Tough Love for Low Back Pain

Abstract & Commentary

Source: Staal JB, et al. Ann Intern Med. 2004;140:77-84.

Synopsis: Graded activity can return a patient to full employment faster than usual care.

The lifetime prevalence of low back pain (LBP) is 58-84% with a point prevalence of 4-33%.1 It ranks fifth in physician visits.2 The American Productivity Audit revealed that in a 2-week period 3.20% of US employees were absent from work due to back pain, second only to headache at 5.43%.3 Into this arena strides Staal and colleagues from the Netherlands with a single-blind, randomized, controlled trial of a graded activity intervention for LBP. The study was performed at KLM Royal Dutch Airline’s occupational health services center at Schiphol Airport in Amsterdam and funded by the Dutch Health Insurance Executive Council.

Nonspecific graded activity interventions have several components. The foundation is the concept that pain behaviors (lying around, whining, and missing work) are subject to operant conditioning (ie, these are learned behaviors) and that recovery from pain is subject to unlearning the behaviors. A second, and equally important, concept is that pain does not necessarily mean harm. Exercise is the nemesis of pain behavior. Patients are encouraged to exercise, even if it hurts, because improvement in function, not pain relief, is the primary goal. Operationally, the physiotherapist puts the patient through his paces for 1 hour twice weekly. During each session the patient does a combination of common exercises. These exercises are aerobic (cycling or rowing), floor abdominal sit-ups, dynamic back extension, leg-press, latissimus pull-down, and standing from a low chair. Individually tailored exercises mimic the patient’s work; in this study, that could be lifting and moving suitcases. In the first 3 sessions, the patient does each exercise to his limit of pain tolerance. These limits are then averaged, and the averages become the baseline for future sessions. Together, the patient and the therapist decide at what level the exercises must be done at the end of therapy. The patient decides on a return-to-work date. Then an exercise quota is established; at each session the bar is set a little higher. The patient exercises to the quota, even if it is painful.

In this study, 134 KLM employees were randomized to 2 groups that were stratified by work type and level of pain. An occupational physician made the first determination whether a worker who was absent with LBP was eligible for inclusion into the study. These people were then referred to a research assistant who determined whether they met the inclusion (full or partial work absence secondary to nonspecific LBP with symptoms > 4 weeks) and exclusion criteria (LBP with radiculopathy below the knee, heart disease precluding exercise, consideration of legal action, or pregnancy). The control group received usual care, which consisted of guidance about work-related problems and barriers to return to work and advise on ergonomics and prevention from the occupational physician. The intervention group received usual care plus graded activity. The outcomes of interest were number of days absent from work due to LBP, functional status, and pain.

The researchers chose a difference of 5 days of work absence between the control and intervention groups as clinically and economically significant (the reduction in days absent pays for the cost of the intervention) and used this to calculate how many participants they would need to show statistical significance. That number was 70.

There were 67 workers in both groups. The groups were well matched. The workers’ average age was in the late 30s, and men made up greater than 90%. The various work categories (baggage handlers, maintenance, cargo, cockpit, and passenger services) were evenly distributed between to the 2 groups. In both groups the workers were absent from work for an average of 6 weeks before randomization. After randomization, the workers in the graded activity group were absent 58 days vs 87 days for the usual care group. This was statistically significant. The workers in both groups returned to work at essentially the same rate up to 50 days; thereafter, the graded activity group returned in greater frequency. Both groups experienced improvement in pain and functional status with a nonsignificant trend in favor of the graded activity group.

Comment by Allan J. Wilke, MD

Methodologically, this study raises some questions. Acute low back pain is usually self-limited with most people recovering and returning to work within 1 month.4 These workers were off for 6 weeks even before entering the study. Were they more severely injured? Or could it be that the Netherlands’s disability system that pays workers full salary for the first year of absence encourages a slow return to work? The occupational physician made the first cut. What interior criteria did that physician use to select eligible participants? Were the workers who were not referred to the research assistant in some way different than the workers who were? Although their power calculations determined that they needed 70 participants in each group to show a 5-day difference, they had only 67. Because they were able to demonstrate a 29-day difference, the results were still significant.

Economically, this type of intervention makes a lot of sense, but when a patient presents to me complaining of back pain, the first thing out of his mouth is not, "Gee, doc, get me back to work!" The patient sees me looking for pain relief. In the long run, though, pain relief without return to work is an unsatisfactory outcome. Treatment for LBP takes many forms, some effective, others not, and still others harmful. A recent meta-analysis found that massage is effective for back pain, but spinal manipulation was not any more effective than conventional therapy, and studies of acupuncture are so poor as to make evaluation of its effectiveness unclear.5 Muscle relaxants reduce pain, decrease muscle tension, and increase mobility, but drowsiness, dizziness, and dependency are common.6 Back exercises are not effective for acute LBP, but may allow people with chronic LBP to resume normal daily activities and work.7 Nonsteroidal anti-inflammatory drugs are effective for short-term relief in patients with acute LBP; there is no NSAID clearly more effective than another.8 Bed rest is not effective and may delay recovery. Simply advising patients to stay active and to continue ordinary activities9 and telling them "hurt does not mean harm" 10 may be the easiest method to get them back to work. However, the US Preventive Health Services Task Force "concludes that the evidence is insufficient to recommend for or against the routine use of interventions to prevent low back pain in adults in primary care settings."11

Dr. Wilke, Assistant Professor of Family Medicine, Medical College of Ohio, Toledo, OH, is Associate Editor of Internal Medicine Alert.


1. Woolf AD, Pfleger B. Bull World Health Organ. 2003;81:646-656.

2. Deyo RA, Weinstein JN. N Engl J Med.. 2001;344:363-370.

3. Stewart WF, et al. JAMA. 2003;290:2443-2454.

4. Pengel LH, et al. BMJ. 2003;327:323-327.

5. Cherkin DC, et al. Ann Intern Med. 2003;138:898-906.

6. van Tulder MW, et al. Cochrane Database Syst Rev. 2003;(2):CD004252

7. van Tulder M, et al. Spine. 2000;25:2784-2796.

8. van Tulder MW, et al. Spine. 2000;25:2501-2513.

9. Waddell G, Feder G, Lewis M. Br J Gen Pract. 1997; 47:647-652.

10. Weinstein J. Ann Intern Med.. 2004;140:142-143.

11. U.S. Preventive Services Task Force. 2004.