Aim: To estimate the incidence and mortality rates for unintentional fall injuries in children under 5 years of age in three developing regions of the world.
Methods: This is a systematic review of literature on unintentional childhood fall injuries. A computerized PUBMED search of literature published between 1980 and 2006 was conducted and a manual search of journals was also completed.
Results: Over 140 000 injuries to children under 19 years were reported in 56 studies (21 from Asia, 20 from Africa and 15 from South America); on an average 36% of injuries (52 575) were due to falls. The median incidence is estimated at 137.5 fall injuries per 100 000 children. The incidence of falls specific to the under-5 age group was reported in 16 studies with a median incidence of 40.6 falls per 100 000. The overall average incidence rate for childhood falls is highest in South America at 1315 followed by Asia at 1036 and Africa at 786 per 100 000, respectively. Average mortality rates were highest for Asia at 27 followed by Africa at 13.2 per 100 000, respectively.
Conclusion: This review demonstrates that the burden of falls on children has not been well documented, and is most likely under-reported. With the large and growing population of children in developing countries, the public health implications of the observed results are tremendous. Appropriate prevention relies on accurate statistics.
Unintentional injuries have been recognized as one of the leading causes of death among children and young adults (1). Over 700 000 children die every year as a result of injuries, especially in developing countries where 13% of the total burden of disease among children less than 15 years old has been attributed to injuries (2). Environmental, social and developmental factors make children highly vulnerable to injuries (3). Among children under 5 years of age the impact of injury is often obscured by the substantial burden of perinatal and infectious diseases; injuries are the leading cause of death for children after their first birthday. There is also high morbidity associated with childhood injuries: for every injured child who dies, there are several thousand children who live with varying degrees of disability. A large proportion of these injuries occur in either the home or in leisure environments.
The vast majority of child injuries are unintentional injuries, which encompass a wide range of conditions such as road traffic injuries, falls, burns, drowning and poisoning. Globally, an estimated 391 000 people of all ages died due to falls in 2002, making it the second leading cause of unintentional injury death globally after road traffic injuries. In contrast to high-income countries (HIC), the burden and pattern of child injuries in developing countries is poorly known and not well studied. The problem of child injuries is further compounded by the population distribution in these countries. According to data for 2005, 141 million children aged 0–4 years lived in Africa, comprising approximately 23% of the world's child population (2). By contrast, only 10% of the world's child population is found in HIC.
This article attempts to quantify the burden of falls in three developing regions for children aged less than 5 years based on published literature. The overall purpose of this article is to define the mortality and morbidity due to falls among children less than 5 years old in low-income countries in Asia, Africa and Latin America. The article does this by estimating injury-specific incidence and fatality rates for these regions. This work will attempt to highlight the implications of falls for child health policies in these regions.
A systematic PUBMED (http://www.ncbi.nlm.nih.gov), EMBASE (http://gateway2.ovid.com) and POPLINE (http://db.jhuccp.org/popinform/index.stm) search of the literature published between 1980 and October 2006 was conducted to retrieve data on unintentional child injuries in the developing nations in Africa, Asia and Latin America. Combinations of following keywords were used: falls, injury, paediatric, childhood and child with country names from each region (see Appendix S1). The search was extended by using the ‘related articles’ link on PubMed. During review of the articles, additional papers were identified from the respective reference lists; and a manual search of journals was also done. Only articles published in English or at least with an English abstract were considered for review; approximately 10% of papers in the international peer reviewed literature were in another language, predominately Spanish and Portuguese from Central and South America.
A search for accessible web-based data was also conducted through the search engine Google (http://www.google.com) and selected websites such as the World Health Organization (http://www.who.int) and the Transport Research Laboratory (http://www.trl.co.uk/800/SEARCHPUBS.asp). A search for reports of country-based or nationally representative samples in such ‘grey’ literature was also undertaken.
Studies considered for inclusion were restricted to original reports of facility- or community-based studies giving estimates of child fall injuries, studies giving proportional mortality data, critical reviews that provided reports of original research results and papers with quantitative data. The review focused on studies that reported data on children under the age of 5 (0–4 years). Studies excluded from the review included those that did not give the size of study population (such as small case series), in which the criteria for recruiting subjects and age groups were not defined, did not give estimates for incidence or mortality rates, and those that focused on special populations such as refugees or low-income populations within developed countries. Studies that were limited to clinical management, contained data prior to 1980, included intentional injuries (violence, suicide), those limited to adult populations, limited to specific paediatric subgroups, such as children with epilepsy, studies of birth trauma, and studies related to natural events or disasters were also excluded. Although the review focused on reported data on children under 5 years of age (0–4 years), studies vary in the age groupings of their results and thus any data that included the age group of interest (such as 0–19 years, or cohorts with older children) have been included.
