On the Concept of the “Normal” Day: Quality Control of Occupational Hygiene Measurements

Abstract

Measurement data must be without significant bias or the size of the bias must be known, allowing for corrections. Occupational hygiene measurements are often performed on a single day or a few days chosen for various reasons, including purely practical ones. When quantitative exposure data from the same exposure period are lacking, the representativeness of the measurement days (MDs) must be assessed based on nonquantitative exposure data. The ability of workers, managers, and occupational hygienists to assess concentration levels, without quantitative exposure data, has been questioned in recent literature and the question can be raised about how it is possible to secure that MDs are “normal.” This article describes a quantitative method for verifying, if the data collected are representative for a longer period than a single day or a few days. The method consists of a categorical description of the working days by workers keeping logs about when, and for how long a time, they perform a set of predefined processes. By assigning a concentration level to each process, the time-weighted average concentration (TWAC) for the shift can be calculated for each day, usually based on an 8-hour shift. The underlying assumption of all occupational hygiene measurements that measurement days are normal can be evaluated by comparing the TWACs calculated for the measurement day(s) with the calculated TWACs for a log period of sufficient length. It is of note that this article is not about how to estimate exposure, but about how to test whether or not the MDs are representative for a period of time longer than a single day or a few days. The proposed method presumes that the full-shift exposure level is independently estimated (e.g., by using passive samplers). The test of representativeness is illustrated by styrene exposure data from two fiberglass-reinforced polyester plants. Test of representativeness adds to the costs of measurements. These additional costs, however, should be compared with the costs of either not detecting a potentially harmful exposure or investing money in control measures that might have been used more effectively elsewhere.