incidence of combined moderate and severe rash increased significantly with increasing number of bowel movements per day (p=0.0004). Breastfed infants tended to have a lower incidence of moderate and severe rash (p=0.015).

Different types of nappies have been studied with regard to their effect on infant skin. Four double blind clinical studies were undertaken on behalf of a commercial organisation to test conventional disposable nappies, with absorbent gelling material disposables and with home laundered cloth nappies (Campbell et al. 1987). The total test population of 1614 infants were uniformly stratified by age and diet and then randomly allocated to one of the study groups. Rash grades were assigned to all infants in each study. Skin pH measurements were made in two studies and transepidermal water loss measurements were made in a separate clinical study. Skin wetness as measured by transepidermal water loss was significantly lower in the absorbent gelling material group at all points in time (p<0.05). Change in skin pH inside and outside of the diapered area was significantly smaller in the nappies with absorbent gelling material. Grading of nappy rash was by the same three trained registered nurses every 2 weeks throughout the studies. The graders were blinded to the products used. Infants wearing the absorbent gelling material nappies had a significantly lower overall subjective rash grade than did infants wearing conventional disposable or home-laundered cloth diapers (p<0.05).

Another RCT was conducted at the University of Rochester in which 149 infants were randomised to conventional disposable nappies or to those with absorbent gelling material (Lane, Rehder, & Helm 1990). The infants were evaluated from 1 day to 14 weeks of age for nappy rash. The study was double blinded. Statistically significant differences between the infants wearing the two nappy types were identified at 14 weeks for the waistband (p=0.03), genitals (p=0.05) and leg (p=0.02), with the absorbent gel nappies demonstrating lower rash scores. At no other time during the study were statistically significant differences seen.

A clinical study undertaken in Italy(Longhi et al. 1992) followed infants for up to 44 weeks. Infants aged between 3 and 24 months were examined by dermatologists at fortnightly intervals for the first 14 weeks of the study. One hundred of the infants continued in the study for a further 30 weeks during which time they were examined every 4 weeks. The report does not describe allocation concealment or blinding. A total of 2169 clinical dermatological exams were carried out. Some effort was made to standardise clinical criteria for rash but inter-rater reliability statistics were not provided. All infants used only “test nappies” both containing hydrogel superabsorbent (HS) materials but with a different distribution of the HS in the structure of the nappy. The authors found that the incidence of napkin dermatitis in this study was 15.2% and that both endogenous factors such as atopy and age and exogenous factors such as illness and treatment drugs, influenced the incidence of diaper dermatitis.

Two British studies evaluated treatment of napkin dermatitis. One double blind trial was carried out in 62 infants with moderate to severe rash to assess treatment with the antibacterial/antifungal agent miconazole(2%) and hydrocortisone (1%) with nystatin/benzalkonium chloride/dimethicone/hydrocortisone preparation which was specifically designed to treat napkin dermatitis (Bowring, Mackay, & Taylor 1984). Treatments were randomly assigned and cultures taken from each infant. The miconazole group had 80% resolution of rash in 7 days and the nystatin group 84% resolution. The nystatin combination preparation failed to cure 2 (25%) of the 8 infants with mycoses whereas all 5 infants from whom fungi were isolated in the miconazole group responded. The nystatin also caused staining of the nappy in 63% of cases.

A non-comparative multi-centre study was carried out among 112 infants with napkin dermatitis to assess the effectiveness of 1% clotrimazole plus 1% hydrocortisone cream. Symptoms were significantly improved for patients overall (p<0.001) and all but a few patients became symptom free within 14 days (Jaffe & Grimshaw 1985).

A study from 1967 (Whitehouse, Bannan, & Ryan 1967) looked at the effect of bleaching on bacteria levels and rash. On a randomly selected basis, 129 mothers were given either liquid bleach or placebo bleach and new cloth nappies. The bleach products were crossed over at the end of the fourth week and new nappies supplied. Bacterial cultures were taken from the nappies at 2, 4, 6, and 8 weeks. The bleached nappies had an average of 2 organisms/sq inch, whereas the placebo-bleached nappies had an average of 277 /sq inch. Infants had less irritation while wearing the bleached nappies (p<0.005).

An evidence-based clinical guideline produced by the Association of Women’s Health, Obstetric and Neonatal Nurses makes the following recommendations:

To maintain optimal skin environment

  • Change nappies frequently (based on animal model studies which demonstrated the effect of urine and feces on skin)
  • Use nappies made with absorbent gel materials

To prevent nappy rash

  • Use petrolatum based lubricants or barrier creams containing zinc oxide (recommendation based on expert opinion)

To treat rash resulting from contact irritants

  • Protect skin with a barrier cream containing zinc oxide (recommendation based on expert opinion)

Identify Candida albicans rash by presence of red satellite lesions.

  • Treat with antifungal creams. Constipation

No research studies were identified which addressed constipation in the newborn. One American study (Loening-Baucke 2005) reviewed records of 4,157 children <2 years of age. The prevalence rate for constipation in the first year of life was 2.9%. Advice about constipation is provided in Birth to Five (Department of Health. 2005). Diarrhoea

The Paediatric Accident and Emergency Research Group based at the University of Nottingham developed an evidence based guideline for the management of children presenting to hospital with diarrhoea (Armon et al. 2002). They have defined diarrhoea as, “…a change in bowel habit for the individual child resulting in substantially more frequent and/or looser stools (Grade D, expert opinion).” Although the guideline targets hospital management, several recommendations are also relevant to the newborn in the primary care setting:

The following factors in the history of a child presenting with diarrhoea should alert the clinician to a high risk of dehydration:

  • Infants < 6 months
  • More than 8 significant diarrhoeal stools in the last 24 hours
  • More than 4 significant vomits associated with diarrhoea in the last 24 hours (No evidence. Based on Delphi consensus).
  • Signs of dehydration are unlikely when the weight loss is below 3%
  • Dry mucous membranes, sunken eyes, diminished skin turgor, altered neurological status and deep breathing are present at mean weight losses of 3–8%. (Based on Level 2+ evidence).
  • Oral rehydration should be the standard treatment for children with mild-moderate dehydration secondary to gastroenteritis. (Based on Level 1+ evidence).
  • Breast fed infants should continue to breast feed through the rehydration and maintenance phases of their acute gastro-enteritis illness. (Based on Level 1+ evidence).
  • Infants who are not weaned should recommence full strength lactose containing formula following rehydration with oral rehydration solution (ORS). (Based on Level 1+ evidence).

The Nutrition Committee of the Canadian Paediatric Society updated their recommendations on the treatment of diarrheal disease in 2003. A quality assessment scheme was applied to the evidence, although the details of the evidence tables were not published. The Committee recommended:

  • ORS should be used routinely in the treatment of watery diarrhoea and dehydration. (Based on Level I, A evidence).
  • Feeding should be continued throughout rehydration to help maintain gut nutrition (Based on Level II-2, A)
  • Loperamide and Lomotil antimotility drugs should not be used in children because of safety considerations (Based on Level III, E evidence).

Two American research groups performed meta analyses to determine the efficacy of probiotic treatments such as Lactobacillus on the duration of diarrhoea (Huang et al. 2002; Van Niel et al. 2002). Huang et al (2002) evaluated 18 studies, five of which included newborn infants. Their meta-analysis suggested that co-administration of probiotics with standard rehydration therapy reduced the duration of acute diarrhoea by 0.8 days (p<0.001). In the VanNiel et al (2002) analysis, five of eleven studies which met their inclusion criteria evaluated newborns. The summary point estimate from the meta analysis indicated a significant reduction in diarrhoea duration of 0.7 days (CI 0.3–1.2 days) in subjects who were given Lactobacillus compared with control when mean differences in days of diarrhoea. Colic

Infantile colic is defined as excessive crying in otherwise healthy infants and is a common problem during the neonatal period. A comprehensive systematic review of the literature was conducted to evaluate the effectiveness of treatments for infantile colic including diets, drug treatment and behavioural interventions with crying or presence of colic as the primary outcome measures (Lucassen et al. 1998). Twenty seven controlled trials were included in this review. Five trials studied the effect of eliminating cows’ milk on colic. Three trials used soya milk substitutes and two used a hypoallergenic formula. When data were pooled, the effect size of eliminating cows’ milk was 0.22 (CI 0.10–0.34). Hypoallergenic formula had a clear effect, 0.22 (CI 0.09–0.35). However, the effect of soy was not significant. Lowering the lactose content of formula had no effect. Comparison of breast milk with standard cows’ milk in infants who were weaned showed no significant difference. In one trial herbal tea containing chamomile, vervain, liquorice, fennel and balm mint seemed to be effective. Drug treatment with dicycloverine (dicyclomine) showed clear benefit, 0.46 (CI 0.33–0.60) but as 9 out of 177 (5%) of treated infants had side effects, some potentially serious such as breathing difficulties and coma use of this medication cannot be recommended. Simethicone showed no benefit. Trials of behavioural interventions had durations of 2 weeks to 3 months. Increased carrying and other specific management techniques did not reduce crying.

