[Bibliometrics and visualization analysis of land use regression models in ambient air pollution research].

Objective: To quantitatively analyze the current status and development trends regarding the land use regression (LUR) models on ambient air pollution studies. Methods: Relevant literature from the PubMed database before June 30, 2017 was analyzed, using the Bibliographic Items Co-occurrence Matrix Builder (BICOMB 2.0). Keywords co-occurrence networks, cluster mapping and timeline mapping were generated, using the CiteSpace 5.1.R5 software. Relevant literature identified in three Chinese databases was also reviewed. Results: Four hundred sixty four relevant papers were retrieved from the PubMed database. The number of papers published showed an annual increase, in line with the growing trend of the index. Most papers were published in the journal of Environmental Health Perspectives. Results from the Co-word cluster analysis identified five clusters: cluster#0 consisted of birth cohort studies related to the health effects of prenatal exposure to air pollution; cluster#1 referred to land use regression modeling and exposure assessment; cluster#2 was related to the epidemiology on traffic exposure; cluster#3 dealt with the exposure to ultrafine particles and related health effects; cluster#4 described the exposure to black carbon and related health effects. Data from Timeline mapping indicated that cluster#0 and#1 were the main research areas while cluster#3 and#4 were the up-coming hot areas of research. Ninety four relevant papers were retrieved from the Chinese databases with most of them related to studies on modeling. Conclusion: In order to better assess the health-related risks of ambient air pollution, and to best inform preventative public health intervention policies, application of LUR models to environmental epidemiology studies in China should be encouraged.目的： 定量分析基于土地利用回归（LUR）模型的大气污染研究现状和发展趋势。 方法： 检索PubMed数据库2017年6月30日前收录的相关文献，采用书目共现分析系统（BICOMB 2.0）统计文献计量学指标，运用引文空间信息可视化分析软件（CiteSpace 5.1.R5）生成关键词共现网络、聚类图谱和时间线图谱，并分析3个中文数据库的相关文献。 结果： PubMed数据库检索到464篇文献，年份发文量符合指数增长趋势。发文最多的期刊是Environmental Health Perspectives。共词聚类分析得到5个聚类（根据CiteSpace 5.1.R5软件自动生成序号）：#0基于出生队列的产前空气污染暴露及健康效应；#1 LUR模型构建及暴露评估；#2交通暴露的流行病学；#3超细颗粒物暴露及健康效应；#4黑碳暴露及健康效应。时间线图谱显示聚类#0和#1是主要研究领域，聚类#3和#4是前沿研究领域。中文数据库共检索到94篇文献，主要为建模研究。 结论： 我国应开展更多基于LUR模型的环境流行病学研究，精细评估大气污染的健康风险。.