Using Microsoft® Excel 2003, data tables were created for each region. The estimated population of children in the study area was taken directly from the study if included. Otherwise, the population less than 5 years was calculated using the population percentage in the 0–4 age range from UNICEF's country-level statistics (http://www.unicef.org/infobycountry/index.html), and either the total study catchment population (if included) or the city population taken from the worldwide web (http://www.citypopulation.de/country.html). Incidence and mortality rates were subsequently calculated using this derived population under age 5 as the denominator.
The search for articles in indexed journals yielded 56 papers from developing nations in Africa, Asia and Latin America (Table S1). Twenty articles were retrieved for Africa, from 9 countries; 15 articles for South and Central America, from 8 countries; and 21 articles for Asia, from 14 countries, with one article providing data for countries in all three regions. Due to the variability in data and broad range of rates reported by community-based studies, the imperative to estimate a single national rate was challenging. Country data were therefore summarized at the regional level to estimate annual regional incidence and fatality rates. Table S1 presents the methods used in each of the relevant studies pertaining to the under-5 age groups. The majority of the papers that meet the inclusion criteria for this study are facility-based studies. Of the 56 studies in this analysis, 17 are community-based studies and 35 are facility-based studies; 4 studies used both methods (Table S1). The majority of community-based studies employ household surveys for data collection, whereas facility-based studies primarily rely on medical records or surveillance systems. While 13 studies use coding systems, such as the International Classification of Diseases (ICD), for classifying the childhood injuries included in the study, most studies do not discuss a definition of injury or use an informal definition. Only eight of 50 studies use a standardized definition for coding injuries resulting from falls. While all papers report epidemiological descriptive statistics, only 12 use logistic regression analysis to analyze risk factors for injury or falls.
The median incidence of fall injury in Asia in those aged 0–18 years was found to be 170 per 100 000 children averaging 43% of all injuries (Table S2) (see Tables S1 and S2 for citation of references 14–55). Among those less than 5 years in Asia there was a median incidence of 58.2 falls per 100 000 children, comprising 35% of all injuries. A study by Bener et al. from the United Arab Emirates revealed the highest incidence at 1923 per 100 000 far outnumbering injury by any other cause, whereas the study by Mukhopadhyay from India revealed only 271 fall injuries per 100 000 (4,5). Only two studies of children under 5 years of age cited mortality rates, giving a median of 38.5 per 100 000. The majority of children injured by falls (65%) in this region were male.
In Africa, the median incidence of fall injury in those aged 0–21 years was found to be 40.6 per 100 000 averaging 41% of all injuries. Among those 0–5 years there was a median of 4.6 falls per 100 000 comprising 25% of all reported injuries. In looking at the studies among those aged 0–5 years, the study by Mwaura et al. from 1994 revealed the highest incidence at 109 per 100 000, followed by 43 per 100 000 in the study by Abantanga et al. from Ghana, yet neither contained mortality data (6,7).
In South and Central America, unlike the other regions, only Gaspar et al.'s study included an under 5-year-old subset analysis (8). The study calculated an incidence of 500 per 100 000, comprising 53% of all injuries in Ipatinga, Brazil. For all children and adolescents less then 20 years, additional studies are available. For example, Del Ciampo et al. report an incidence of falls of 2700 per 100 000 (constituting 47% of all injuries) and Bangdiwala et al.'s study in Cuba revealed falls of 1378 per 100 000 fall injuries (representing 52% of all injuries; 9,10). Only one study by Baracat measured mortality due to falls (0–14 years) at 0.44 per 100 000 in Brazil (11).
The regional injury rates from falls in the 0–4 age group differ dramatically between Asia (median rate of 58.2 per 100 000) and Africa (4.6 per 100 000). The rate of 501 per 100 000 in South America is only based on a single study from Brazil, and though much higher, could be considered an ‘outlier’ at this stage in the absence of further support in the literature. Fall regional mortality rates stood at 38 and 1 per 100 000 children aged 0–4 years, respectively, for Asia and Africa, paralleling the large differences seen in fall injury rates. No mortality estimates for the 0–4 age group could be found in South or Central America.
