Article Text


Child protection

Statistics from


C. Hamer, J. Heron, A. Emond the ALSPAC Study Team Centre for Child and Adolescent Health, University of Bristol; Hampton House, Cotham, Bristol BS 6 6JS

Aim: To investigate the incidence and the causes of burns and scalds in young children, using an ecological model based on data from a representative population based cohort.

Method: With ethical approval, questionnaire data on injuries were collected from parents in the Avon Longitudinal Study of Parents and Children (ALSPAC), when their children were 6, 15, and 24 months of age. Developmental scores were calculated from responses to questions modified from the Denver Developmental Screening Test.

Results: A total of 1648 burn and scald incidents were reported in the 13 971 children in the study in the first 2 years. Burns were reported in 1.5% (173) infants at 6 months, 6.3% (682) at 15 months, and 7.8% (793) at 24 months. Hands were the part most frequently injured at all ages, with hot water and hot drinks being the most common cause at 6 and 15 months and kitchen appliances at 24 months. There were more boys than girls at each time point, and children from ethnic minorities were over represented at 6 and 15 months, but not at 24 months. The children with the best motor scores at 18 months were more likely to sustain a burn at 15 and 24 months, whereas those with the best fine motor scores had less burn incidents at 15 months. Family factors significantly (p<0.05) associated with burns included single and teenage mothers, and having additional adults (other than partner) in the house. Maternal educational level was positively associated with burn rates at 6 months, and negatively related at 15 and 24 months. Environmental factors significantly associated with burns included housing tenure, overcrowding, and lack of a car. After multivariable analysis, the following factors remained independently significant: male sex, maternal age and educational level, housing tenure, and higher gross motor and lower fine motor scores.

Conclusions: Minor burns and scalds are common in young children, and are associated not only with adverse environmental and familial factors, but also with the developmental progress of the child.


P. M. Barnes, M. James-Ellison, A. Maddocks, K. Wareham, W. Dickson, R. Lyons, H. Hutchings. Swansea NHS Trust, Welsh Regional Burns Centre, Morriston Hospital, University of Wales, Swansea

Introduction: Thermal injuries are an important cause of morbidity and mortality in childhood. Over 50% of burns occur in children aged under 3 years. Little is known about the health and development of young children in the UK in the years following a burn. We believe that a burn in early childhood is a marker for future problems in relation to the child’s health and developmental progress.

Methods: A retrospective case control study of children under 3 years old, who were admitted to the Welsh Regional Burns Unit, over the period 1994–1997 was conducted. Controls were matched for age, sex, and socioeconomic background. Information was gathered from hospital notes, social service records, schools, and community child health records to determine health and developmental progress up to the end of the first year in full time education.

Results: A total of 145 case control pairs were studied (58% boys). Burns occurred at a mean age of 17 months and were typically sustained at home (88%) by scalding (68%). The majority were thought to be accidental. By the age of 6 years, significantly more developmental concerns were identified in cases (25% v 18%), and uptake of developmental screening was poorer in case children. Hospital admissions (unrelated to the burn) were more common in case children, and more cases were known to social services, excluding referrals made at the time of the burn (32% v 18%) (all p<0.05).

Conclusion: Following a burn, young children suffer significantly poorer physical and developmental health than a matched control. In the aftermath of a burn, all young children should receive a full health and developmental assessment, with close supervision indicated in the years that follow.


C. Breathnach, E. Finan, F. Sharif, M. Kelly, A. J. Nicholson. Our Lady of Lourdes Hospital, Drogheda, Ireland

Background: Despite legislation, many young children in Ireland travel in cars without appropriate child restraints.

Aims: (1) To study the impact of an antenatal educational intervention on car seat use for the journey home and (2) to study all car passenger injuries and fatalities under 5 years of age presenting to hospital over a 6 year period (1997–2003).

Methods: A prospective cross sectional observational study of the means of transport of newborn infants home was carried out for four 6 week periods from 1997 to 2002. In 1997, no information was given and we observed whether mothers used a car seat for the journey home. From January 1998, a 30 minute educational intervention highlighting the importance of car seats and demonstration of their use was delivered by a midwife at the final antenatal class. Prospective electronic injury surveillance took place using the minimal dataset and all car passenger injuries under 5 years old were studied. Data collected included the date, day and time of the injury, a text description, injury diagnosis, whether admitted or not, and use of appropriate child restraints. Data analysis was via an extractor software package.

Results: Car seat use on discharge from the maternity unit rose from 16% (32/198) in 1997 to 93% (264/283) in 2002 (p<0.001). In 1997, car seat use was associated with higher social class (p<0.01) and breastfeeding (p<0.05). Subsequent cohorts demonstrated car seat use among all social classes but a persistent non-compliant group (asylum seekers and those leaving hospital in a taxi) was evident. Of 157 under 5 year old car passenger injuries, 100 (64%) occurred between 12.00 and 20.00 hours and the highest numbers (54%) presenting at weekends. There were two deaths, both unrestrained rear seat passengers. Child restraint use rose from 32/70 (46%) in 1997–2000 to 57/87 (66%) in 2001–2003 (p<0.01). Front seat position dropped from 10/70 (14%) in 1997–2000 to 1/87 (1%) in 2000–2003. Admissions halved over the same time periods (18.5% v 9%). Head injuries were reduced in 2000–2003 cohort (p<0.05).

