Anda di halaman 1dari 6

ORIGINAL ARTICLE

Targeting Pediatric Pedestrian Injury Prevention Efforts: Teasing the Information Through Spatial Analysis
Mindy Statter, MD, Todd Schuble, MS, Michele Harris-Rosado, RN, Donald Liu, MD, PhD, and Kyran Quinlan, MD, MPH

Background: Pediatric pedestrian injuries remain a major cause of childhood death, hospitalization, and disability. To target injury prevention efforts, it is imperative to identify those children at risk. Racial disparities have been noted in the rates of pediatric pedestrian injury and death. Children from low-income families living in dense, urban residential neighborhoods have a higher risk of sustaining pedestrian injury. Geographic information systems (GIS) analysis of associated community factors such as child population density and median income may offer insights into prevention. Methods: Using trauma registry E-codes for pedestrian motor vehicle crashes, children younger than 16 years were identied, who received acute care and were hospitalized at the University of Chicago Medical Center, a Level I pediatric trauma center, after being struck by a motor vehicle from 2002 to 2009. By retrospective chart review and review of the Emergency Medical Services run sheets, demographic data and details of the crash site were collected. Crash sites were aggregated on a block by block basis. A hot spot analysis was performed to localize clusters of injury events. Using Gi* statistical method, spatial clusters were identied at different condence intervals using a xed distance band of 400 m ( 14 mile). Maps were generated using GIS with 2000 census data to evaluate race, employment, income, density of public and private schools, and density of children living in the neighborhoods surrounding our medical center where crash sites were identied. Spatial correlation is used to identify statistically signicant locations. Results: There were 3,521 children admitted to the University of Chicago Medical Center for traumatic injuries from 2002 to 2009; 27.7% (974) of these children sustained injuries in pedestrian motor vehicle injuries. From 2002 to 2009, there were a total of 106 traumatic deaths, of which 29 (27.4%) were due to pedestrian motor vehicle crashes. Pediatric pedestrian motor vehicle crash sites occurred predominantly within low-income, predominantly African-American neighborhoods. A lower prevalence of crash sites was observed in the predominantly higher income, nonAfrican-American neighborhoods. Conclusions: Spatial analysis using GIS identied associations between pediatric pedestrian motor vehicle crash sites and the neighborhoods served by our pediatric trauma center. Pediatric pedestrian motor vehicle crash sites occurred predominantly within low-income, African-American neighborhoods. The disparity in prevalence of crash sites is somewhat attributable to the lower density of children living in the predominantly higher income, non-African-American neighborhoods, including the community immedi-

ately around our hospital. Trafc volume patterns, as a denominator of these injury events, remain to be studied. Key Words: Pediatric pedestrian motor vehicle crashes, GIS, Injury prevention. (J Trauma. 2011;71: S511S516)

Submitted for publication December 1, 2010. Accepted for publication September 28, 2011. Copyright 2011 by Lippincott Williams & Wilkins From the Section of Pediatric Surgery, University of Chicago, Chicago, Illinois. Presented at the Injury Free Coalition for Kids (IFCK) Conference, November 1214, 2010, Chicago, Illinois. Address for reprints: Mindy Statter, MD, Division of Pediatric Surgery, Childrens Hospital at Monteore, 111 E. 210th Street, Bronx, NY 10467; email: mstatter@monteore.org. DOI: 10.1097/TA.0b013e31823a4b70

ediatric pedestrian injuries remain a major cause of childhood death, hospitalization, and disability. To target injury prevention efforts, it is imperative to identify those children at risk. Racial disparities have been noted in the rates of pediatric pedestrian injury and death. Compared with adults, 60% to 70% of children injured and killed as pedestrians are male. Black and American Indian/Alaska Native children have higher rates of pedestrian injury and death.1 Walkablility, exposure to trafc, and safety may differ depending upon the socioeconomic status or ethnic composition of the community. The potential benets of walking to school may be undermined by dangers due to trafc, crime, and poor maintenance of sidewalks.2 As previously reported in other cities, we hypothesized that children from lowincome families living in dense, urban residential neighborhoods have a higher risk of sustaining pedestrian injury.3 6 Descriptive epidemiology involves the examination of person, place, and time in the occurrence of disease or injury. Geomatics is the science of acquiring and integrating spatially referenced data for mapping and analysis. Geomatics includes geographic information systems (GIS), a tool used for spatial analysis. GIS methods link spatial features with descriptive attribute data (e.g., social, economic, and demographic variables) at various geographic scales. GIS is a useful tool in injury prevention in its ability to map and identify specic patterns and spatial clusters. Geocoding locations of injury occurrence translates the addresses of injury events (pedestrian motor vehicle crashes) into geographic coordinates which can then be correlated with other spatial data layers such as census data.7 A method used to identify concentrations of high or low values within a distance, clusters of injury events, or hot spots is the Gi*. Gi* uses a spatial neighborhood based on a set distance from an event. A group of features with high Gi* values indicates a cluster or concentration of features with high attribute values. A more recent version of Gi*, developed by Keith Ord and Art Getis in 1995, combines the original Gi* statistic and z score in a single measure. The reported statistic is the z score.
S511

The Journal of TRAUMA Injury, Infection, and Critical Care Volume 71, Number 5, November Supplement 2, 2011

Statter et al.