Twenty-three papers included in this review report one or more risk factors for childhood falls. The most frequently reported risk factors are age of child, sex of child and location of injury. Seven studies report that children under 1 year are at a greater risk of suffering a fall. Five studies report that falls are more likely to occur at home, compared to two studies that report that falls occur more often outside the home. A greater rate of falls was reported among males in four studies, whereas no studies showed a greater risk for females. Additional risk factors reported in multiple studies include regional risk differentials, maternal education level, presence of a mental disorder in the child's caregiver and housing quality or material.
This study is a systematic literature review of the burden of injuries due to falls that aims to calculate regional injury and mortality incidence. In a search of peer-reviewed literature published over 25 years, only 56 studies measuring unintentional injuries in children were found. Within these studies, less than 20% enumerated mortality due to falls, only 12% of studies use standardized or formal definitions for falls and no data on the severity of injuries or disability were available. There was a significant lack of data, especially population based, from Asia, Africa and South America. Many studies failed to include a 0–4 year old age breakdown or comment specifically on falls by age category even if it was the leading cause of injury. Less than half of the included papers analyze or identify factors associated with an increased risk of falls.
The estimates of regional fall rates presented here need to be interpreted with caution. Although they do reflect a significant burden on children under the age of 5 years, they are supported by few studies and vary widely between the regions. Consequently, the true burden of falls among children may be substantially underestimated. Furthermore, the lack of robust empirical data impede efforts to consider the most effective (and cost-effective) strategies to address falls in these regions. Comprehensive recording of the mechanism of the fall and a complete analysis of risk factors for falls are missing from almost all papers in this review.
According to the World Health Organization (WHO) data (12) from 2002 onwards, fall mortality for Africa and Southeast Asia was 1.7 and 1.9 per 100 000 population aged 0–4 years, respectively. These rates are a fraction of those in the study for Asia and several fold less than that for Africa (4.6 per 100 000). Overall fall morbidity or disability-adjusted life years (DALYs) were not calculated by age group in the WHO data; however, those aged 0–4 years contributed to 23% (second largest age group) of DALYs lost from falls globally. Among all ages, the Southeast Asian region contributed 26% of the total DALYs lost from falls, almost three times that of Africa (9%). Overall, the WHO data parallel the trend in the results of this literature review, supporting higher fall morbidity rates in Asia over Africa. However, without disaggregate age group analysis for DALYs lost per region it is difficult to compare the literature review with WHO morbidity data.
The reviewed studies often did not contain specific 0–4 year age groupings for injury data and frequently had no mention of study site populations. When population sizes were not included, the total population of the city or town was utilized. This is expected to be larger than a particular hospital's catchment area, giving a lower than expected incidence. Many studies listed the percentage of falls but failed to breakdown falls by age group and therefore could not be included. There was an extreme dearth of literature on injuries of children under 5 years in Central and South America.
The review was limited to articles published in English or at least with an English abstract, which results in a publication bias. We made an effort to exclude data on intentional injuries when possible, but many studies did not distinguish between unintentional and intentional injury, and there is a potential for misclassification. Most studies were facility based, and cases presenting to a hospital or gaining admission are not likely to be representative of all injuries experienced in the community. Most of the facility-based studies were retrospective reviews of hospital admissions, and there were very few prospective studies with follow-up of patients to study the type of treatment given and its outcome. Further, it is difficult to generalize the experience of one or two facilities to an entire country or continent. The quality of care, and therefore the case-fatality and mortality risk, likely differ among the various facilities and over the time periods reviewed. Similarly, community-based study sites represent diverse economic, cultural, geographic and environmental settings that will determine the magnitude, type and severity of injuries observed. Extrapolating from a limited number of studies is, therefore, problematic. However, although it is challenging to draw conclusions from studies of such diverse nature, it is clear that there is a stream of childhood injuries coming to health care facilities in these regions.
This study is an important step in exploring the burden of injuries to children in the developing world. The main goal of this article is to emphasize the lack of data on child fall injuries and as a consequence, the potential to undercount the impact of this public health problem. There are a limited number of studies that include fall data for children and even fewer that include mortality data. This is despite WHO's work that ranked falls as the leading cause of injury burden in those aged 0–4 years (13). This is a significant, and preventable, public health burden and deserves urgent and immediate action. However, the development of adequate interventions is predicated on accurate and detailed data on risk factors for falls.
This article was supported in part by the Child Health and Nutrition Research Initiative (http://www.chri.org). We would also like to thank the following individuals for their valued revisions: Margie Peden, Olive Kobusinjye, David Bishai and Andrea Gielen.