Bed rest is usually recommended for acute low back pain. Although the optimal duration of bed rest is uncertain, a given prescription may directly affect the number of days lost from work or other activities. In a randomized trial, we compared the consequences of recommending two days of bed rest (Group I) with those of recommending seven days (Group II). The subjects were 203 walk-in patients with mechanical low back pain; 78 percent had acute pain (less than or equal to 30 days), and none had marked neurologic deficits. Follow-up data were obtained at three weeks (93 percent) and three months (88 percent). Although compliance with the recommendation of bed rest was variable, patients randomly assigned to Group I missed 45 percent fewer days of work than those assigned to Group II (3.1 vs. 5.6 days, P = 0.01), and no differences were observed in other functional, physiologic, or perceived outcomes. For many patients without neuromotor deficits, clinicians may be able to recommend two days of bed rest rather than longer periods, without any perceptible difference in clinical outcome. If widely applied, this policy might substantially reduce absenteeism from work and the resulting indirect costs of low back pain for both patients and employers.