¹Department of Occupational and Environmental Medicine, Bispebjerg Hospital, Copenhagen, Denmark

Text

Introduction: Epidemiological studies of occupational risk factors for rare disorders require large study populations with adequate exposure estimates. Job exposure matrices (JEMs) linked to national information on standardized job titles may enablesuch large studies. We aimed to establish and validate a JEM for occupational hand‐wrist exposures that could be linked to Danish national register data on job titles and hand‐wrist disorders.

Methods: We developed a JEM for hand‐wrist repetition, force, vibration, and computer work in 96 job groups covering 91% of the 2227 occupational titles in the Danish version of the International Standard Classification of Occupation‐88, and examined inter‐rater reliability of five expert ratings. Poisson regression models were used to estimate incidence rate ratios for the association of carpal tunnel syndrome (CTS) with the level of repetitive movements, force, vibration, and hours of computer work described by the JEM, adjusted for relevant confounders.

Results: Interrater Spearman correlations of expert ratings ranged from 0.52-0.73 across the four exposures. We followed 2.6 million persons for 55.3 million years and found 71,718 cases of CTS (incidence rate (IR) of 13/10,000 person‐years). Crude IRRs for levels of repetition, force, and vibration showed a steady increase with increasing levels of exposure; the opposite trend was observed for computer work. Adjusted for levels of force, IRRs for repetition were still statistically significant although they were considerably attenuated except for the highest level for which the IRR increased; IRRs for vibration were attenuated to close to unity and no longer statistically significant. After adjustment for repetition, IRRs for force showed modest decreases except for the highest level for which the IRR was unchanged. In a sub‐cohort with low exposure to repetition and force, increasing computer use was still negatively associated with CTS and the pattern and risk estimates were similar to the results of the main analyses.

Discussion: The JEM based on expert ratings had fair to good interrater reliability. The incidence of CTS increased with increasing levels of force, hand‐wrist repetition, and vibration, Exposure‐response patterns for repetition and vibration became less consistent after adjustment for force. The interaction between repetition and force was complex and did not support an overall positive interaction. Computer work was negatively associated with incident CTS.

Conclusion: The JEM was able to identify known risk factors for CTS consistent with current evidence, and provided further associations on exposure‐response patterns, mutual exposure adjustment, and interaction effects between repetition and force. The reliability of expert assessments of hand‐wrist physical exposures was fair to good.