The Journal of TRAUMA Injury, Infection, and Critical Care Volume 71, Number 5, November Supplement 2, 2011

The interpretation remains the same, high Gi* values reect statistically signicant cluster of high values.8

MATERIALS AND METHODS


Using E-codes for pedestrian motor vehicle crashes, children younger than 16 years were identied from the trauma registry, who received acute care and were hospitalized at the University of Chicago Medical Center, a Level I pediatric trauma center, after being struck by a motor vehicle from 2002 to 2009. By retrospective chart review and review of the Emergency Medical Services run sheets, demographic data and details of the crash site were respectively collected.

Crash sites were aggregated on a block by block basis. A hot spot analysis was performed to localize clusters of injury events. Using Gi* statistical method, spatial clusters were identied at different condence intervals using a xed distance band of 400 m ( 14 mile). This allowed the separation of spatial outliers and the further elucidation as to why crashes were happening in only certain locations. In addition, maps were generated using GIS with 2000 census data to evaluate race, income, and density of children living in the neighborhoods surrounding our medical center where crash sites were identied. Spatial correlation is used to identify statistically signicant locations. Our institution is the only

Figure 1. All crashes 20022009. S512 2011 Lippincott Williams & Wilkins

The Journal of TRAUMA Injury, Infection, and Critical Care Volume 71, Number 5, November Supplement 2, 2011 Targeting Pediatric Pedestrian Injury Prevention Efforts

Level I pediatric trauma center on the south side of Chicago. Very few signicant pediatric traumas that occur on the south side of Chicago are treated at other institutions and would therefore be missed in our trauma registry.

RESULTS
There were 3,521 children admitted to the University of Chicago Medical Center for traumatic injuries from 2002 to 2009; 27.7% (974) of these children sustained injuries in pedestrian motor vehicle injuries. During this study period, there were a total of 106 traumatic deaths, of which 29 (27.4%) were due to pedestrian motor vehicle crashes.

The geographic distribution of all pediatric pedestrian motor vehicle crashes treated at the University of Chicago Pediatric Trauma Center during the study period from 2002 to 2009 is mapped in Figure 1. Clusters of injury events are localized to the neighborhoods served by the medical center. Fewer injury events were noted in the Hyde Park neighborhood where the medical center is located. The relative density of African-American families residing in the surrounding neighborhoods and the localization of injury events in these neighborhoods is shown in Figure 2. Figure 3 illustrates the average income of residents in the neighborhoods served by the medical center relative to the clusters of injury events.

Figure 2. Density of African-American families. 2011 Lippincott Williams & Wilkins S513

Statter et al.

The Journal of TRAUMA Injury, Infection, and Critical Care Volume 71, Number 5, November Supplement 2, 2011

Figure 3. Average income.

The density of children residing in the neighborhoods surrounding the University of Chicago relative to the distribution of clusters of injury events in these neighborhoods is shown in Figure 4. The prevalence of crash sites is less in the vicinity of private schools relative to public schools (maps not shown).

DISCUSSION
This analysis provides a rich picture of the social determinants of child pedestrian safety incorporating how these factors interact with pedestrian trauma risk spatially in a section of a large US city. This analysis and, in particular,
S514

the maps generated will be useful to planners, advocates, trafc safety practitioners, and others in understanding and preventing child pedestrian trauma. In a preliminary study published in 2006, data from the trauma registry at the University of Chicago, a Level I pediatric trauma center, were used to map out sites where clusters of pediatric pedestrian motor vehicle events, hot spots, had occurred over 3 consecutive years (20022004). GIS analysis showed a signicant relationship between community area and frequency of pediatric pedestrian injury. The relatively afuent university community immediately adjacent to our hospital center had no child pedestrian crashes that
2011 Lippincott Williams & Wilkins

The Journal of TRAUMA Injury, Infection, and Critical Care Volume 71, Number 5, November Supplement 2, 2011 Targeting Pediatric Pedestrian Injury Prevention Efforts

Figure 4. Child density.

