Bed rest has been shown to be an ineffective treatment for non-specific low back pain (LBP). Despite this, during a new episode of pain some patients still rely on bed rest. Which patients choose bed rest is however unknown. The objectives of the present study were, firstly, to assess characteristics of patients choosing bed rest in (sub)acute pain and secondly to study whether prolonged bed rest in the (sub)acute phase of pain will result in long term disability. A prospective longitudinal cohort study included 282 patients with non-specific LBP for less than 7 weeks. Main outcome measures were duration of bed rest (in three categories) and disability. Results showed that 33% of patients with (sub)acute LBP had bed rest, but only 8% stayed in bed for more than four days. An ordinal regression analysis revealed that behavioural factors (catastrophizing (OR = 1.05 per bed rest category p <0.01)) and fear of injury (OR = 1.05 per category p <0.01) rather than specific pain related factors (pain history (OR = 0.61 per category p = 0.16) and pain intensity (OR = 1.00 per category p = 0.63)) were associated with bed rest. Patients with prolonged bed rest in an early phase of pain were still more disabled after one year (p <0.01). Based on these results we conclude that prolonged bed rest in the early phase of pain is associated with a higher long term disability level. In preventing low back disability, GP screening for catastrophizing and fear of injury in LBP patients who had prolonged bed rest merits consideration.