Objectives To investigate the importance of multi-site musculoskeletal pain as a predictor of sickness absence days due to musculoskeletal disorders (MSD) among blue-collar employees, and to study what extent such a relationship might be confounded by physical loading at work.

Method Survey responses from 901 employees were linked to a food industry company’s record of sickness absence due to MSD (≥ four days). Generalised Linear Models (GLM) with negative binomial distribution assumption was used in order to determine associations between the occurrence of multi-site pain (no pain, one-site and multi-site pain), individual variables, work related variables and sickness absence days due to MSD during a four-year follow-up.

Results The high exposure group had about 92 and the low exposure about 72 all-cause sickness absence days yearly, and corresponding figures for absence due to MSD were 36 and 28. The share of MSD absence is about 40% irrespective of the exposure. Single site pain did not predict absence, whereas multi-site pain turned out as an independent predictor. Multi-site pain predicts absence in the group with low biomechanical exposure, but not in the group with high exposure. The p-values for interaction show that the groups differ significantly both in case of repetitive movements and in case of awkward postures.

Conclusions This prospective cohort study revealed very high level of sickness absence in biomechanically strenuous work, represented by manual work in food industry. On average, the employees were absent over 80 calendar days, i.e. almost three months, yearly.