Background. Self-reported task durations are frequently used as input when assessing exposures related to musculoskeletal disorders (MSDs), in spite of being associated with errors. This work aimed to estimate the extent to which low-back pain status, true task duration, biomechanical load and pace influence errors in self-reported task durations.

Methods. 48 workers with—and 48 workers without— low-back pain, matched by gender, age group and job type, were recruited through a health service provider. T-tests con-firmed that matching was successful. Each worker performed three standardized tasks—i.e. shelving boxes, filing journals and typing texts—in a combination of the following conditions: one of three durations (60, 80 or 100 minutes); two paces in shelving (walking at 3 km/hr vs. 6 km/hr); and two loads in shelving (box weight 1.25 vs. 2.5 kg). Partici-pants were asked about the perceived duration of each task immediately after the work session while being aware of the total duration of the session. Posture and kinematics of the back (iLMM™) and heart rate (portable Polar®) were monitored throughout sessions.

Results. Regression analyses indicated that task type, true task duration and low-back pain status affect errors in self-reported task durations. Workers with low-back pain overestimated the shelving task more than workers without pain, by 15 to 36 minutes, depending on the true duration of the task. This occurred at the expense of a larger underestimation of the other two tasks, and mainly the typing task.

Discussion. Since errors in self-reported task duration appear to be significantly de-pendent on the worker’s musculoskeletal pain status, as well as factors in the job, we recommend that efforts be made to correct such errors by calibration modeling, or, at a minimum, that researchers be aware of this potential effect on exposure assessments and on epidemiological research that deals with work-related MSDs.