Inter-examiner reliability of a proposed decision-making treatment based classification system for low back pain patients

Evolving evidence has shown increased clinical outcomes, when low back pain (LBP) patients are classified and receive matched physical treatment. The present study aimed to examine the inter-examiner reliability of a proposed new decision-making classification system for non-specific LBP patients, using a mixed simultaneous and independent examiner design. With minimal familiarization, two pairs of experienced physiotherapists trained in Orthopedic Manual Therapy (OMT) at two different out-patient clinics in primary care, examined and classified 64 consenting consecutive patients. Further, inter-examiner reliability on five examination items was examined. The agreement between examiners was expressed by percentage of agreement (%) and by the un-weighted (κ) or weighted (κ(w)) kappa coefficient. The overall % agreement, categorizing patients into one of four classifications was 80% and κ = 0.72. For each classification, pain modulation, stabilization exercise, mobilization and training, agreement was 90%, 83%, 58% and 89% (κ = 0.77, 0.67, 0.11 and 0.75), respectively. Agreement on five individual examination items was; irritability 82% (κ(w) = 0.41), specific movement pattern 68% (κ = 0.38), specific segmental signs 67% (κ = 0.28), uni- or bilateral signs 62% (κ = 0.42), and neurological signs and symptoms 92% (κ = 0.84). This study demonstrated that this new classification system had substantial inter-examiner reliability when used by clinically experienced OMT-trained physiotherapists. Agreement within classification was substantial, except for mobilization which was poor. Inter-examiner reliability for the individual examination items varied from fair to almost perfect. Further studies are needed to investigate utility and validity of this new classification system.