Kumagai S, Matsunaga I, Sugimoto K, Kusaka Y, Shirakawa T
Department of Occupational Health, Osaka Prefectural Institute of Public Health.
Sangyo Igaku. 1989 Jul;31(4):216-26. doi: 10.1539/joh1959.31.216.
A method based on interday fluctuation of contaminant concentrations for evaluating employee's exposure averages (8-h TWAs) was proposed in our previous report. The method was established on the assumption that daily exposure averages of the workers are lognormally distributed in actual workplaces. The study was conducted to elucidate whether the distribution of daily exposure averages is statistically lognormal or not and to examine the relationship between sample geometric standard deviation (sg) of worker's daily exposure averages and its estimate (sg2) calculated by measurements for two consecutive days. These are critical for our proposed method. The data on daily exposure concentrations over five to six weeks were collected from workers exposed to cobalt, acetone, n-hexane, toluene, xylene, ethylbenzene and ethylacetate. The data on organic lead, inorganic lead and mercury exposures reported by Cope et al. and Lindstedt et al. were also used for the study. The result can be summarized as follows: 1. Lognormal distribution of daily exposure averages was confirmed by plotting on normal probability paper and chi 2-test. 2. Median of sg2 on daily exposure averages obtained from individual worker was smaller than sg. 3. sg calculated by a set of measurements for two consecutive days in every worker can be corrected by the equation: (sg2) 1.48 in obtaining a better estimate of sigma g. 4. Statistical analysis on daily exposure averages of all workers showed that median of sg2 was smaller than that of sg, and 88% upper limit of sg2 was equal to that of sg. 5. Therefore, in evaluating TWA obtained by only single day's measurement using the proposed method, median of sg representative of industrial hazardous substance exposure workplaces could be also corrected by the equation described above. However, correction of 90% upper limit of sg2 is not necessary.
我们之前的报告中提出了一种基于污染物浓度日间波动来评估员工暴露平均值（8小时时间加权平均值）的方法。该方法基于这样的假设：在实际工作场所中，工人的每日暴露平均值呈对数正态分布。本研究旨在阐明每日暴露平均值的分布在统计学上是否呈对数正态，并检验工人每日暴露平均值的样本几何标准差（sg）与其通过连续两天测量计算得到的估计值（sg2）之间的关系。这些对于我们提出的方法至关重要。收集了五到六周内接触钴、丙酮、正己烷、甲苯、二甲苯、乙苯和乙酸乙酯的工人的每日暴露浓度数据。还使用了科普等人和林德施泰特等人报告的有机铅、无机铅和汞暴露数据。结果总结如下：1. 通过在正态概率纸上绘图和卡方检验，证实了每日暴露平均值的对数正态分布。2. 从个体工人获得的每日暴露平均值的sg2中位数小于sg。3. 每个工人连续两天的一组测量计算得到的sg可以通过以下公式校正：(sg2)×1.48，以更好地估计西格玛g。4. 对所有工人的每日暴露平均值进行统计分析表明，sg2的中位数小于sg的中位数，且sg2的88%上限与sg的上限相等。5. 因此，在使用所提出的方法评估仅通过单日测量获得的时间加权平均值时，代表工业有害物质暴露工作场所的sg中位数也可以通过上述公式校正。然而，sg2的90%上限无需校正。
