[Interrater studies of evaluating the reliability of somatic findings].

20 patients (11 female, 9 male) without low back pain were included in two interrater studies carried out by three physical therapists. The following variables were investigated: shape of the spine, iliac crest heights, twisting of the pelvis, spine test, length of legs, Patrick sign, Schober test, fingertip-to-floor measurements, flexion and extension of the lumbar spine, straight-leg raising, and length of the iliopsoas and the rectus muscles. Intraclass coefficients for the ordinal variables, and kappa coefficients for the nominal variables were calculated for evaluation of interrater agreement. In both investigations, the Schober sign, lumbar flexion, fingertip-to-floor measurements, straight-leg raising of the left leg, and lengths of both legs were almost perfectly reliable. All other variables exhibited a lower reliability between the three therapists.