Conclusions: Observed car seat use improves with an antenatal educational intervention with subsequent reductions in car passenger head injuries and admissions.


F. Khan, M. C. J. Rudolf, B. Mulley, Z. Smith, T. Gaussen, K. Naylor. Growth and Nutrition Team, Belmont House, East Leeds PCT

It is current practice in the UK to screen for failure to thrive, even though weight monitoring does not fulfil the screening criteria. As part of an audit on failure to thrive, the growth of children seen in a secondary referral clinic was reviewed to ascertain the extent of their change in weight, and how it is affected by the severity of the growth pattern and the presence of eating difficulties. It was hoped that in this way references could begin to be established that might be used for future audit and research.

Methods: The Growth and Nutrition Team database was reviewed for the period January 2001 to October 2003. This database records children’s weight over time along with the paediatrician’s assessment at the first appointment of the severity of the growth pattern and any eating difficulties. The data were analysed for change in weight SD at 6 months, 12 months, and at discharge for the sample as a whole, and then for those with concerning growth alone, significant eating difficulties, and those with a combination of the two.

Results: 164 children (64 boys) aged 24 months (SD 25.2 months) were referred, of whom 99 were seen on more than one occasion. 67 had a growth pattern that had been considered as definitely concerning, 54 had significant eating difficulties, and 38 a combination of the two. Weight (SD) at referral was −2.1 (1.3) with 56 below the 2nd centile. By 6 months weight had risen by 0.17 (95% CI −0.01 to 0.34) and by 12 months there was a significant improvement with an increase in weight of 0.28 (95% CI 0.05 to 0.5, p<0.02). For those discharged from the clinic the weight had increased by 0.46 (95% CI 0.23 to 0.7, p<0.0001). Those with a combination of eating and growth concerns did best with a weight increase of 0.7 (95% CI 0.2 to 1.2, p<0.01) at discharge.

Conclusion: These data give some parameters for anticipated weight change for children referred with failure to thrive and/or eating difficulties. It suggests that follow up beyond 12 months may be required, and that, perhaps surprisingly, children with significant eating difficulties may show more weight gain than those without.


R. Gilbert, S. See, M. Harden, S. Logan. Institute of Child Health

Aims: To determine the effect of research evidence on recommendations about sleeping position, the prevalence of infant sleeping position, and mortality due to sudden infant death syndrome (SIDS) from 1940 to 2002.

Methods: Systematic review of comparative studies of the effect of infant sleeping position on sudden unexpected infant death and historical review of advice to clinicians and parents in UK publications, 1940–2002.

Results: Systematic review: 25 studies were included (2937 articles reviewed). None showed a beneficial effect of prone sleeping. In 1965, evidence favoured a protective effect of supine sleeping. By 1986 there was evidence of significant harm from prone sleeping. Historical review: 87/189 books referred to sleeping position. Recommendations favoured supine/side sleeping until the mid 1950s, prone/side position thereafter until 1990, supine/side till 1994, and then supine. Prone sleeping position increased in the 1970s–80s in Western industrialised countries, and decreased sharply in the early 1990s, except in the USA where prone sleeping was higher for longer. SIDS incidence followed a similar pattern. Health promotion campaigns lagged behind evidence of harm by at least 5 years.

Conclusion: A major shift to prone sleeping position was contrary to the available evidence and increased the risk of SIDS fourfold. Lack of a culture of evidence appraisal and systematic review delayed the effective health promotion of supine sleeping and the prevention of deaths due to SIDS.


A. M. Cairns, J. Y. Mok, R. R. Welbury. Glasgow Dental Hospital and School, Royal Hospital for Sick Children, Edinburgh

Aims: To identify the incidence of orofacial (head, neck, or face) injuries found within a cohort of physically abused children and to compare the results with previously published studies. To examine the demographic data surrounding this subject in relation to the alleged perpetrator, the location in which the alleged assault occurred, the mechanism of injury, and the actual orofacial injury sustained.

Methods: Ethical approval was granted by the Lothian Region ethics committee. 750 cases of suspected child physical abuse over the previous 5 years were identified using the database held by the CPS at the Royal Hospital for Sick Children in Edinburgh. A data sheet with collection points based on previous studies was devised and information extracted from the case records.

