Lasecki C H, Mujica F C, Stutsman S, Williams A Y, Ding L, Simmons J D, Brevard S B
From the Departments of Surgery and Geography, University of South Alabama, Mobile, AL.
J Trauma Acute Care Surg. 2018 Jan;84(1):70-74. doi: 10.1097/TA.0000000000001720.
Geographic information systems (GIS) have proven effective in studying intentional injury in various communities; however, GIS is not implemented widely for use by Level I trauma centers in understanding patient populations. Our study of intentional injury combines the capabilities of GIS with a Level I trauma center registry to determine the spatial distribution of victims and correlated socioeconomic factors.
One thousand ninety-nine of 3,109 total incidents of intentional trauma in the trauma registry from 2005 to 2015 had sufficient street address information to be mapped in GIS. Comparison of these data, coupled with demographic data at the block group level, determined if any clustering or spatial patterns existed. Geographic information systems delivered these comparisons using several spatial statistics including kernel density, ordinary least squares test, and Moran's index.
Kernel density analysis identified four major areas with significant clustering of incidents. The Moran's I value was 0.0318. Clustering exhibited a positive z-score and significant p value (p < 0.01). Examination of socioeconomic factors by spatial correlation with the distribution of intentional injury incidents identified three significant factors: unemployment, single-parent households, and lack of a high school degree. Tested factors did not exhibit substantial redundancy (variance inflation factor < 7.5). Nonsignificant tested factors included race, proximity to liquor stores and bars, median household income, per capita income, rate with public assistance, and population density.
Spatial representation of trauma registry data using GIS effectively identifies high-risk areas for intentional injury. Analysis of local socioeconomic data identifies factors unique to those high-risk areas in the observed community. Implications of this study may include the routine use of GIS by Level I trauma centers in assessing intentional injury in a given community, the use of that data to guide the development of trauma prevention, and the assessment of other mechanisms of trauma using GIS.
Epidemiological, level IV.
地理信息系统（GIS）已被证明在研究不同社区的故意伤害方面是有效的；然而，一级创伤中心在了解患者群体时并未广泛应用GIS。我们对故意伤害的研究将GIS的功能与一级创伤中心登记处相结合，以确定受害者的空间分布以及相关的社会经济因素。
2005年至2015年创伤登记处3109起故意伤害事件中的1099起有足够的街道地址信息，可在GIS中进行映射。将这些数据与街区组层面的人口统计数据进行比较，以确定是否存在任何聚集或空间模式。地理信息系统使用包括核密度、普通最小二乘法检验和莫兰指数在内的几种空间统计方法进行这些比较。
核密度分析确定了四个事件显著聚集的主要区域。莫兰指数I值为0.0318。聚集呈现正的z分数和显著的p值（p < 0.01）。通过与故意伤害事件分布的空间相关性对社会经济因素进行检查，确定了三个显著因素：失业、单亲家庭和没有高中学历。测试的因素没有表现出实质性的冗余（方差膨胀因子< 7.5）。不显著的测试因素包括种族、与酒类商店和酒吧的距离、家庭收入中位数、人均收入、公共援助率和人口密度。
使用GIS对创伤登记数据进行空间表示可有效识别故意伤害的高风险区域。对当地社会经济数据的分析确定了观察到的社区中那些高风险区域特有的因素。本研究的意义可能包括一级创伤中心在评估特定社区的故意伤害时常规使用GIS，利用该数据指导创伤预防的发展，以及使用GIS评估其他创伤机制。
流行病学，四级。
