Clinical tests
Age and Comorbidities Three studies reported the sensitivity for the patient characteristic age [22, 40, 41]. The sensitivity of the patient characteristic younger than 60 years was 15%; the specificity was 62% [41]. The sensitivity of the patient characteristic older than 65 years was 77%; the specificity was 69% [40]. The sensitivity for the presence of orthopaedic conditions was 18%; the specificity was 91% [56].
Symptoms Four studies reported the sensitivity for different pain locations [12, 22, 40, 50]. Radiating pain, calf pain, thigh pain, and moderate back pain all had a high sensitivity (>90%) [12, 22, 50]. Bilateral symptoms and leg pain that was worse than back pain had a specificity of 98% and 92% [12, 22]. Exacerbation while standing up had a sensitivity of 92% [56]; no pain when seated and improvement when bending forward had a specificity of 93% and 92% [40, 41].
The presence of symptoms related to cauda equina syndrome had a specificity of 99%. The specificity of urinary disturbance was 98% [41]. However, these symptoms were insensitive and present in only a small percentage of patients.
Physical examination In general, physical examination tests had a lower sensitivity than clinical symptoms. One study found that the specificity of a wide-based gait was 97%, and the specificity of an abnormal Romberg test result was 91% [40].
Diagnostic support tools Two studies used predictor variables that were independently associated with LSS to create risk scores for diagnosing LSS [39, 41, 56]. The sensitivity of a score of 7 or higher on a clinical diagnostic tool including history and examination findings was 95% [39] and 93% [41]; the specificity was 40% [39] and 72% [41]. Sensitivity was optimized by the combination of history and examination findings, but this resulted in a lower overall specificity. The sensitivity of a score of 5 or higher on a diagnostic tool including only questionnaire-based items was 81%; the specificity was 58% [56]. On testing in a validation sample, the sensitivity was 75%, and the specificity was 51% [56].
One study created a diagnostic tool to indicate the likelihood of the presence of LSS, composed of patient characteristics and clinical symptoms [22]. Having 4 of 5 positive findings had a sensitivity of 6% and a specificity of 98%.
One other study investigated the diagnostic accuracy of a Pain Response to Activity and Position Questionnaire (PRAP) [50]. A positive PRAP had a sensitivity of 52%, and a specificity of 74%.
Pain drawings One study investigated the diagnostic accuracy of pain drawings with a clinical reference standard [43-45]. The sensitivity of statistical analysis of the pain drawings in a two category prediction was 34%; the specificity was 83% [43]. The sensitivity of an expert evaluation of the pain drawings was 58%; the specificity was 88% [44].
Treadmill Two studies investigated the diagnostic accuracy of a treadmill test. One study used CT or MRI as a reference standard; the sensitivity was 77%, and the specificity was 95% [31]. The other study had myelography as a reference standard; the sensitivity ranged from 38% to 100%, and the specificity ranged from 33% to 87% [37].
Gait-analyses One study investigated the diagnostic accuracy of gait-analyses; this study showed a sensitivity of 97% and a specificity of 80% [48].
Physiotherapist assessment One study investigated the diagnostic accuracy of a physiotherapist assessment to detect anatomic lumbar stenosis; this study showed a sensitivity of 23% [42].