Another systematic review (Garrison & Christakis 2000) searched only Medline and Cochrane for clinical trials or RCTs on management of colic. Twenty two studies were identified. Quality assessment was summarized but evidence grading was not provided. Three studies of simethicone treatment were reviewed but the evidence presented did not demonstrate any conclusive benefit. In three RCTs, dicycloverine (dicyclomine) performed significantly better than placebo. However, due to potential side effects including apnoea, seizures and coma, the manufacturers contraindicated use in infants less than 6 months old. One RCT of methylscopolamine to treat infant colic found no significant impact on symptoms. Data regarding utilisation of hypoallergenic diets by breastfeeding mothers were inconclusive but suggest there may be some therapeutic benefit. A stratified analysis of the Hill study (1995) showed that infants randomised to the hypoallergenic formula had significantly greater improvements in colic than placebo (p<0.01). A second study reported significant improvements in colic while infants were receiving hypoallergenic formulae but data were not presented in a manner that allowed for comparable interpretation by the reviewers. Two RCT’s of soy based formulas were reviewed. One study measured the mean weekly duration of colic symptoms during treatment with soy formula as 8.7 hours compared to 18.8 hours during control periods (RR 0.33, CI .017–.65). The other study did not report the data in a way that allowed for analysis of treatment effect and therefore was not included. Neither of the 2 RCTs that studied the effects of lactase enzymes on colic found any significant difference between treatment and placebo. Behavioural studies which measured colic symptoms and carrying the infant (2) and car riding (1) showed no significant effect. Decreasing infant stimulation was studied in one trial and did appear to be significant (RR 1.87, CI 1.04–3.34). However there were methodological problems with this study as inclusion criterion were not well defined and concealment was questionable. Herbal tea containing chamomile, vervain, liquorice, fennel and balm mint was studied in one RCT. After 7 days of treatment 57% of infants no longer met the Wessel criteria for colic as opposed to 26% of infants in the placebo group. Two final randomised trials looked at sucrose as a treatment for colic. The response to sucrose in these samples appears to have lasted for <30 minutes in the first trial and only 3 minutes in the second trial. The reviewers concluded that an evidence based approach to colic might include dietary changes, particularly to hypoallergenic formula, treatment with herbal tea and reduction in environmental stimulation. Fever

There are many variables which affect both the definition and measurement of normal and elevated body temperature. In 2004 the Brighton Collaboration Fever Working Group developed a case definition and guidelines for fever in an attempt to produce practical guidelines to standardise the collection, analysis and presentation of data on temperature measurement for clinical trials of vaccine safety (Marcy et al. 2004). They defined fever as the endogenous elevation of at least one measured body temperature of ≥ 38° C.

Fever in young infants under 3 months of age has generated great interest and concern in the medical community. Strategies which encouraged extensive diagnostic testing, hospitalization and antibiotic treatment of all febrile infants less than 60 days old were developed to avoid the consequences of failing to detect serious bacterial illness. A large prospective cohort study carried out in the U.S. (Pantell et al. 2004) surveyed the spectrum of disease in 3066 infants aged 3 months or younger with temperatures of at least 380 C who were seen in the outpatient office setting. The top ten primary diagnoses were:

  • Upper respiratory tract infection
  • Unidentified source
  • Otitis media
  • Bronchiolitis
  • Gastroenteritis
  • Urinary tract infection
  • Pneumonia
  • Viral syndrome, nonspecific
  • Viral meningitis (nonherpes)
  • Well child

The majority (64%) of these infants were treated exclusively outside of the hospital. Although this large sample allowed the researchers to assess the frequency of high risk bacteremia/bacterial meningitis in infants the sample included few African American, Hispanic or inner city infants.

A practice guideline for management of infants and children with fever of unknown origin was developed by a group of American physicians (Baraff et al. 1993). Their literature search was limited to Medline and the quality assessment measures were not described. However, meta analyses were performed where appropriate and modified Delphi techniques were employed for consensus. The definition of the lowest temperature which constitutes fever was based on expert medical opinion. A rectal temperature of 38 0 C was considered febrile. Because clinical evaluation of febrile infants is inadequate to reliably exclude serious bacterial infection, the guideline panel members recommended that all toxic appearing infants and children and all febrile infants less than 28 days of age should be hospitalised for parenteral antibiotic therapy. Febrile infants 28 to 90 days of age defined as low risk by specific clinical and laboratory criteria (previously healthy, no focal bacterial infection on physical exam and negative lab screening), may be managed as outpatients, if close follow up is assured.

There are two Cochrane systematic reviews on treatment of fever in children. Neither of these reviews includes trials with infants. The BNF recommends that paracetamol be given to children under 3 months only after consultation with a physician. Ibuprofen is only recommended for children over 7 kg.

Rectal thermometry has traditionally been considered the gold standard for temperature measurement. However in small infants axillary temperature is easy to measure and avoids the possibility of rectal perforation and contamination via stool. Recently tympanic thermometry and single use disposable thermometers have become available. The NHS no longer recommends the use of mercury thermometers due to its potential as an environmental hazard. A number of studies compare the accuracy of various temperature measurement techniques.

Two comprehensive systematic reviews were conducted at the Institute of Child Health in Liverpool (Craig, V et al. 2002; Craig et al. 2000) comparing axillary temperature measurement with the rectal route and subsequently analysing ear thermometry compared with rectal thermometry. In the first review, data on 3201 participants were included in a meta-analysis. In the studies, overall mean axillary temperature was always lower than mean rectal temperature. A sub-analysis of 652 neonates was carried out but only two studies used electronic thermometers. One study showed narrow limits of agreement. The other, with wide limits of agreement, used a pre-1980 device. In the second review, the relation between age and differences in rectal and ear temperature were investigated for 2611 children. No association was noted. Overall, there was significant heterogeneity between mean differences (p<0.0001) and limits of agreement (p<0.0001). The researchers concluded that the findings implied measurements taken with infrared ear thermometry cannot be used as an approximation of rectal temperature.

Two additional studies which evaluated methods of temperature assessment in newborns were reviewed. Sganga et al (2000) compared newborn temperatures obtained by glass, digital disposable, electronic and tympanic thermometers in 184 healthy newborns aged between 1 and 168 hours of age. Left axillary and left tympanic temperatures were measured. For each infant, glass mercury thermometers were used first, followed by the other methods in random order. The glass mercury axillary temperature measurement served as the baseline. The digital thermometer correlated most closely (r=0.84); the electronic thermometer correlation was r=0.74 and the tympanic thermometer correlation was r=0.35.

A study from Rochester, New York (Loveys et al. 1999) evaluated paired temperature measurements taken by digital rectal and infrared tympanic thermometers. 1175 pairs of measurements were obtained, with 200 of these >380 C. Calculating from a linear regression model, rectal and ear temperatures were identical at about 37.70 C. However, lower rectal temperatures were associated with slightly higher ear temperatures (maximum difference 0.40C) and high rectal temperatures were associated with slightly lower ear temperatures (maximum difference of 0.50C).

7.6.2. When should women be asked about health problems in their babies?

There are no studies which address the timing of health assessments in newborn infants. Recommendations have been made by the National Screening Committee (see

7.7. Vitamin K

Narrative Summary

Available evidence indicates that Vitamin K prophylaxis is effective and prevents significant morbidity and mortality due to Vitamin K deficiency bleeding (VKDB) (Ross & Davies 2000). VKDB may present in one of three ways:

  • Early onset within the first 24 hours of birth
  • Classical within the first week after birth and typically presenting with oral, umbilical, rectal or circumcision bleeding.
  • Late onset after the first week, almost exclusively in breast fed infants and often in those with liver disease or malabsorption. Intracranial bleeding occurs in more than 50% of babies who are diagnosed with late onset VKDB. Is there an association between childhood cancer and IM vitamin K?

In 1992, Golding et al reported an association between childhood cancer and intramuscular (IM) administration of Vitamin K (Golding et al. 1992). This finding was based upon a case control study of 195 cases of childhood cancer and 558 controls. There was a significant association (p=0.002) with IM Vitamin K (OR 1.97, CI 1.3–3.0) when compared with oral vitamin K or no vitamin K.

Following Golding’s report, additional studies were conducted by other investigators which did not support a clinical relationship between newborn parenteral administration of Vitamin K and childhood cancer. An expert group convened in the UK in October 1997 by the Medicines Control Agency on its own behalf and for the Department of Health and Committee of Safety of Medicines reviewed the studies linking Vitamin K and childhood cancer. The Working Group considered eight case control and four ecological studies and concluded that overall, the available data did not support an increased risk of cancer, including leukaemia, caused by Vitamin K (Department of Health 1998).

