Student Research: Carolyn Salazar
MS, Safety and Ergonomics (Saf Erg), 2003
Faculty Advisor: Peter W. Johnson
Evaluation of Reliability and Validity of the Hand Activity Level (HAL) and the Strain Index for Use in Epidemiological Studies
The purpose of the current study, which was a sub-study of a large prospective study, was to assess the utility of the ACGIH’s Hand Activity Level) and the Strain Index (SI) for use in epidemiological studies for assessing hand activity. Three experienced ergonomists and one novice observer, working in pairs, formed six groups and evaluated the hand activity of 122 subjects working in industries that included an exercise equipment manufacturer, a heating and air-conditioning manufacturer and a sawmill. Measurement methods include 1) observations by using the HAL and Strain Index measures, 2) the worker’s self-report of hand activity using the BORG scale, and 3) direct measurements of the hand forces via dynamometer and pinch gauge measurements. The weighted kappa and the Spearman Rho were used to assess inter-rater reliability, compare the objective measurement and subjective estimates of hand forces, and determine whether there were differences between the novice’s and experienced ergonomist’s observations. Validity was assessed by 1) rater accuracy, 2) correlations between the dynamometer and pinch gauge measurements and the observers’ and workers’ estimates of these forces, and 3) comparing the observations of the novice and the experienced ergonomists.
The categories of HAL and speed of work from the SI yielded the highest inter-rater reliability achieving kappa values of 0.34-0.45, thus these categories may have the greatest utility in epidemiological studies when assessing hand exposures. Duration of effort and efforts per minute had the lowest rater accuracy and the highest rater variability, indicating the difficulty in measuring these categories. The rater pairs that had the novice rater had the lowest inter-rater reliability (average weighted kappa = 0.24) compared to the pairs comprised of experienced ergonomists (average weighed kappa = 0.36), indicating the novice rater tended to rate differently than the more experienced ergonomists. When examining rater accuracy, the novice rater tended to rate exposures higher compared to the other raters and the most experienced ergonomist tended to rate the exposures lower. The novice rater also had slightly more variability when rating than the other raters. When comparing objective and subjective estimates of hand forces, the subject was better able to estimate the hand forces than the observer and the subject-observer (r = 0.30) and between-rater agreement (r = 0.11) was poor, indicating the challenge of trying to subjectively estimate hand forces