Childhood injury in Tower Hamlets, one of the most deprived inner-city London boroughs: prospective audit data from the Royal London Hospital

Abstract

BackgroundThe most socioeconomically deprived children are at greatest risk of injury. Tower Hamlets is one of the most deprived areas in England. Data obtained for injuries are routinely inadequate to inform or assess prevention strategies. Enhanced injury data collection could contribute towards the establishment of national injury surveillance. MethodsWe did a prospective audit of injuries in patients aged less than 16 years presenting to paediatric A&E, Royal London Hospital, between June and December, 2012. Staff gathered patient data by hand 24 h a day, 7 days a week using a questionnaire based on the enhanced injury dataset of the College of Emergency Medicine. The sample gathered was an opportunistic sample and the inclusion criteria were all children presenting with injury within the specified period. Seriously injured patients, assessed by the trauma team, had data recorded in the paediatric intensive care unit database, which was then used to supplement our audit sample. The questionnaire was used to identify injury characteristics, date and time of the injury, and the patient's medical record number (MRN). Age, sex, ethnic origin, and hour of arrival were extracted from the NHS Care Record Service (CRS) dataset with the MRN. Primary outcomes were location and activity at the time of injury, sport involvement, mechanism of injury, substance involvement, and whether intentional or unintentional. The association between location and injury rates was tested by Spearman's rank correlation coefficients. FindingsThe audit captured 471 of the 6358 under-16 injury-related attendances recorded on CRS, of which 414 (6·5%) had valid MRN numbers. Of these 414, 97 were subsequently admitted. The audit dataset was representative of the CRS dataset for sex but not for age or arrival time at A&E; figures compared with the CRS dataset were men 64·0% versus 59·8% (p=0·0813); mean age 8·3 years versus 6·7 years (p<0·0001); and mean hour of arrival 1545 h versus 1510 h (p=0·0035). Leisure (32·9%, 95% CI 28·3–37·4), sport (18·1%, 14·4–21·8), and road traffic collisions (11·4%, 8·3–14·4) were the three most common activities. Falls (49·8%, 95% CI 44·9–54·6) and blunt force (16·2%, 12·6–19·7) were the most common injury mechanisms. Injuries occurred mainly at home (32·9%, 95% CI 28·3–37·4), in schools (22·0%, 18·0–26·0%), and on roads (20·3%, 16·4–24·2). Patients subsequently admitted were more likely to have been involved in a road traffic collision than those simply attending A&E (29·9% vs 11·4%; p<0·0001) or to have had a fall from a height above 1 m (24·7% vs 7·5%; p<0·0001), and 32·0% (95% CI 22·7–41·3) of admissions had head injuries. There was a positive correlation between injury rates by home address and income deprivation (p<0·0001). InterpretationThis study highlights the importance of including routine injury data in existing computerised admission systems. Data collection was labour intensive and needed staff commitment. Despite efforts made to inform all medical staff and promote a uniform collection of data, data collection proved suboptimum with variability across shifts, and our final sample was small and unrepresentative of age and time of arrival. The retrospective entry of the trauma patients might have led to over-representation of serious injuries. FundingNone.