A more recent UK Department of Health (Roman et al. 2002) report published in the British Journal of Cancer conducted a pooled analysis of 2431 children developing cancer before age 15 and compared them with 6338 controls matched for sex and year (but not place) of birth. The analysis confirmed that solid tumours were no commoner in children given IM vitamin K at birth. The situation with regard to childhood leukaemia was less clear and since almost every baby now gets some form of prophylaxis it is unlikely to be clarified by further data collection. The increased risk was small with an unadjusted OR of 1.25 (CI 1.06–1.46) and could be due to that fact that those selected for prophylaxis (because of prematurity, operative delivery, etc.) were already more at risk for some unknown reason.

A recently conducted large national United Kingdom Childhood Cancer Study (UKCCS) included an updated pooled analysis with data for 7017 children (1174 with leukaemia). It found no association between IM Vitamin K and any diagnostic group, with the pooled odds ratio for leukaemia diagnosed between 12 and 71 months of age of 0.98 (CI 0.79–1.22) (Fear et al. 2003). The authors concluded that in the light of all available evidence, chance was the most likely explanation for early findings regarding the link between Vitamin K and childhood cancer.

Only a controlled trial could resolve any uncertainty about the association between Vitamin K and childhood leukaemia and this would have ethical as well as methodological implications. When and in what doses should Vitamin K be administered?

A Cochrane review (Puckett & Offringa 2000) evaluated eleven RCTs comparing biochemical indices of coagulation status between oral and IM routes of administration. A single oral compared with a single IM dose resulted in lower plasma Vitamin K levels at two weeks and one month whereas a 3 dose oral schedule resulted in higher plasma vitamin K levels at two weeks and two months than a single IM dose.

A study undertaken in the north of England, between 1993 and 1998, (Wariyar et al. 2000) treated 182,000 healthy babies with 1 mg of an oral Vitamin K preparation (Orakay) at birth. Babies judged to be at high risk, (13,472) received 0.1 mg/kg IM Vitamin K after birth. Regardless of their treatment at birth, it was further recommended that the parents of all breastfed babies should then be given three 1 mg. capsules of Orakay and told to give their baby the contents of one capsule at fortnightly intervals after discharge. None of the babies treated with Orakay developed any sign suggestive of Vitamin K deficiency bleeding in the first seven days of life. Four documented cases of late Vitamin K deficiency bleeding occurred to babies in the study. Two babies had underlying alpha-1-antitrypsin deficiency and the mothers of two others did not receive instructions or capsules for prophylaxis. A postal survey of 458 mothers (61% response rate) and interviews with 173 randomly selected mothers both indicated that 93% of the babies still being breastfed had all four doses as recommended and that 98% had had at least three doses. What is the most clinical and cost effective route (IM or oral) of delivering Vitamin K?

Narrative Summary

The intramuscular route was the most common route of administration of Vitamin K until the Golding report. Subsequently, the British Paediatric Association (now known as the Royal College of Paediatrics and Child Health) recommended that infants should receive Vitamin K supplementation orally with several doses to be given to infants who are breast fed.

As there was uncertainty about the optimum dosage, regimens in Britain became quite varied. In addition some inherent problems with oral dosing, which potentially compromised effectiveness, were identified. These include:

Compliance: Several doses of oral Vitamin K were needed over several weeks. A hospital in London surveyed mothers in June 1993 to evaluate its new policy of prescribing three 0.5 mg doses of Vitamin K for breast fed babies. The survey (n=207) showed a compliance rate of 99% for the first dose, 88% for the second at 1 week of age and 39% for the third at 6 weeks (Croucher & Azzopardi 1994).

Absorption: There was potential unreliability of absorption of oral Vitamin K, e.g. variable absorption or regurgitation

In December 1992 the Australian College of Paediatrics and the Royal Australian College of Obstetricians and Gynaecologists recommended replacing IM Vitamin K with three oral doses of 1 mg. each. However, the decision was rescinded in 1994 due to increased incidence of VKDB, thereby emphasising the efficacy of a regimen of IM prophylaxis.

In several European countries, oral regimens of Vitamin K administration were initiated in light of the Golding report. Surveillance studies on late VKDB (days 8–84) in countries which adopted oral therapy are reported below:

  1. 325,000 newborns given 1 mg Vitamin K IM at birth; 0 cases of VKDB
  2. 1,200,000 newborns in Germany given 1 mg Vitamin K orally at birth and two additional 1 mg doses between days 4–10 and days 28–42; 32 cases of late VKDB (2.7 per 100,000)
  3. 800,000 newborns in Germany given 2 mg Vitamin. K orally at birth and two additional 2 mg doses between days 4–10 and days 28–42; 7 cases of late VKDB (0.9 per 100,000).
  4. 325,000 newborns in Australia given three doses of 1 mg oral Vitamin K at birth and again between days 3–5 and days 21–28; 8 cases of late VKDB (2.5 per 100,000)
  5. 83,000 newborns in Switzerland given 2 mg of mixed micellar preparation of Vitamin K (Konakion MM- better absorbed and produces higher plasma K levels than the standard oral preparation) by mouth on days 1 and 4; four cases of late VKDB (4.8 per 100,000)
  6. 439,000 newborns in the Netherlands given 1 mg oral Vitamin K at birth and 25 micrograms daily from week 1 to week 13; 5 cases of late HDN (two did not receive the prophylaxis as recommended and three had “predisposing illness” and did not receive additional Vitamin K which would have been appropriate). (1 per 100,000).
  7. 396,000 newborns in Denmark given 1 mg oral Vitamin K at birth and 1 mg weekly; 0 cases of VKDB (Hansen, Minousis, & Ebbesen 2003)
  8. The efficacy of 3 oral 2 mg doses with Konakion MM remains to be established (von Kries 1999)
Cost Effectiveness
Current Advice from the DH

The DH continues to advise that all new born babies receive an appropriate Vitamin K regimen to prevent the rare, but serious and sometimes fatal, disorder of VKDB.

The recommendations state that Vitamin K can be given in a single dose of 1 mg IM. Oral regimens for vitamin K, on the other hand, must be repeated. Arrangements must be in place to ensure that all recommended oral doses are given at the appropriate ages. Doses of Vitamin K given at or shortly after birth should suffice for formula fed babies. Breastfed babies need additional Vitamin K. Konakion MM is licensed for oral use in 2 doses of 2 mg to be given in the first week for all babies and, for exclusively breastfed babies, a third dose of 2 mg is to be given at one month of age. Konakion MM is formulated in glass ampoules and currently it is necessary for a health care professional to administer the doses (Department of Health 1998). Orakay is not yet licensed in the UK.

Cost Perspective

There are two major costs associated with the administration of Vitamin K in neonates. The first is the cost of the drug and the second is the cost of the HCP required (for certain situations) to administer it.

In this case, issues surrounding effectiveness are of less importance than those of cost. The first reason is that Vitamin K Deficiency Bleeding (VKDB) is highly unusual among babies who had successful administration of a Vitamin K regimen (so either regime will have a similar outcome). The second is that intramuscular treatments are generally considered more likely to eliminate the threat of VKDB than an oral approach. Since this costing summary will show a lower cost level for the IM approach, the inclusion of effectiveness will only accentuate the advantages of the IM strategy.

The comparison undertaken here is the cost of an oral region of Konakion MM against an intra-muscular (IM) regimen of the same drug. Orakay is a possible future alternative but is, as yet, unlicensed. General guidance on the effect of introducing Orakay on cost effectiveness issues are outlined later for information.

Another major issue is the suggested link between IM Vitamin K prophylaxis and childhood leukaemia. Other than the Bristol study (Golding, Birmingham, Greenwood, & Mott 1992), there is a wide consensus that no solid link has been established. Therefore, this view was followed.

The oral regimen that seems to be prevalent is a series of three doses of 2mg, one at birth, one within the first week and one after around one month for breastfed babies (since breastfed infants do not receive the Vitamin K supplementation found in formula). This contrasts with the IM route which requires one dose of 1mg soon after birth. The cost of a dose of oral Konakion MM Paediatric is £1.00 (as of April 2006) while the intramuscular option (Konakion Neonatal) costs 21p. On the 31st of March 2006, Konakion Neonatal was discontinued. Thus, the results of the analysis for the alternative intramuscular option, namely Konakion MM Paediatric, are reported.

The drug cost of the oral approach is therefore £2.00 or £3.00 depending on whether the baby is breastfed. Thus, the total drug cost is (£2.00 x (1 - breastfeeding rate) x 640 000) + (£3.00 x breastfeeding rate x 640 000). Assuming a breastfeeding rate of 50%2, this gives a total drug cost of £1.6 million.

As of March 2006, the IM route costs £1.00 per baby. This is based on the assumption that the second half of each 2 mg vial is wasted. Thus, the total implementation of this strategy would cost £640 000 (the drug cost is the total cost). It was felt that the non-usage of the second half of the vial was best representative of current practice.

The assumption made on the cost side was that the administering of the IM injection and the first and second oral was effectively costless as it can be tied-in with other routine visits (in hospital or otherwise) from the HCP.

