Handal Alexis J, McGough-Maduena Alison, Páez Maritza, Skipper Betty, Rowland Andrew S, Fenske Richard A, Harlow Siobán D
a Public Health Program, Department of Family and Community Medicine, University of New Mexico , Albuquerque , New Mexico , USA.
Arch Environ Occup Health. 2015;70(4):232-40. doi: 10.1080/19338244.2013.879563.
Self-reported measures of residential pesticide exposure are commonly used in epidemiological studies, especially when financial and logistical resources are limited. However, self-reporting is prone to misclassification bias. This pilot study assesses the agreement between self-report of residential pesticide exposure with direct observation measures, in an agricultural region of Ecuador, as a cross-validation method in 26 participants (16 rose workers and 10 controls), with percent agreement and kappa statistics calculated. Proximity of homes to nearby flower farms was found to have only fair agreement (kappa =.35). The use of discarded plastics (kappa =.06) and wood (kappa =.13) were found to have little agreement. Results indicate that direct observation or measurement may provide more accurate appraisals of residential exposures, such as proximity to industrial farmland and the use of discarded materials obtained from the flower farms.
在流行病学研究中，尤其是在资金和后勤资源有限时，通常会采用自我报告的方式来衡量居住环境中的农药暴露情况。然而，自我报告容易出现错误分类偏差。这项试点研究评估了在厄瓜多尔一个农业地区，居住农药暴露自我报告与直接观察测量之间的一致性，以此作为一种交叉验证方法，对26名参与者（16名玫瑰种植工人和10名对照者）进行了研究，并计算了一致百分比和kappa统计量。结果发现，住宅与附近花卉农场的距离仅有中等程度的一致性（kappa = 0.35）。使用废弃塑料（kappa = 0.06）和木材（kappa = 0.13）的一致性较低。结果表明，直接观察或测量可能会对居住暴露情况提供更准确的评估，例如与工业农田的距离以及使用从花卉农场获得的废弃材料的情况。
