Computed tomography (CT) of the medial clavicular epiphysis has been well established in forensic age estimations of living individuals undergoing criminal proceedings. The present study examines the influence of the examiner's qualification on the determination of the clavicular ossification stage. Additionally, the most frequent sources of error made during the stage assessment process should be uncovered. To this end, thin-slice CT scans of 1,420 clavicles were evaluated by one inexperienced and two experienced examiners. The latter did the evaluations in consensus. Two classification systems, a five-stage system and a substaging system for the main stages 2 and 3, were used. Prior to three of his six assessment sessions, the inexperienced examiner was specifically taught staging of clavicles. Comparison of the examiners' results revealed increasing inter- and intraobserver agreements with increasing state of qualification of the inexperienced examiner (from κ= 0.494 to 0.674 and from κ= 0.634 to 0.783, respectively). The attribution of a not-assessable anatomic shape variant to an ossification stage was identified as the most frequent error during stage determination (n= 349), followed by the overlooking of the epiphyseal scar defining stage 4 (n= 144). As to the clavicular substages, classifying substage 3a instead of 3b was found to be the most frequent error (n= 69). The data of this study indicate that κ values must not be considered as objective measures for inter- and intraobserver agreements. Furthermore, a high degree of specific qualification, particularly the knowledge about the diversity of anatomic shape variants, appears to be mandatory and indispensable for reliable evaluation of the medial clavicular epiphysis.
