Abstract

This study investigated sources of variance in exposure to risk factors for knee pain in a variety of highway construction trades, operations, and tasks. Over 15,000 discrete observations of leg postures and weights handled were made on 120 construction workers in five construction trades, in nine operations over 79 days. The contributions of trade, operation, task, and worker to the variability in work time spent kneeling, squatting, and carrying loads were evaluated with multilevel random effects models. Construction operation and task explained about 20% to 30% of total variation in kneeling, squatting, and carrying loads. There was a large unexplained component of variance thought to represent day-to-day variability of exposure within task. Reliable assessments of knee exposures require multiple days to accommodate the high variability of exposures among operations and tasks and over time. These sources of variability should be carefully considered in efforts to estimate exposures to knee loading for epidemiologic or intervention studies. Homogenous exposure groups are not easily defined from the readily available organizational features of construction work.