For breastfed babies, this assumption of costlessness of provision was much less certain as the third dose is at a later stage. It seems that the strategy across England and Wales for administering this third dose is rather mixed. It is sometimes the case that a HCP would go to the baby to administer it (a health visitor in this case). Cost figures for this were taken from Curtis and Netten, Unit Costs of Health and Social Care 2004 (Curtis & Netten 2004).

The total annual cost of a health visitor (wages plus NI contributions and overheads) was calculated and divided by the suggested workload (37.5 hours multiplied by 42 weeks) to provide an hourly cost. Assuming the trip takes half an hour and adding the cost of travelling (taken from the same source), this gives a total of £11.86 per visit. The non-drug costs of the oral Konakion MM regime were calculated thus,

£11.86 x breastfeeding rate x 640 000

If we use the previous assumption of a 50% breastfeeding rate, this means a non-drug cost of £3.795 million. If the guardian administers the third dose, this cost would be eliminated but further costs would be incurred through dealing with VKDB cases due to uneven compliance.

The summing of these produces a cost of an oral Konakion MM regimen of £1.6 million (£5.395 million if the third dose is HCP administered) and an IM regimen cost of £640 000.

The option most commonly cited as an alternative to these two approaches is Orakay. At present, this is unlicensed. Therefore, it can be recommended only if the clinical and cost-effectiveness evidence is compelling in favour. If, and when, Orakay is licensed, the effect of this would be to lower the cost of the oral approach (since the cost is around 31.7p per dose)

If Orakay was to be recommended and administered by parents, it would be necessary to repeat the analysis allowing for a non-compliance rate. It is important to note that, a recommendation for Orakay based on cost-effectiveness criteria would necessarily require either significantly better clinical outcomes than the intramuscular approach or lower costs. At present, the evidence suggests neither is proven.

Economic Summary

From a costing side, an IM regimen of Vitamin K prophylaxis is less expensive than a licensed oral regimen. The respective annual costs are £640 000 and £1.6 Million. If we follow the view on the lack of a link between IM prophylaxis and childhood leukaemia, adding a measure of effect to this analysis will make IM prophylaxis relatively more cost-effective compared to its oral alternative.

If Orakay is to be included in the calculation, the evidence on cost-effectiveness relative to the IM route is uncertain but it is likely to be preferred to an oral Konakion MM system due to a lower level of costs.

7.8. Safety

7.8.1. Accident prevention

Narrative Summary Incidence

Accidental injuries are a major cause of childhood morbidity and mortality throughout the United Kingdom. Three children die as the result of a home accident every week (Royal Society for the Prevention of Accidents 2004). Those most at risk from a home accident are in the 0–4 year old group. The figures in Table 7-1 were provided by the Royal Society for the Prevention of Accidents, and identify the most common fatal and non fatal accidents incurred by children from 0–1 year of age.

Table 7-1. Common fatal and non fatal accidents incurred by children from 0–1 years.

Table 7-1

Common fatal and non fatal accidents incurred by children from 0–1 years.

ALSPAC (Avon Longitudinal Study of Parents and Children, formerly the Avon Longitudinal Study of Pregnancy and Childhood) collected data on injuries resulting from accidents in pre-mobile infants (Warrington, Wright, & Team 2001). A total of 11466 responses were collected by postal questionnaire. When their infants reached the age of 6 months, mothers were asked to describe any accident that had occurred to their babies since birth. Falls and burns were the two events reported by the ALSPAC study group. A total of 3357 falls were reported for 2554 children (22% of the total). The most common types of falls were from beds or settees. Only 166 infants sustained a burn or scald (1.5% of the total).

A Canadian epidemiological analysis of injuries experienced by infant children less than 12 months old presented data by circumstances and age (Pickett et al. 2003). In babies 0–2 months, falls accounted for 52% of injuries. Sixteen percent of injuries were unspecified. Falling objects (7.4%) and motor vehicle accidents (7.4%) were the next most common causes of accidental injury to infants. Risk

According to the Child Accident Prevention Trust the number of children’s accidents has been declining steadily, however, children of all ages from the poorest families are still at much greater risk than those from the most affluent households.

Reading et al (Reading et al. 1999) investigated the relationship between social disadvantage and accidental injury in preschool children, attempting to separate the risks associated with individual family circumstance from those associated with wider environmental and area characteristics. The research group identified 21 large socially homogeneous areas in Norwich and collected accident records for 3580 accident attendances of which 1134 were for moderate or severe injuries. The researchers used a multilevel modelling technique particular to geographic epidemiology where information is available on individuals nested within a hierarchical structure of areas, or units such as hospitals and clinics, all of which may independently influence the outcome of interest. In this study, factors which explained the variation in accident rates at the level of individuals included child’s sex, the number of siblings, the age of the mother and whether the mother was a lone parent. For all accidents and for moderate and severe accidents, the proportion of variance explained in the model by individual characteristics was about 90%, the remaining 10% being explained by the social area. The significance of the effect of living in a deprived neighbourhood was greater in the model for more severe accidents with an odds ratio for risk of an accident which was equivalent to 1.49 times greater (CI 1.01–1.08) in the most deprived social area compared to the least deprived. The authors suggest that these finding may have implications for preventative measures which might include social policy changes to improve child care among unsupported families and targeting accident prevention measures at a local level towards deprived neighbourhoods. Several studies have assessed risk factors for injury in young children. Ramsay et al (Ramsay et al. 2003) used a health visitor questionnaire to investigate the physical, social and psychological environment of 79 families with preschool children under 5 years of age presenting to an A & E department with an injury sustained in the home. These families were matched with 128 control families by postcode, age and sex of children. Matching on postcode effectively matched for deprivation and therefore this factor could not be analysed. The main carers of cases had a lower level of educational attainment than controls (p<0.01). Socket cover utilization was also lower in case families.

A prospective American study (Harris & Kotch 1994) interviewed 367 mothers six to eight weeks after delivery and again one year later, when 132 infants (36%) were reported as injured. The infants were recruited through a high priority infant tracking program (HPIP). One low risk infant was recruited for every four “HPIP” infants. In the crude analysis maternal unemployment, number of siblings living in the home, families receiving welfare, maternal stress, maternal depression, family conflict and seeking spiritual support were associated with unintentional injury. After adjusting for potential confounding only three variables remained as significant risk factors for injured infants: family conflict, maternal unemployment and families that had fewer than two other children, OR 1.3 (CI 1.1–1.4). Among mothers with high levels of social support the odds of injury were 0.38 (0.04–3.44) in contrast to those with low levels of social support, where the odds of injury were significantly increased, 2.41 (1.37–4.24).

In another American study (Gielen et al. 1995)150 mothers of babies aged 6 and 36 months were interviewed about injury prevention practices and personal beliefs about safety. The interviewees were residents of a disadvantaged inner city community. In this population factors significantly associated with the number of injury prevention practices implemented were family income, housing quality and environmental factors such as moving house frequently or living alone. Prevention

In 2001 the Health Development Agency published an updated systematic review of health promotion intervention to prevent unintentional injuries in childhood and young adolescence (Towner et al. 2001). The review included studies published between 1975–2000. One hundred and fifty five studies were evaluated. A range of interventions shown to reduce injury or to change behaviour were identified. Those which might apply to young infants have been extracted and are presented in Table 7-2:

Table 7-2. Interventions to Reduce Injury.

Table 7-2

Interventions to Reduce Injury.

The HDA report also reviewed ten studies which evaluated the effectiveness of community based childhood injury prevention programmes. Six of these were based on the WHO Safe Communities model which combines local surveillance systems and a reference group to coordinate activities. Eight of the programmes were at least partially effective in either reducing injury or increasing knowledge about safety. A meta-analysis was not performed.

The U.S.-based Institute for Clinical Systems Improvement (2004) developed a guideline for childhood preventive services. A systematic review of the literature was not done. The evidence which was reviewed was graded as ‘good,’ ‘ fair’ or ‘insufficient.’ They concluded that there is fair evidence to support counselling on several safety topics which relate to young infants including child safety seats, flame resistant sleepwear, protection from UV light and infant sleep position and SIDS.

An RCT was conducted in a large urban teaching hospital in the US to develop and evaluate an injury prevention anticipatory guidance training programme for paediatric residents (Gielen et al. 2001). The goal was to improve parents’ safety practices for the prevention of burns, falls and poisoning in young children aged 0–2 years old living in low-income inner-city neighbourhoods. Thirty one paediatric residents were randomised either to standard safety training plus a 5 hour experiential instruction programme on injury prevention and counselling skills or to the standard one hour seminar about injury prevention alone. Parents seen by physicians in the intervention group (n=117) received significantly more injury prevention counselling and were significantly more satisfied but their knowledge, beliefs and home safety behaviours did not differ significantly from parents seen by the control group of doctors (n=73). The researchers concluded that because low-income families face many barriers to carrying out recommended safety practices, supplemental strategies are needed to ensure safer homes.

