The study is a randomized, controlled, clinical trial in the Swedish province of Gotland, with primary data collection from 1994 to 1998. The recruitment population consisted of the 19000 employed persons born in Sweden, 20-55 years of age. Additional inclusion criteria were:
Acute or subacute perceived low back pain with or without pain radiating to one or both legs, not requiring acute surgical or rheumatological care, low back pain was required to be the dominating symptom, symptom duration of 3 months or less, preceded by at least 2 months of relative freedom from symptoms, consent to treatment and follow-up for 10 weeks, agreement not to consult other therapists during the treatment period, absence of conditions or circumstances that might jeopardize completion of treatment and follow-up, no previous treatment of current complaints with specific mobilization or manipulation, and no previous participation in the present study.
In Sweden, the National Social Insurance Offices (a government agency running the mandatory national social insurance scheme applicable to all Swedish residents) handle sick leave with a duration of two weeks or more. They as well as general practitioners (GP) and other physicians referred patients. The recruiting physician examined all patients, performed a physical examination, and made the final assessment whether they fulfilled the inclusion criteria. Of the 316 patients who were referred to the study, 111 did not meet the inclusion criteria and 45 declined participation. The 160 patients that gave informed consent were entered in the study. After baseline evaluation they were randomized by the study monitor, using sealed pre-prepared envelopes with group assignment derived from a random number table. The envelopes were inaccessible to anyone but the monitor. A design with four treatment groups was used: two experimental and two reference groups with treatment items added successively. The two-group comparison (reference vs experimental treatment) was the primary planned analysis, and pain, disability rating and sickness absence were the main outcome measures. A weighted randomization procedure was used, with 45% of the patients to reference and 55% to the experimental therapies. The Research Ethics Committee at Uppsala University approved the study.
The stay active concept was the basic management strategy in all study groups. Treatment was individual, in groups or both. Treatment modalities were chosen from a group-specific 'toolbox' after clinical assessment of the patients and according to need. Two orthopedic surgeons at Visby Hospital and eight physiotherapists treated the reference patients. In accordance with the study design, muscle stretching was a treatment option in 51% of the reference group, 41% actually received muscle stretching. Two GPs and nine physiotherapists treated the experimental patients. They received the full reference treatment, plus specific mobilization, spinal manipulation, and auto-traction when indicated. Steroid injections were a treatment option in 52% of the experimental group, 36% actually received injections.
The study population was followed over ten weeks with measurements of outcome and progression variables. Outcome measures were self-reported back pain, disability rating, return to work as assessed from sick leave records, quality of life measures, and pain drawings. Moreover, sick leave data during two years of follow up from baseline were obtained.
