We conducted a prospective multicentre observational study of patients with sciatica. The sample size was calculated based on an expectation that surgical treatment would be necessary for 30% of the patients and that 70% of those who were surgically treated and 50% of those who were not surgically treated would experience a good outcome . Power analysis indicated a sample size on 300 patients requiring 90% power with a two-sided exact Fisher test of 5% significance level. Taking loss-to-follow-up into consideration we ended up with a study sample of 400 patients. The study protocol was approved by the Regional Committee for Medical Research Ethics and The Ombudsmann for Privacy in Research at the Norwegian Social Science Data Services.
Patients were recruited from specialty back clinics at four hospitals in southeast Norway (Sykehuset Østfold, Sørlandet Sykehus, Oslo Universitetssykehus Ullevål and Sykehuset Innlandet). Patients were referred to the specialty back clinics from the primary health care service. Assessment and treatment were conducted as usual in the back clinics. A physician or physiotherapist informed consecutive eligible patients about the study. Inclusion criteria were age ≥18 years and radiating pain below the knee level, and/or paresis and a disc herniation at the corresponding level and side according to magnetic resonance imaging (MRI) or computed tomography (CT). Exclusion criteria were prior surgery at the same disc level, fracture, infection, malignancy, pregnancy, and lack of fluency in Norwegian. All patients signed an informed consent form after receiving oral and written information.
At the first visit, a questionnaire on sociodemographic factors (age, gender, education length, smoking status and work status) was completed. Patients also reported their history of previous sciatica and the duration of relevant sciatica episodes. A clinical examination was carried out by trained physiotherapists or physicians at the departments of rheumatology, physical medicine and rehabilitation, or orthopaedics.
The assessment included motor function of the muscles of the knee, leg and toe, the Trendelenburg test, sensory loss in the leg, reflexes of the Achilles tendon and patella (all deemed abnormal if reduced) and the straight leg raising test (deemed abnormal if <60°). Number of tender-points were recorded. Standardisation of testing procedures was done in meetings with the participating centres.
A follow-up questionnaire and a prepaid envelope were sent to the patients after 3, 6 and 12 months. The date was recorded for patients who underwent back surgery. A study nurse sent a reminder after 2 weeks if no reply was obtained.
Leg pain and low back pain were measured using the visual analogue scale (VAS) 0–100. The Sciatica Bothersomeness Index (SBI), which is a composite score of four questions (each score ranging from 0–6) that include elements of leg pain and sensory and motor disturbances, was used [7, 16]. Total score ranges from 0 to 24 and higher scores indicate worse symptoms. The Maine-Seattle Back Questionnaire (MSBQ) [2, 7] which was developed from the RMDQ and modified for sciatica patients, was used to measure functional status. The scale is composed of 12 items and the score ranges from 0 to 12, higher scores indicating greater disability. The MSBQ mainly assesses disability, but also includes questions about pain. The Norwegian versions of the SBI and the MSBQ were recently validated . The bodily pain and physical functioning subscales of SF-36  are parts of a generic quality of life questionnaire in widespread use, including studies on sciatica [1, 23, 24]. Scores range from 0 to 100 and higher values indicate better health.
The follow-up questionnaires included a 7-item global change scale for leg and back pain with the following categories: Completely recovered, Much better, Better, A little better, No change, A little worse and Much worse. Patients were categorised as success if they responded with Completely recovered or Much better. Satisfaction to spend the rest of life in the current state was registered using a 5-item Likert scale with the wording Very satisfied, A little satisfied, Neither satisfied nor not, A little dissatisfied, and Very Dissatisfied.
Baseline and follow-up comparisons between groups were analysed using an independent t-test for continuous data and χ2 for categorical data. Floor or ceiling effects were defined as >15% of scores in the lowest or highest parts of the scales, respectively .
A receiver operating characteristic (ROC) curve was obtained by plotting every possible cut-off score’s sensitivity on the y-axis against 1-specificity on the x-axis. Sensitivity was defined as the proportion of patients who were correctly classified in the success group (Completely recovered or Much better). Specificity was defined as the proportion of patients who were correctly classified in the non-success group. The value on the curve closest to the upper left corner shows the highest sensitivity and specificity and was considered the best cut-off score for distinguishing between success and non-success. The area under the ROC curve (AUC) reflects the scoring system’s ability in differentiating between success and non-success. An AUC value >0.70 was considered satisfactory .
McNemar’s test was used when comparing sensitivity of two diagnostic measures where the cut-off point was set to specificity 80% for each measure.
The Statistical Package for Social Sciences Version 17.0 (SPSS Inc., Chicago, IL) was used for data analysis.