A study was conducted in the UK at a GP practice in Nottingham (Clamp & Kendrick 1998) to assess the effectiveness of counselling on injury prevention by a general practitioner in conjunction with access to low cost safety equipment for families on a low income. One hundred and sixty five families responded to a questionnaire on child safety practices. They were then randomly assigned to an intervention or a control group. The intervention consisted of standardised advice and safety leaflets. Families receiving means tested benefits were offered a smoke alarm for 50p and two window locks, three cupboard locks, six socket covers, or a door slam device for 20p, all available at the surgery at the time of the consultation. Stair gates and fireguards were offered at £5 per item. After the intervention, significantly more families in the intervention group used fireguards (RR 1.89, CI 1.18–2.94), smoke alarms (RR 1.14, 1.04–1.25), socket covers (RR 1.27, 1.10–1.48), locks on cupboards for storing cleaning materials (RR 1.38, 1.02–1.88) and door slam devices (RR 3.60, 2.17–5.97). Also significantly more families in the intervention group showed very safe practice in storage of sharp objects (RR 1.98, 1.38–2.83), storage of medicines (RR 1.15, 1.03–1.28), window safety (RR 1.30, 1.06–1.58), fireplace safety (RR 1.84, 1.34–2.54), socket safety (RR 1.77, 1.37–2.28), smoke alarm safety (RR 1.11, 1.01–1.22), and door slam safety (RR 7.00, 3.15–15.6). Stratifying results by receipt of state benefits showed that the intervention was at least as effective in families receiving benefits as others.

Kendrick et al (Kendrick 1999) evaluated a package of safety advice provided at child health surveillance consultations; provision of low cost safety equipment including stair gates and fireguards, cupboard locks and smoke alarms; and home safety check and first aid training by health visitors. Overall 44 health visitors from 36 practices in Nottingham took part. Eighteen practices were randomly allocated to the intervention group (n=823) and 18 practices comprised the controls (n=771). No significant difference in injuries was demonstrated. However, as severe injury is a rare event, a larger study may be required to assess the effect of the intervention in these cases. Role of Home Visiting in Prevention of Childhood Injury

In 1983 a study was published in the Lancet which evaluated the first seven years of the Sheffield Intervention Programme, from 1973–1979 (Carpenter et al. 1983). In this study 39,452 infants born to parents resident in Sheffield were scored at birth for risk of unexplained infant death. Infants in Sheffield were normally seen by their health visitor at 10 days, 4 weeks and 3 months. High risk infants were seen every 2 weeks up to 3 months and every month up to 6 months. Thus high risk infants received at least five extra home visits in the first 6 months. A campaign to promote breast feeding and to ensure that bottle feeds were made up correctly decreased the risk of hypernatraemia. Before 1973 post-perinatal mortality in Sheffield was, on average, 11.5% above the rate for England and Wales. Since 1973 it has only once exceeded the rate for England and Wales. “Possibly preventable” deaths have fallen from 5.2 to 1.9 per 1000. Twelve percent of this decline is associated with a rise in the average age of the mother and a fall in the number of pregnancies, 9% with a reduction in precipitate deliveries, 24% with an increase in breastfeeding, 18% with extra care given by health visitors to high-risk infants, and 36% with other factors.

A systematic review of home visiting and the prevention of childhood injury was conducted by Ian Roberts of the Institute of Child Health and published in 1996 (Roberts, Kramer, & Suissa 1996). Eleven RCTs were reviewed. Eight of these examined the effectiveness of home visiting in the prevention of childhood injury. The pooled OR for the eight trials was 0.74 (CI 0.60–0.92). Four studies examined the effect of home visiting on injury occurrence in the first year of life only. The pooled OR was 0.98 (0.62–1.53). Most of the trials in this review used non-professional home visitors. In all but one of the trials the intervention was targeted at groups considered to be at increased risk for adverse child health outcomes.

A Canadian multicentre randomised controlled trial was conducted to evaluate the effect of a single home visit by a research assistant which included the provision of an information package, discount coupons, and specific instruction regarding home safety measures (King et al. 2001). Initially, 1172 homes were inspected for safety. They were then randomised to intervention or control. The median age of children in the homes was 2 years. Outcomes included (1) parental injury awareness and knowledge; (2) the extent that families used home safety measures (3) the rate of injury and (4) the cost effectiveness of the intervention. There were no significant differences in prenatal injury awareness and knowledge from baseline between the intervention and control groups. Significant changes were observed in the number of intervention homes who had hot water not exceeding 54 degrees centigrade (OR 1.31, CI 1.14–1.50) and who had smoke detectors on some or all levels (OR 1.45, CI 0.94–2.22). There was a significant difference, favouring the control group, in the number of homes with a fire extinguisher (OR 0.81, CI 0.67–0.97). The intervention group reported significantly less injury visits to the doctor compared with the non-intervention group (rate ratio: 0.75, CI 0.58–0.96). The total costs of care for injuries were significantly lower in the intervention group compared with the controls with a cost of $372 per injury prevented.

7.8.2. Reducing the risk of SIDS SIDS Risk (a review of modifiable risk for SIDS after initiation of the “Back to Sleep” campaign)

The study of risk of Sudden Infant Death Syndrome has resulted in the identification of a number of modifiable risk factors which have subsequently been targeted in public health campaigns. Advice from Chief Medical Officer in England formed the basis of a national intervention campaign in November 1991 called “Back to Sleep” to reduce the prevalence of modifiable risk factors for SIDS. In 1992 the U.K. Confidential Enquiry into Stillbirths and Deaths in Infancy (CESDI) was established to ‘improve understanding of ways in which the risks of death might be diminished with particular regard to avoidable factors’ (Saunders 1997). Early studies identified sleeping position, maternal smoking and bedding as factors affecting the risk of SIDS (Beal & Finch 1991; Dwyer et al. 1991; Fleming et al. 1990; Mitchell et al. 1992). Other factors such as breast feeding, use of dummies and bed sharing were not consistently found to be associated with risk of death (Ford et al. 1993). In 1993 the Department of Health funded a large scale detailed case control study of sudden unexpected deaths in infancy (SUDI study) in three NHS regions in the United Kingdom (Southwest, Yorkshire and Trent, population 17.7 million) combined with a confidential inquiry into such deaths over a two year period starting one year after the national risk reduction campaign (Blair et al. 1999; Blair et al. 2000; Blair et al. 1996; Fleming et al. 1999; Fleming et al. 1996). Cases (n=195) were reported within 24 hours and data were collected on a standard questionnaire by research interviewers. Each bereaved family was visited twice within two weeks of the death. Four controls were selected for each case (n=780), with babies matched on age and geographical location. Controls were visited within a week of the case baby’s death. In the multivariable analysis the following environmental risk factors were significant:



Prone sleeping Side sleeping

An important new observation in this data was that the side sleeping position carried significantly increased risk when compared with supine. Total head covering emerged as the most potent of all risk factors (21.58; 6.21–74.99).

Subgroup analysis of bed sharing showed that the associated risk was significant only among smokers, OR 9.25 (2.51–34.02). There is no suggestion that taking the baby into bed for a short time for feeding or for comfort poses any risk except that the mother may then fall asleep and keep the baby in her bed all night long. There were no differences in the proportions of babies sharing a room with an adult or another child. There were factors related to bed sharing among young babies that seem to be associated with increased risk. These included maternal alcohol consumption (OR 3.40; 1.88–6.16), parental fatigue (OR 1.61–3.63), overcrowded housing conditions (OR 18.49; 3.62–94.48) and infant being under a duvet (OR 3.97; 2.43–6.46).

The risk associated with smoking during pregnancy was significant (1.78; 1.04 – 3.05). Paternal smoking had an additional independent effect (2.50; 1.48 −4.22). If one or both parents smoked the OR for SIDS was 3.79 (2.09 – 6.88) in a multivariable analysis. The additive effect of smoking in pregnancy and postnatal exposure was also significant (2.93; 1.56 – 5.48).

No protective effect of breastfeeding was identifiable in the multivariable analysis.

The SUDI study was carried on for a third year in which pacifier use was recorded. The proportion of infants who had ever used a pacifier for day or night sleeps was exactly the same (66%) but significantly fewer SIDS infants used a pacifier for the last/reference sleep, OR 0.41 (0.22–0.77), giving an apparent protective effect against SIDS.

Prospectively recorded weight observations for cases (n=325) and controls (n=1300) were collected in the SUDI study as well. Poor weight gain in the first six weeks emerged strongly as a risk factor only for those babies with birth weights above the 16th centile (OR 1.97, 1.49–2.61, p<0.0001 ).

An American study published in 2005 (Alexander & Radisch 2005) presented the results of a retrospective review of 102 SIDS deaths. The researchers aimed to evaluate the overlap of known risk factors such as sleep position, sleep surface and co-sleeping. Of particular interest was the frequency of infants who died with no risk factors. Of the 102 deaths, 67(65.7%) were not in a crib, 63 (61.8%) were prone and 48 (47.1%) were co-sleeping. Overall, 94 (92.2%) of these deaths had at least one risk factor present. Only 8 (7.8%) of infants had none of these risk factors, leading to the conclusion that for infants in this study a SIDS death was rare if the baby was sleeping alone, in a crib, on its side or back.

