Optimising sampling strategies: components of low-back EMG variability in five heavy industries

Direct/ measurement of work activities is costly, so researchers need to distribute resources efficiently to elucidate the relationships between exposures and back injury.

This study used data from full-shift electromyography (EMG; N=133) to develop three exposure metrics: mean, 90th percentile and cumulative EMG. For each metric, the components of variance were calculated between- and within-subject, and between-group for four different grouping schemes: grouping by industry (construction, forestry, transportation, warehousing and wood products), by company, by job and by quintiles based on exposures ranked by jobs within industries. Attenuation and precision of simulated exposure-response relationships were calculated for each grouping scheme to determine efficient sampling strategies.

As expected, grouping based on exposure quintiles had the highest between-group variances and lowest attenuation, demonstrating the lowest possible attenuation with this data.

There is potential for grouping schemes to reduce attenuation, but precision losses should be considered and whenever possible empirical data should be employed to select potential exposure grouping schemes.