Results: 394 records were available for review. 57.1% of cases were girls and 41.6% were boys. 52.5% of all alleged abuse occurred in the child’s home. Other locations included outside in a public place, school, and at the home of the alleged abuser. The alleged perpetrator was the child’s mother in 26.1% of cases, the father in 25.1%, and the mother’s partner in 12.1%. 58.6% of the children had obvious signs of abuse on the head, face, or neck. 22.5% had been punched or slapped around the orofacial region, 16% had been stuck by an object, and 15.6% had allegedly sustained multiple modes of injury. Bruising to the head, neck, or face was seen in 79.6% and abrasions in 29.4% of the children with orofacial signs. 64.5% of the children with orofacial signs had visible bruising on their face and 22.9% had obvious abrasions on their face.

Conclusions: This study is the largest of its kind in the UK, and illustrates that many of the injuries caused by physical child abuse are visible to the dental practitioner. It concurs with similar studies carried out in the US, which have found that around 60% of physically abused children have orofacial signs of abuse. Intraoral signs were infrequently reported but routine examination by a paediatric dentist was not standard protocol.

G114 Which fractures are seen in physical child abuse?: a systematic review

S. Maguire, A. Kemp, M. Mann, S. Harrison, F. Dunstan, J. R. Sibert Welsh Child Protection Systematic Review Group University of Wales College of Medicine, Academic Centre, Llandough Hospital, Penlan Road, Cardiff CF64 2XX

Background: Clinicians have to decide whether a fracture has arisen from child abuse or not. This systematic review aims to inform this decision.

Methods: An all language literature search of original articles, references, textbooks, and conference abstracts 1970–2003 using ASSIA, Caredata, Medline, Child Data, CINAHL, Embase, PsychINFO, Social Science Citation Index, and TRIP databases. Inclusion criteria was distribution of non-accidental fractures in children <17 years. We excluded review articles and expert opinion. Two independent reviews (+/− third review), standardised criteria for study definition, data extraction, and critical appraisal were used.

Results: 79 papers were included. Abusive fractures are often multiple, 80% occur before 18 months whereas 85% of accidental fractures occur after 5 years. Forest plots of key fractures highlight the heterogeneity of studies. The probability of femoral fractures being due to abuse is 12.4%, excluding studies of isolated femoral fracture increases this risk to 24%. Spiral fractures are the commonest abusive fracture in non-walking infants but are not discriminatory once they walk. Rib fractures due to abuse are prevalent in infants, involve multiple ribs, are often bilateral, the majority (80%) have other associated injuries. The commonest accidental and abusive skull fracture is a linear parietal fracture. Complex or depressed fractures appear more common but this disappears when fatally abused are excluded. Up to 90% of abusive humeral fractures occur under 15 months, only 6% of abusive fractures are supracondylar v 66% of accidental. Less common but significant abusive fractures are: vertebrae, clavicle, pelvis, hands, and feet.

Conclusions: Child abuse must be excluded as a cause of fractures in children under 18 months. Multiple fractures, spiral fracture of femur in non-walking infants, mid shaft and metaphyseal fracture of humerus, and rib fracture in the absence of major trauma or underlying bone disease are strong indicators of abuse. More detailed case control studies of fractures in abused and non-abused children are urgently needed.


S. Maguire, A. Kemp, M. Mann, S. Morris, J. R. Sibert Welsh Child Protection Systematic Review Group University of Wales College of Medicine, Academic Centre, Llandough Hospital, Penlan Road, Cardiff CF64 2XX

Background: Fractures constitute 20−50% of injuries in physically abused infants and toddlers. They are commonly multiple, of different ages, and occult. The optimal investigations to identify these fractures are not defined. This systematic review addresses this question.

Methods: We performed an all language literature search of original articles, their references, and conference abstracts, 1970−2003. We used ASSIA, Caredata, Medline, Child Data, CINAHL, Embase, PsychINFO, Social Science Citation Index, HealthSTAR, and TRIP databases. We conducted two independent reviews for each study, with a third review if disputed, and used standardised criteria for study definition, data extraction, and critical appraisal. We included papers addressing radiology investigation in suspected child abuse and excluded review articles and personal practice.

Results: Forty one of 373 studies reviewed were included. Skeletal survey (SS) and bone scan (RNS) technique varied over time. Nine of the 12 papers comparing SS with RNS were published before 1989. Skeletal survey missed 3.9 times as many fractures as a bone scan, excluding skull (SS missed 54 and RNS missed 14 fractures in 417 children screened). Oblique view of ribs in skeletal survey increases the yield of fractures by 7% (p 0.004) (1 RCT). Repeat skeletal survey increased the yield by 27% (p 0.005) (1 longitudinal study). Yield for occult fractures in children<1 year of age was 50−80%, in 1−2 year olds there was 12.5−17%, and in 2−3 year olds there were insufficient data.

Conclusion: All children under 2 years with suspected abuse should be investigated to exclude occult fractures. As SS or RNS alone will miss fractures, the highest yield is with both investigations. Addition of oblique chest x ray or repeat survey will increase the yield from SS alone. RNS alone must be accompanied by skull x ray and coned views of metaphyses should be considered. We recommend further research using current imaging techniques to define diagnostic accuracy of each of these investigation strategies.

View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.