The relationship between routine infant care practices and Sudden Infant Death Syndrome was investigated in a four year case control study by the Scottish Cot Death Trust (Brooke et al. 1997). Questionnaires were completed by parents recruited for this study, with the assistance of a fieldworker at a home visit. Forms were completed for 147 cases out of a total of 201 reported SIDS deaths and on 276 controls. Sleeping prone (OR 6.96; 1.51 to 31.97) and drug treatment the previous week (OR 2.33; 1.10–4.94) were more common in cases than controls in the multivariable analysis. Smoking was a significant risk factor if the mother smoked (OR 5.05; 1.85 to 13.77), the risk increasing if both the mother and father smoked (OR 5.19; 2.26 to 11.91). However paternal smoking on its own was not significant. The risk of SIDS increased with the number of cigarettes smoked by the mother or the father (p=0.0001) and also with bed sharing (P<0.005). Sleeping on a previously used mattress was also significant (OR 2.51; 1.39 to 4.52), although the risk was not established for mattresses completely covered by polyvinyl chloride.

A study conducted in southeast Norway and published in 2001 compared changes in risk factors for SIDS during the time of increasing SIDS rate (1984–1989), the period of rapid decline (1990–1992) and the time of stable low rates (1993–1998) (Arnestad et al. 2001). A questionnaire was distributed by mail to 266 parents who had lost a child to SIDS during the period 1984–98 and to 698 control parents. The controls were chosen at random from the national population register, and matched on sex, date of birth and place of birth. The questionnaires were distributed in 1993 and 1998. The response rate was 69% in the SIDS parents group and 75% in the control parents group. The mean time between death and completion of the questionnaire was three years. The SIDS rate in southeast Norway decreased from 1.8/1000 in 1989 to 0.6/1000 in 1993 at which point it stabilised. During this time, the proportion of infants found dead in the prone sleeping position decreased from 93% to 63%, however, over half of the SIDS victims were found in the prone position. No increased risk of side sleeping was demonstrated by this study. Although co-sleeping as a usual mode of sleep increased over the time period studied, it could not be related to changes in the SIDS rate for the region.

There was a positive interaction between co-sleeping and maternal smoking during pregnancy (OR 8.63, 1.87–39.85, p<0.01). Maternal smoking before, during and after pregnancy was significantly associated with SIDS in this study in both the crude and multivariable analysis. The OR postnatally was 3.7 (2.5 to 5.5). When paternal smoking was adjusted for maternal smoking in pregnancy there was no effect shown. A clear dose response relationship was demonstrated for maternal smoking with the highest rates of SIDS occurring in the children of the heaviest smokers.

Caffeine during or after pregnancy was not found to be an independent risk factor for SIDS after adjusting for maternal age, education, parity and smoking during pregnancy.

Heavy postnatal alcohol intake was significantly associated with SIDS (OR 5.9; 1.0 to 33.9) but was based on a very small sample of 7 cases and 2 controls.

Both the usual age peak for SIDS deaths between 2 and 4 months and the seasonal peak during autumn/winter became less pronounced during the time period of this study.

The European Union Concerted Action of SIDS (ECAS) was founded in January, 1994 to combine data from across Europe with the aim of reexamining the epidemiology of SIDS (Carpenter et al. 2004). Data were derived from 20 centres between September, 1992 and April, 1996. The Nordic study and three regions in England which made up the first 2 years of the SUDI study were included. New studies following the same ECAS protocol were also set up in 12 centres, six of which were in Eastern Europe. Interview data were collected for 745 cases and 2411 controls. Highly significant risks were associated with prone sleeping (OR 13.1; 8.51 to 20.2); head covered (OR 12.5; 6.47 to 24.1); maternal alcohol consumption in last 24 hours of 3 drinks or more (OR 2.33; 1.28 to 4.21); illegal drug use by the mother (OR 1.92; 1.00–3.70). Maternal smoking was significant and there was a dose response (< 10 cigarettes per day OR 1.52; 1.10 to 2.09; >10 cigarettes per day OR 2.43; 1.76 to 3.36).

In this study the OR associated with bed sharing all night with an adult was progressively greater for younger infants whether or not the mother smoked, the increase being 5.2% per week in both groups (p=0.002). Adjusted OR’s for smokers who bed shared was 11.3 times greater than for non smokers and ranged from OR 27 (13.3–54.9) at 2 weeks to 7.5 (4.3–13.2) at 26 weeks. For non smokers, the risk of bed sharing was significant only at less than 8 weeks (OR 2.4; 1.2–4.6).

Regular dummy use and use in the last sleep was significantly less than one in this data analysis (OR 0.74; 0.58–0.95 and OR 0.44; 0.29–0.68, respectively). Similarly infants who regularly shared a room with their parents and did so on the last occasion were at significantly less risk (OR 0.48; 0.34–0.69 and OR 0.32; 0.19–0.55, respectively). Duvet use on the last occasion was significant with an adjusted OR 1.82; 1.30–2.58).

Overall, sleeping position, type of bedding used, and sleeping arrangements were the modifiable factors identified in this study for future SIDS reduction campaigns.

A five year case control study was conducted using data on SIDS victims in the Republic of Ireland between January 1994 and December 1998 (McGarvey et al. 2003). Despite dramatic reductions in SIDS deaths after the “Back to Sleep” campaign, SIDS remained the leading cause of infant death in Ireland. The purpose was to examine the role of factors relating to the infant’s sleeping environment which might constitute modifiable risk factors for SIDS. A total of 203 SIDS cases and 622 control infants were included in this study. Parents were interviewed in their homes within 6 weeks of the case infant’s death. Only the use of the prone sleeping position (OR 11.47; 1.24–106.06), co-sleeping (OR 16.47; 3.73–72.75) and absence of routine soother (dummy/pacifier) use (OR 5.86; 2.37–14.36) were statistically significant in the multivariable analysis.

The risk of co-sleeping for infants whose mothers did not smoke was significant with a CI ranging between 1.42 and 13.91. However in cases where the mothers smoked, the OR varied between 13.21 and 711.69. In fact, 90% of all SIDS cases who were co-sleeping during the last sleep period had mothers who smoked during pregnancy. The mechanism by which maternal smoking exerts a fetal biological effect is generally thought to be by impaired fetal growth, chronic hypoxia and impaired brain stem function with subsequent poorer cardio-respiratory control postnatally. Thermal stress of the co-sleeping may combine with physiological vulnerability to prove a lethal combination.

In this study more cases than controls habitually used a soother. However, 47% of these babies did not have their soother on the night they died. Infants who usually used a soother and did not have it during the last sleep were almost six times more at risk than regular users who did have it. It appears that the absence of habitual soother may present a risk.

Secondary analyses of the New Zealand Cot Death Study (Mitchell et al. 1992) were conducted by Ford et al (1993, 1996, 1997). These researchers looked at lack of social support and the experience of stressful life events as well as breastfeeding and illness/acute medical care as risk factors for SIDS (Ford et al. 1996). The original data for the New Zealand Cot Death Study was collected by a comprehensive parental questionnaire answered one month after the death of the case infant. The study ran from 1 November 1987 to 31 October 1990. Subject ascertainment was 81% (n=393) for case families and 88.4% (n=1259) for controls families interviewed. A full set of data for the analysis of social support and life events was available for 390 cases and 1592 controls. Mother’s social support index (MSSI) was used and questions relating to family stress were derived from previous studies. In this study there were only minor differences between SIDS families and control families for social support. Families who had experienced more than six stressful life events in the preceding year had three times the risk for a SIDS event. This result was nonsignificant when socio economic and demographic confounders were included in the model. The researchers concluded however that it was possible that sociodemographic influences on SIDS may work through additional life event stresses. These two factors are difficult to disentangle.

The analysis of duration of breastfeeding and the risk of SIDS (Ford et al. 1993) utilised data from 356 cases and 1529 controls. In a multivariate analysis, exclusive breastfeeding at discharge from the hospital appeared to be protective (OR 0.52, 0.35–0.77) and the protective effect was also present for infants exclusively breastfed in the ‘last two days’ (OR 0.65, 0.46–0.91).

Finally, the analysis of illness and acute medical care (Ford et al. 1997) was evaluated using data from 390 cases and 1592 controls. Severe illness (based on BabyCheck) was associated with an increased risk of SIDS (OR 2.36, 1.14–4.90). SIDS infants were taken to the GP more often than control infants for the same severity of illness but the difference was not significant on multivariate analysis. Only 1.3% of all SIDS cases had symptoms suggestive of severe illness and had not seen a GP. The researchers concluded that lack of appropriate medical contact for illness could not be implicated as a risk factor for SIDS.

