Marotta Phillip L, Leach Benjamin Cb, Hutson William D, Caplan Joel M, Lohmann Brenna, Hughes Charlin, Banks Devin, Roll Stephen, Chun Yung, Jabbari Jason, Ancona Rachel, Mueller Kristen, Cooper Ben, Anasti Theresa, Dell Nathaniel, Winograd Rachel, Heimer Robert
Washington University in St. Louis, MO, USA; Brown School, Washington University in St. Louis, St. Louis, MO, USA; Social Policy Institute, Washington University in St. Louis, MO, USA.
Washington University in St. Louis, MO, USA; Brown School, Washington University in St. Louis, St. Louis, MO, USA; University of California San Francisco, Department of Medicine, Division of Health Equity and Society, San Francisco, California, United States.
Int J Drug Policy. 2024 Dec;134:104611. doi: 10.1016/j.drugpo.2024.104611. Epub 2024 Nov 2.
The objective of this study was to identify place features associated with increased risk of drug-involved fatalities and generate a composite score measuring risk based on the combined effects of features of the built environment.
We conducted a geospatial analysis of overdose data from 2022 to 2023 provided by the St. Louis County Medical Examiner's Office to test whether drug-involved deaths were more likely to occur near 54 different place features using Risk Terrain Modeling (RTM). RTM was used to identify features of the built environment that create settings of heightened overdose risk. Risk was estimated using Relative Risk Values (RRVs) and a composite score measuring Relative Risk Scores (RRS) across the county was produced for drugs, opioids, and stimulants, as well as by Black and White decedents.
In the model including all drugs, deaths were more likely to occur in close proximity to hotels/motels (RRV=39.65, SE=0.34, t-value=10.81 p<.001), foreclosures (RRV=4.42, SE=0.12, t-value = 12.80, p<.001), police departments (RRV=3.13, SE=0.24, t-score=4.86, p<.001), and restaurants (RRV=2.33, SE=0.12, t-value=7.16, p<.001). For Black decedents, deaths were more likely to occur near foreclosures (RRV=9.01, SE=0.18, t-value =11.92, p<.001), and places of worship (RRV= 2.51, SE=0.18, t-value = 11.92, p<.001). For White decedents, deaths were more likely to occur in close proximity to hotels/motels (RRV=38.97, SE=0.39, t-value=9.30, p<.001) foreclosures (RRV=2.57, SE=0.16, t-value =5.84, p<.001), restaurants (RRV=2.52, SE=0.17, t-value=5.33, p<.001) and, auto painting/repair shops (RRV=0.04, SE=0.18, t-value =3.39, p<.001).
These findings suggest that places of worship, the hospitality industry, and housing authorities may be physical features of the environment that reflect social conditions that are conducive to overdose. The scaling up of harm reduction strategies could be enhanced by targeting places where features are co-located.
本研究的目的是确定与药物相关死亡风险增加相关的场所特征，并基于建筑环境特征的综合影响生成一个衡量风险的综合评分。
我们对圣路易斯县法医办公室提供的2022年至2023年过量用药数据进行了地理空间分析，以使用风险地形建模（RTM）测试药物相关死亡是否更有可能发生在54种不同的场所特征附近。RTM用于识别建筑环境中导致过量用药风险升高的特征。使用相对风险值（RRV）估计风险，并针对药物、阿片类药物和兴奋剂以及黑人和白人死者生成全县范围内衡量相对风险评分（RRS）的综合评分。
在包括所有药物的模型中，死亡更有可能发生在酒店/汽车旅馆附近（RRV = 39.65，SE = 0.34，t值 = 10.81，p <.001）、止赎房屋附近（RRV = 4.42，SE = 0.12，t值 = 12.80，p <.001）、警察局附近（RRV = 3.13，SE = 0.24，t分数 = 4.86，p <.001）和餐馆附近（RRV = 2.33，SE = 0.12，t值 = 7.16，p <.001）。对于黑人死者，死亡更有可能发生在止赎房屋附近（RRV = 9.01，SE = 0.18，t值 = 11.92，p <.001）和宗教场所附近（RRV = 2.51，SE = 0.18，t值 = 11.92，p <.001）。对于白人死者，死亡更有可能发生在酒店/汽车旅馆附近（RRV = 38.97，SE = 0.39，t值 = 9.30，p <.001）、止赎房屋附近（RRV = 2.57，SE = 0.16，t值 = 5.84，p <.001）、餐馆附近（RRV = 2.52，SE = 0.17，t值 = 5.33，p <.001）以及汽车喷漆/修理店附近（RRV = 0.04，SE = 0.18，t值 = 3.39，p <.001）。
这些发现表明，宗教场所、酒店业和住房管理部门可能是环境的物理特征，反映了有利于过量用药的社会状况。通过针对特征共存的场所扩大减少伤害策略，可以加强这些策略。
