Introduction Low back pain (LBP) is one of the leading causes of disability all over the world. We performed a secondary prevention program of LBP among employees that reported mild or moderate level low back symptoms in a large forestry industry complex.

Methods First, respondents of an employee survey (n=2480; response rate 71%) were eligible into this study, if they fulfilled pre-defined low back (LB) specific risk assessment criteria. Secondly, eligible employees (n=505, 66% males, 45 y) were divided into two sub-cohorts, ‘Mild’ and ‘Moderate’ LBP, according to recent LB pain intensity. Sub-cohort Mild (n=181, 47 refused) was randomised into two intervention arms, both receiving back book information and the other arm also additional face-to-face patient information. Sub-cohort Moderate (n=126, 17 refused) was randomised into three groups, receiving either one of two active exercise interventions or LB specific advice from their occupational health (OH) physician. All intervention arms in Mild and Moderate were controlled by their respective natural course (NC) of LBP groups (n=83 and n=50, respectively). Primary outcomes were disability (Roland-Morris Disability Questionnaire (0–18) and Oswestry Disability sum index, 0–50), LB pain (Visual Analogue Scale, 0–100 mm) and total sickness absence days (SA).

Results Mild: Compared to NC, pain, disability and SA decreased after both interventions and back book information alone was also cost-effective. Moderate: Compared to NC, pain and disability decreased after both active interventions but SA did not. OH physician’s advice was not effective. Interventions in Moderate were not cost-effective in two years.

Discussion Simple patient information was effective and also cost-effective in mild LBP. Active LB specific interventions were effective but not cost-effective after two years in moderate LBP. OH physician’s advice was not effective. Population based LB specific risk assessment seems feasible. In general, proactive management of LBP is recommendable in the OH setting.