A before and after cohort study conducted in Tasmania assessed the effect of supine positioning on SIDS rates in the state (Dwyer et al. 1995). Infants were assessed for risk of SIDS using a locally constructed predictive model. Of 5534 high risk infants included in the study, 39 died of SIDS. The cohort was divided into pre-intervention (May 1, 1988 through April 30, 1991) and post-intervention (May 1, 1991 through October 31, 1992) cohorts. Three interviews were conducted, at 4 days of age, at home during the fifth postnatal week and by telephone during the 10th postnatal week. The SIDS rate in Tasmania from 1975 through 1990 was 3.8/1000. In 1992 the rate of 1.4/1000 was significantly lower (p<.01). During the pre-intervention period the SIDS rate in the high risk cohort study was 7.6/1000. The rate decreased to 4.1/1000 and when adjusted for season of birth the decline was significant (p=.04). Multiple logistic regression analysis was used to estimate the individual contribution of the various risk factors. The largest individual contribution to the decline in SIDS was made by the change in usual prone position, accounting for 70% of the decline in the cohort. All other factors accounted for less than 10% of the reduction in SIDS rate. The decrease in SIDS rate after the intervention, adjusted for season of birth and usual prone position, was not materially altered by further adjustment for any of the individual factors.

Two additional studies were reviewed which assessed the role of other potential risk factors for SIDS. The role of breastfeeding (McVea, Turner, & Peppler 2000) was the subject of a meta analysis performed at the University of

Nebraska. This analysis included 23 studies. Medline and dissertation abstracts were searched. Cohort and case control studies were included. The studies were heterogeneous and the majority were of fair or poor quality. The result of the meta-analysis suggested that breastfeed is associated with a 50% reduction in SIDS risk but this result is highly suspicious due to severely flawed methodology.

The seasonality of SIDS has been noted since the 19th century. A Scottish study evaluated this factor in the light of the fall in the incidence of SIDS after advice on infant sleeping position was issued (Douglas et al. 1996). An analysis was carried out looking at the role of seasonality for the years 1988–1990 and for 1992–1994. Seasonality persisted as a risk factor. There is a lowering of risk, by one third, amongst babies born in February-May compared to those born in August-November. Once a mother has had an infant with SIDS there is an estimated risk of recurrence, possibly three to 10 times greater than in the rest of the population. The authors suggest that advice on subsequent pregnancy delivery date should be given to families who have already experienced SIDS. What reduces/eliminates the risk of death among infants in the first 6–8 weeks?

Six key health messages came out of the SUDI studies. These include the following:

  • Back to sleep: Babies should be put down to sleep lying on their backs. Sleeping on the back is preferable to sleeping on the side and sleeping prone should be avoided
  • Feet to foot: Babies should sleep in such a way that their head does not become covered during sleep. This is most easily achieved by putting a baby to sleep with his or her feet close to or touching the foot of the cot. Blankets are preferred to duvets and should be tucked in so that the baby’s head is exposed and uncovered without a hat.
  • Not too hot: Although it is important to prevent a baby becoming cold, becoming too hot is also a danger. Room heating is not required at night except when the weather is very cold. Babies’ bedrooms should be kept at approximately 16–200 C.
  • Smoke free zone: Cigarette smoking in pregnancy and around babies increases the risk of cot death. There should be no smoking around the baby’s sleeping place.
  • Prompt medical advice: The risk of cot death may be reduced by seeking prompt medical advice for babies who become unwell, particularly those with a raised temperature, breathing difficulties and who are less responsive than usual.
  • Bed sharing for comfort, not sleep: While it is likely to be beneficial for parents to take their baby into bed with them to feed or comfort, it is preferable to place the baby back into a cot to sleep. This is especially important if the parents smoke or have consumed alcohol.

The National SIDS Council of Australia reviewed the evidence concerning risk factors of SIDS and developed recommendations based on this review (Henderson-Smart, Ponsonby, & Murphy 1998). The National Health and Medical Research Council of Australia provides a methodology for guideline development which includes a systematic identification and synthesis of the evidence including quality assessment. However, the evidence tables and grading were not provided in the review article on SIDS. Thus the recommendations are categorized as ‘expert opinion.’ The forum developed three recommendations for which there was strong evidence:

  • Put your baby on the back to sleep, based on at least 19 retrospective case control studies with OR’s ranging from 1.2 to 14.1.
  • Make sure your baby’s head remains uncovered during sleep
  • Keep your baby smoke free before and after birth

This group also found that most studies suggest that there is no significantly increased risk of SIDS for babies of non smoking parents who bed share. They found that if infants become either too hot or too cold the risk of SIDS is increased, although this appeared to be related to sleeping position as well. The review group did suggest that it was important to try to maintain infants in a comfortable temperature zone. They found no consistent evidence that breastfeeding decreases the risk of SIDS.

Clinical Evidence published by the BMJ reviewed the effects of interventions to reduce the risk of sudden infant death (Creery & Mikrogianakis 2002). A broad based literature search was performed. Studies were chosen based on critical appraisal of abstracts and then summarised. Each topic was subsequently peer reviewed. A formal quality assessment of chosen studies was not undertaken. Based on this review, advice to avoid prone sleeping was supported. This was based on one non- systematic review and 12 observational studies which found that eight campaigns encouraging non-prone positioning and seven campaigns involving, among other recommendations, advice to encourage non-prone sleeping positions, were followed by a reduced incidence of SIDS.

Interventions including avoidance of tobacco smoke exposure, bed sharing and over heating were all based on studies which evaluated campaigns to reduce several risk factors and therefore it was not clear whether the effects were specifically due to any particular intervention. What are the risks and benefits of co-sleeping?

An investigation of the characteristics of parent-infant bed-sharing prevalence in England was published in 2004 (Blair & Ball 2004). Using sleep logs and interviews information on night time care giving was obtained for a broadly representative sample of 261 families in Stockton-on-Tees. Data on 1095 infants from the CESDI study of five English health regions were also used. Data from both studies found that almost half of all infants bed-shared, that is, slept with an adult for at least part of their night sleep, at some time with their parents (46% locally; 47% nationally). On any one night in the first month over a quarter of parents slept with their baby (27% locally; 30% nationally). Statistical analysis showed that bed sharing was not significantly related to younger mothers, single mothers or larger families and was not more common in the colder months, at weekends or among the more socially deprived families. Breastfeeding among young infants was strongly associated with bed sharing (p<0.0001).

The evidence about co-sleeping reviewed thus far indicates that:

  • Risk was significant only among smokers in the SUDI study, but not in the ECAS study.
  • Risk is increased if smoking and bed sharing are combined (Scottish Cot Death Trust).

Although co-sleeping as a usual mode of sleep increased over the time period studied, it could not be related to changes in the SIDS rate for the region. There was a positive interaction between co-sleeping and maternal smoking during pregnancy (Nordic study and European Union study).

In the European Union study the OR associated with bed sharing all night with an adult was progressively greater for younger infants whether or not the mother smoked.

A paper published in 1999 reported the results of the SUDI study specifically on the risk of babies sleeping with parents (Blair et al. 1999). In a three year period 325 cases and 1300 controls were identified. There was no increased risk of death for infants who shared the bed but were put back in their own cot. There were significant associations with infants who were found in the parent’s bed (9.78; 4.02–23.83), who shared a sofa (48.99; 5.04–475.60) and who slept alone (10.49; 4.26–25.81). Removal of three variables including death in the prone sleeping position, placed on a pillow or head covered halved the strength of the association of bed sharing and SIDS (4.62; 2.34–9.09). As previously noted there was no increased risk if parents were non smokers. If at least one parent smoked and the bed was shared, the OR was 12.35 (7.41–20.59) for SIDS. There was also no risk for infants older than 14 weeks. Among younger infants risk appears to be associated with recent parental consumption of alcohol (3.40; 1.88–6.16), overcrowded housing conditions(18.49; 3.62–94.48), extreme parental tiredness (2.42; 1.61–3.6) and the infant being under a duvet (3.97; 2.43–6.46).

Bed sharing and SIDS was assessed in a case control study reported in the BMJ in 1995 (Klonoff-Cohen & Edelstein 1995). Two hundred cases and two hundred matched controls were identified. The parents were interviewed by telephone. Routine bed sharing was assessed independently for day and night. The adjusted OR’s for both daytime and night routine bed sharing were not significant (OR 1.38; 0.58–3.22 and 1.21; 0.59–2.48 respectively). Bed sharing was adjusted for routine sleep position, birth weight, medical conditions at birth, passive smoking, exclusive breast feeding, intercom use and maternal age and education. Prenatal smoking was not included in the analysis. The study was underpowered for OR’s less than 2.9.

A study by Sheers et al (2003) compared the sleep location of 348 cases younger than 8 months at death with the sleep location of 4220 living infants younger than 8 months. The risk of suffocation was approximately 40 times higher for infants in an adult bed than those in cribs. The increase in risk remained high even when overlying deaths were discounted (32 times higher) or the estimate of rates of bed sharing among living infants doubled (20 times higher). The risk to younger infants also appeared to be greater when they slept in adult beds (87 deaths in adult beds at one month of age compared to 56 deaths at 5 months). This study did not control for potential confounders in the analysis.

