INTER- AND INTRA-OBSERVER RELIABILITY OF THE MIRELS’ SCORING SYSTEM FOR THE DETERMINATION OF PATHOLOGICAL FRACTURE RISK IN METASTATIC BONE LESIONS AMONG ORTHOPEDIC SURGEONS WITH DIFFERENT LEVELS OF EXPERTISE

Abstract

Objective: To assess inter- and intra-observer reliability of Mirels scoring for the determination of pathological fracture risk in metastatic bone lesions of 30 patients among six different levels of experienced orthopedic surgeons who were trained in the same university clinic. Material and Methods: Thirty patients were randomly selected from oncology unit consultations. Six observers were selected in accordance to their orthopedic experience. Mirels parameters except pain were evaluated by observers on two different times without the observers being aware of each other. Results: The Fleiss’ Kappa values were detected as κ=0.21, p<0.0001, and κ=0.15, p<0.0001 by inter-observers at the first and second observational points, respectively. The Kappa values were in perfect (κ=0.95), fair (κ=0.27) and fair (κ=0.10) agreements for the region, size and type of the metastatic involvement by inter-observers at the first observational point. The same parameters had similiar scores as κ=0.83, κ=0.13 and κ=0.28 for region, size and type for the second observation. Fair (κ=0.333), moderate (κ=0.413), fair (κ=0.225), slight (κ=0.035), fair (κ=0.369 and poor (κ=0.030) evidence of agreements were detected in the comparison of the first and second observations for total scores in seniority order. Conclusions: Kappa analysis showed perfect agreement for region, but slight to fair for size and type. There was a significant difference in overall scores across experience levels for the most and least experienced observers. A new rating system with revised parameters may be required to predict impending fractures.