Validity and reliability of a method for retrospective evaluation of chlorophenate exposure in the lumber industry.

This paper describes the validity and reliability of a method to retrospectively assess exposure to antisapstain agents used in sawmills (chlorophenates). The method is based on experienced workers' estimates of exposure for each job title at the sawmill where they work. At a pilot mill, 10 randomly selected workers estimated the frequency and duration of exposures to chlorophenates for all 59 job titles. The reliability of their mean exposure estimates was very high, with an intraclass correlation coefficient for all raters of 0.91 (based on a calculated index of exposure). To assess validity, urinary chlorophenate levels were measured for 86% of the workers at the mill during the summer and/or fall, and compared to experienced workers' estimates of exposure. The correlation between workers' exposure estimates and the urinary chlorophenate levels for each job title were consistently above 0.65 for all analyses and greater than 0.72 when summer and fall urine sample results were averaged. The evidence indicates that the validity and reliability of worker exposure estimates are high enough to justify investigation of the method's generalizability to other sawmills.