resulted in injuries. Cluster analysis of an injury site in the vicinity of an elementary school showed that the majority of injury events occurred within 1 to 2 blocks of the school and that victims were the age of school attendees. These data allow for informed efforts toward prevention. We found previously that each year clusters of pediatric pedestrian motor vehicle crashes involving children younger than 16 years occurred at the same locations and the majority of injury events occurred between 4 and 8 PM.9 In our earlier work, we found that there was a signicant clustering of child pedestrian crashes that occurred immediately adjacent to a nearby Chicago Public School serving kindergarten through
2011 Lippincott Williams & Wilkins

eighth grades. According to the school principal, more than 95% of the 858 students enrolled at this elementary school walk to school. Data from this study conrm that walking to school presents a considerable risk to the health of children in this community. These low-income, minority children are referred to as captive walkers in the transportation literature because they have no alternative means of transportation.10 This targeted injury prevention effort has resulted in the designation of this elementary school as a site to receive federal funding through the Safe Routes to School Program for environmental modication. Signicant federal funding has recently been allocated to Safe Routes to School to help
S515

Statter et al.

The Journal of TRAUMA Injury, Infection, and Critical Care Volume 71, Number 5, November Supplement 2, 2011

states develop programs and infrastructure to encourage children to walk to school in a safe environment. This effort has incorporated the school, community members, aldermen, and the Chicago Department of Transportation. Environmental modication is only one component of pedestrian injury prevention. Geomatics applications, through their spatial analysis capabilities, allow for visualization of the distribution of spatial clusters. Although GIS can show associations of community-based data with injury data, this methodology cannot infer causality. Demographic data are useful to develop risk proles. The relationship of population health determinants, e.g., income, level of education, employment, and race to injury patterns continues to be studied. Each factor alone affects injury occurrence in general; however, the interaction of these factors may differ in specic geographic locations (communities).11 Spatial analysis has the potential to determine geographical distribution, map populations at risk, stratify risk factors, plan and target interventions, and monitor interventions over time. Spatial analysis using GIS identied associations between pediatric pedestrian motor vehicle crash sites and the neighborhoods served by our pediatric trauma center. Pediatric pedestrian motor vehicle crash sites occurred predominantly within low-income, African-American neighborhoods. A lower prevalence of crash sites was observed in the predominantly higher income, nonAfrican-American neighborhoods including the community immediately around our hospital. The disparity in prevalence of crash sites is somewhat attributable to the lower density of children living in the latter referenced neighborhoods. The prevalence of crash sites is less in the vicinity of private schools compared with public schools. Future studies will evaluate those factors contributing to fewer pediatric pedestrian injury events around private schools such as the presence of crossing guards and after-school programs that decrease the childs trafc exposure at peak injury times. A limitation of this study is the availability of census data from 2000. 2010 census data (to be released in the summer of

2011) will provide more accurate community information, in light of the recent impact of the economy, on employment, housing, and family density. Trafc volume patterns, as a denominator of these injury events, remain to be studied. Crash sites occur where urban children have a high level of exposure to trafc, often related to routine play. The relationship of crash sites to public parks will allow for targeted park and playground builds. Templates for injury prevention strategies need to be community based to target those children at risk. REFERENCES
1. Committee on Injury, Violence, and Poison Prevention. Policy statement pedestrian Safety. Pediatrics. 2009;124:802 812. 2. Zhu X, Lee C. Walkability and safety around elementary schools. Am J Prev Med. 2008;34:282290. 3. Rao R, Hawkins M, Guyer B. Childrens exposure to trafc and risk of pedestrian injury in an urban setting. Bull NY Acad Med. 1997; 74:65 80. 4. Posner JC, Liao E, Winston FK, Cnaan A, Shaw KN, Durbin DR. Exposure to trafc among urban children injured as pedestrians. Inj Prev. 2002;8:231235. 5. DiMaggio C, Durkin M. Child pedestrian injury in an urban setting. Acad Emerg Med. 2002;9:54 62. 6. Durkin M, Laraque D, Lubman I, Barlow B. Epidemiology and prevention of trafc injuries to urban children and adolescents. Pediatrics. 1999;103:e74. 7. Cusimano MD, Chipman M, Glazier RH, Rinner C, Marshall SP. Geomatics in injury prevention: the science, the potential and the limitations. Inj Prev. 2007;13:5156. 8. Mitchell A. Identifying clusters. In: The ESRI Guide to GIS Analysis, Vol. 2: Spatial Measurements and Statistics. Redlands: Esri Press; 2005. 9. Statter M, Strickland J, Quinlan K, Harris-Rosado M, Glynn L, Liu D. The identication of environmental factors in pediatric pedestrian motor vehicle crashes. J Trauma. 2006;60:1384 1385. 10. ChuX. The relative risk between walking and motoring in the United States. In: Proceedings of the 85th Annual Meeting of the Transportation Research Board, January 2125; Washington, DC, 2006. 11. Schieber RA, Vegega ME, eds. Reducing childhood pedestrian injuries: summary of a multidisciplinary conference. Inj Prev. 2002; 8(Suppl 1):19.

S516

2011 Lippincott Williams & Wilkins