A test–retest design was used.
To assess interexaminer reliability of the McKenzie method for performing clinical tests and classifying patients with low back pain.
Summary of Background Data.
Clinical methods and tests classifying patients with nonspecific low back pain have been based mainly on symptom duration or extent of pain referral. The McKenzie mechanical diagnostic and classification approach is a widely used noninvasive, low-technology method of assessing patients with low back pain. However, little is known about the interexaminer reliability of the method, previous studies having yielded conflicting results.
For this study, 39 volunteers with low back pain, mean age 40 years (range, 24–55 years), were blindly assessed by two physical therapists trained in the McKenzie method. The variability of two examiners for binary decisions was expressed by the kappa coefficient, and by the proportion of observed agreement, as calculated from a 2 × 2 contingency table of concordance.
On the basis of pure observation alone, agreement among clinical tests on the presence and direction of lateral shift was 77% (κ = 0.2;P < 0.248) and 79% (κ = 0.4;P < 0.003), respectively. Agreement on the relevance of lateral shift and the lateral component according to symptom responses was 85% (κ = 0.7;P < 0.000) and 92% (κ= 0.4;P < 0.021), respectively. Using the repeated movements and static end-range loading strategy to define the centralization phenomenon and directional preference, agreement was 95% (κ = 0.7;P < 0.002) and 90% (κ = 0.9;P < 0.000), respectively. When patients with low back pain were classified into the McKenzie main syndromes and into specific subgroups, agreement was 95% (κ = 0.6;P < 0.000) and 74% (κ = 0.7;P < 0.000), respectively.
Interexaminer reliability of the McKenzie lumbar spine assessment in performing clinical tests and classifying patients with low back pain into syndromes were good and statistically significant when the examiners had been trained in the McKenzie method.