The Chicago Infant Mortality Study (Hauck et al. 2003) was designed to examine risk factors for SIDS using a population based case control study. All 260 Chicago infant SIDS cases between November 1993 and April 1996 were evaluated. The adjusted OR for shared bed with mother alone or with mother and father was not significant but maternal smoking was not included as a potential confounder. Shared bed with others, meaning other children or children and parents had an OR of 2.0 (1.2–3.3).

A study of the role of parental bed sharing with consideration to parental weight was undertaken in Cleveland, Ohio (Carroll-Pankhurst & Mortimer 2001). An analysis of 84 SIDS deaths was carried out, 30 of which were bed sharing and 54 of which were non-bed sharing infants. Large maternal size was not associated with increased risk of earlier SIDS death in the absence of bed sharing but there was a significantly younger age of SIDS among infants of large mothers who shared a bed than among those who did not share a bed (p=.01). This study is limited by lack of a control group and by the fact that the population studied was urban and disproportionately minority and low income.

Bed sharing and the infant’s thermal environment was studied by Baddock et al (Baddock et al. 2004). Forty routinely bed sharing infants and 40 routinely cot sleeping infants aged 5–27 weeks were evaluated. Overnight video and physiological data were recorded including rectal, shin and ambient temperature as well as frequency of face covering events. The mean rectal temperature two hours after sleep was higher for bed share infants but not statistically significant (0.05, −0.03–0.14). However the rate of increase in rectal temperature in bed sharing infants was significant (0.01, 0.00–0.02) as was the frequency of face covering events (p<0.001) and the amount of bedding (p<0.001). The researchers conclude that bed share infants experience warmer thermal conditions than cot sleeping infants but appear to be able to maintain adequate thermoregulation to maintain a normal core temperature.

A study published in 2005 by Tappin et al (Tappin, Ecob, & Brooke 2005) described a 1:2 case control study conducted in Scotland which included 123 infants who died of SIDS between January 1996 and May 2000 and 263 controls. Sharing a sleep surface during the last sleep was associated with an OR 2.89 (CI 1.40–5.97) for SIDS. The largest risk was associated with couch sharing (OR 66.9, CI 2.8, 1597). Sharing a bed when <11 weeks (OR 10.20, CI 2.99, 34.8) was associated with a greater risk, p=0.01, compared with sharing when older (OR 1.07, CI 0.32–3.56). The association remained for young infants if mother did not smoke (OR 8.01, CI 1.20, 53.3) or the infant was breastfed (OR 13.10, CI 1.29, 133). Previous studies failed to show an associated risk for bedsharing if the mother was a non-smoker.

However age groups were not stratified. Despite small numbers and wide confidence intervals, these researchers support the revised UK Department of Health advice that the safest place for your baby to sleep is in a cot in your room for the first six months.

The benefits and risks of bed sharing and co-sleeping are presented in a guidance paper published by the Royal College of Midwives (RCM) (Royal College of Midwives 2004). This guidance is based on good practice recommendations. The RCM advises:

  • Parents should be informed of the benefits of bed sharing and co-sleeping including successful breastfeeding and better sleep;
  • Parents should be discouraged from:

    co-sleeping on sofas and armchairs;


    bed sharing and co-sleeping at home if either parent is a smoker, has recently had alcohol, is on sleep inducing medication or illicit drugs, is unwell, excessively tired or unaware of body periphery;


    co-sleeping if the baby is bottle feeding, is bed sharing or co-sleeping with father only or if the father comes to bed later;


    allowing the baby to bed share or co-sleep with children or bed share with unwell or restless children


    using heavy quilts duvets, electric blankets or pillows;


    sleeping on waterbeds, feather, soft or sagging mattresses, beanbags and V shaped pillows;


    permitting gaps between the mattress, furniture and wall.

A qualitative study was conducted in North Tees, UK to examine parents motives for sleeping with their infants (Ball 2002). Sleep logs were recorded and then mothers (n=253) were interviewed at the end of the first and third months postpartum using a semi structured interview format. Forty seven percent of the babies in the sample bed shared with their parents at least once during the first month. The most common reason given by mothers for sleeping with their infants was the ease and convenience of night-time breastfeeding. Among mothers who ever breastfed, 65% of infants had co-slept. Among those who breastfed for a month or more, 72.3% of infants were bed sharers. For the 19% of families where bed sharing was unrelated to breastfeeding, settling the baby who was either irritable or ill, was the reason for bed sharing. Eleven families described having their baby sleep in their bed during periods of overt illness. Other bed sharing motives included:

  • Fear of baby dying
  • For enjoyment to increase time spent with baby
  • To ease maternal discomfort following birth
  • Nowhere else for the baby to sleep
  • Parenting ideology: family bed

7.8.3. Child Abuse What tools exist to identify the child at risk of abuse?

The prevention of child abuse depends upon correctly targeting families who are at increased risk of maltreatment. The use of screening tools to identify such families presents a number of ethical and practical concerns. These include the stigmatisation of parents identified as being at risk and the application of accurate screening measures. The causes of child abuse are multifactorial and developing tools with adequate sensitivity (e.g. percentage of maltreated children correctly identified as being at risk), specificity (e.g. percentage of non-maltreated correctly identified as not being at risk) and positive predictive value (e.g. percentage of the high-risk group who go on to maltreat) is fraught with difficulties. Peters and Barlow (Peters & Barlow 2003) assessed the accuracy of instruments specifically designed for use in the antenatal and postnatal periods. A systematic review resulted in 8 studies which had prospectively evaluated a standardized instrument. One study was exclusively an antenatal intervention and one tool was implemented when a child was admitted to intensive care. Most of the more methodologically rigorous studies yielded sensitivity greater than 50%. However a majority of these studies had a positive predictive value less than 50%, which means that the majority of children categorized as high risk were not actually maltreated. It has been suggested that acceptable criteria for instruments that predict maltreatment are a sensitivity of 40–60%, a specificity of 90% and a positive predictive value of 25%. Only one instrument reached these levels for all three operator characteristics. Despite this level of accuracy 75% of families who have been predicted to be at increased risk will not go on to abuse.

In U.S. law, child abuse and neglect is defined as, “…any recent act or failure to act on the part of a parent or caretaker which results in death, serious physical or emotional harm, sexual abuse or exploitation, or an act or failure to act which presents an imminent risk of serious harm. “ A recent systematic review for the US Preventive Services Task Force (Nygren, Nelson, & Klein 2004) evaluated the benefits and harms of screening children in primary health care settings for abuse and neglect. Evidence on screening instruments and the effectiveness of interventions was examined. Six studies on screening methods met the eligibility criteria but none of them addressed the effectiveness of screening in reducing harm and premature death and disability. A meta analysis was not done. Four of the studies were rated as poor, one fair and one good-fair. Three screening instruments had a fairly high sensitivity but low specificity, that is, they did not correctly identify the non-abusers. Screening instruments require high sensitivity and specificity because falsely implicating a parent as a potential abuser may have serious consequences. A Canadian systematic review update (Macmillan 2000) found that the primary difficulty with approaches to screening for risk of child maltreatment was the unacceptably high false positive rates. They reviewed three studies published between 1993 and 1999. The predictive validity was either not measured or was very poor. They concluded that the harms outweighed the benefits of screening for risk of child maltreatment.

One study (Browne 1989) assessed the predictive value of a checklist of risk factors utilised by home visitors in Surrey. An initial screening checklist of thirteen factors was collected on 62 abuse cases and 124 matched non-abusing families. Risk factors were weighted using discriminant function analysis and it was found that the sensitivity of the checklist was 82% and specificity was 88%. This means that 18% of abusers were missed and 12% of non abusers were falsely identified as high risk. When the checklist was applied prospectively to a sample of over 14,000 births it was found that 6.7% of Surrey families with a newborn were assessed to be high risk. On follow up, only one out of 17 of these ‘high risk’ families went on to abuse a child within two years of birth. The researchers hypothesized that risk factors may be buffered by secure relationships within a family.

A follow up study on maternal attitudes and mother infant interaction comparing high and low risk families was also conducted by the Surrey group. During home visits the mothers and children were observed and a questionnaire was completed. It was found that abusing parents had more negative conceptions of their children’s behaviour than non abusing parents as they perceived their children to be more irritable and demanding.

The Brigid Collins Risk Screener was assessed in a small sample of 49 expectant mothers who were screened on risk factors covering six areas (Weberling et al. 2003). A trained interviewer blind to family risk status completed a three month postpartum measure of parenting stress and the family environment during a one hour home interview and observation. Parental stress scores were not related to the total risk score. The only environmental factor which was significant between high and low risk groups was the availability of learning materials in the home (p<.05). Abuse outcomes were not measured.

NICE guidance entitled ‘Identification of children with suspected abuse’ has recently been commissioned. Publication date is to be announced.



(UNICEF. 2005) gives a breastfeeding rate of 71% at birth for England and Wales (2000) falling to 43% at 6 weeks and 14% at 9 months.