Rosenberg C R, Mulvihill M N, Fischbein A, Blum S
Institute of Environmental Medicine, New York University Medical Center, NY 10016.
Br J Ind Med. 1987 Oct;44(10):702-10. doi: 10.1136/oem.44.10.702.
An investigation was conducted to examine the validity of self reported work histories obtained from a sample of 326 capacitor manufacturing workers who had participated in an epidemiological study relating health abnormalities to exposure to polychlorinated biphenyls. Based on objective personnel records obtained for 288 members of the sample group, validity scores ranged from 20% to 100% with most falling between 70% and 75%. This validity range corresponded to misclassification proportions ranging from 13% to 29%. Exposure misclassification tended to be random for men but non-random for women (overestimation of exposure). By means of multiple regression analysis, the diversity of the job categorical pattern was found to be a significant independent predictor of validity. The factors sex, duration, and time lapse were also significant validity predictors but only interactively with job diversity. There were also indications that interviewer skill could be an important factor. Knowledge of the magnitude of these factor effects could be critical in planning retrospective epidemiological studies.
对326名参与了一项将健康异常与多氯联苯暴露相关联的流行病学研究的电容器制造工人样本所提供的自我报告工作经历的有效性进行了调查。基于从样本组的288名成员处获得的客观人事记录，有效性得分在20%至100%之间，大多数落在70%至75%之间。这个有效性范围对应的错误分类比例在13%至29%之间。男性的暴露错误分类倾向于随机的，但女性的则是非随机的（暴露估计过高）。通过多元回归分析，发现工作类别模式的多样性是有效性的一个显著独立预测因素。性别、工作时长和时间间隔等因素也是有效性的显著预测因素，但仅与工作多样性存在交互作用。也有迹象表明访谈者的技能可能是一个重要因素。了解这些因素影响的程度对于规划回顾性流行病学研究可能至关重要。
