Intra- and inter-observer reliability of Dias-Tachdjian classification in pediatric ankle fractures: do clinical experience and expertise matter?

Abstract

TheDias-Tachdjian classificationis the most commonly used system for the classification of pediatric ankle fractures, but its inter- andintra-observer reliabilityhas not been studied in detail. Also, the impact of theclinician’s experienceandexpertiseon the reliability of this system is unknown. This study aimed: (1) to determine the intra- andinter-observer reliabilityof theDias-Tachdjian classificationand (2) to investigate the effect of theclinician’s experienceandexpertiseon the reliability of this system. Anteroposterior and lateral ankle radiographs of 56 children (34 male, 22 female) with ankle fractures, aged between 3 and 14 years, with open growth cartilages, were retrospectively identified and included in the study. Each patient radiograph was examined by 10 observers from two different specialties with different levels of clinical experience (two orthopedic surgeons with interest in pediatric orthopedics, three orthopedic surgeons with no interest in pediatric orthopedics, three orthopedic residents, and two radiology specialists) from two different specialties (orthopedics and radiology). All observers were then asked to classify pediatric ankle fractures at 6-week intervals per theDias-Tachdjian classificationsystem. Overall,intra-observer reliabilityas substantial to very good (κ = 0.77–0.95,P< 0.01), butinter-observer reliabilityas fair for both assessments (κ = 0.21,P< 0.01 and κ = 0.20,P< 0.01 for the first and second occasions, respectively).Inter-observer reliabilityamong pediatric orthopedic surgeons as very good (κ = 0.90, 95% CI = 0.86–0.94,P< 0.01 and κ = 0.82, 95% CI = 0.71–0.93,P< 0.01 for the first and second occasions, respectively). Orthopedic surgeons with no special interest in pediatric orthopedics demonstrated substantial agreement in the first occasion (κ = 0.63, 95% CI = 0.53–0.72,P< 0.01) but moderate in the second one. Orthopedic residents exhibited moderate levels of agreement in each assessment period (κ = 0.58, 95% CI = 0.47–0.68,P< 0.01 and κ = 0.44, 95% CI = 0.37–0.51,P< 0.01 for the first and second occasion, respectively). Considering that the specialists dealing with pediatric orthopedics show very good consistency forDias-Tachdjian classification, both within and between observers, consistency in the identification of the ankle fracture models increases as the interest in the field of pediatric orthopedics intensifies.