Ferguson Alesia C, Canales Robert A, Beamer Paloma, Auyeung Willa, Key Maya, Munninghoff Amy, Lee Kevin Tse-Wing, Robertson Alexander, Leckie James O
College of Public Health, University of Arkansas Medical Sciences, Little Rock, 72205-7199, USA.
J Expo Sci Environ Epidemiol. 2006 May;16(3):287-98. doi: 10.1038/sj.jea.7500459.
In 1994, Stanford University's Exposure Research Group (ERG) conducted its first pilot study to collect micro-level activity time series (MLATS) data for young children. The pilot study involved videotaping four children of farm workers in the Salinas Valley of California and converting their videotaped activities to valuable text files of contact behavior using video-translation techniques. These MLATS are especially useful for describing intermittent dermal (i.e., second-by-second account of surfaces and objects contacted) and non-dietary ingestion (second-by-second account of objects or hands placed in the mouth) contact behavior. Second-by-second records of children contact behavior are amenable to quantitative and statistical analysis and allow for more accurate model estimates of human exposure and dose to environmental contaminants. Activity patterns data for modeling inhalation exposure (i.e., accounts of microenvironments visited) can also be extracted from the MLATS data. Since the pilot study, ERG has collected an immense MLATS data set for 92 children using more developed and refined videotaping and video-translation methodologies. This paper describes all aspects required for the collection of MLATS including: subject recruitment techniques, videotaping and video-translation processes, and potential data analysis. This paper also describes the quality assurance steps employed for these new MLATS projects, including: training, data management, and the application of interobserver and intraobserver agreement during video translation. The discussion of these issues and ERG's experiences in dealing with them can assist other groups in the conduct of research that employs these more quantitative techniques.
1994年，斯坦福大学暴露研究小组（ERG）开展了首次试点研究，以收集幼儿的微观层面活动时间序列（MLATS）数据。该试点研究包括对加利福尼亚州萨利纳斯山谷农场工人的4名子女进行录像，并使用视频翻译技术将他们的录像活动转换为有价值的接触行为文本文件。这些MLATS对于描述间歇性皮肤接触（即每秒记录接触的表面和物体）和非饮食摄入（每秒记录放入口中的物体或手）接触行为特别有用。儿童接触行为的逐秒记录适合进行定量和统计分析，并能更准确地估计人类接触环境污染物的暴露量和剂量。用于模拟吸入暴露的活动模式数据（即访问的微环境记录）也可以从MLATS数据中提取。自试点研究以来，ERG使用更先进和完善的录像及视频翻译方法，为92名儿童收集了大量的MLATS数据集。本文描述了收集MLATS所需的各个方面，包括：受试者招募技术、录像和视频翻译过程以及潜在的数据分析。本文还描述了这些新的MLATS项目所采用的质量保证步骤，包括：培训、数据管理以及视频翻译过程中观察者间和观察者内一致性的应用。对这些问题的讨论以及ERG在处理这些问题方面的经验可以帮助其他研究小组开展采用这些更定量技术的研究。
