Is it possible to discriminate workers with a higher prevalence of low back pain considering daily exposure time in a work-related lumbar posture? A diagnostic accuracy study