Perspectives on the management of humerus fractures due to gunshot trauma: an inter- and intra-observer agreement and reliability study

Master Thesis

2017

Permanent link to this Item
Authors
Supervisors
Journal Title
Link to Journal
Journal ISSN
Volume Title
Publisher
Publisher

University of Cape Town

License
Series
Abstract
Background: Upper extremity fractures due to gunshot trauma are frequently treated at the level I trauma unit of Groote Schuur Hospital. There is no gold standard for the classification and management of such complex upper extremity fractures available to date and only few retrospective case studies on gunshots of the humerus were available. Interobserver agreement studies reported low levels of intra- and inter-rater reliability (IRR) for the classification of proximal humerus fractures using Neer and AO/OTA classification. The complexity of the fractures, the inconsistency of classification systems outcomes and the wide variety of treatment modalities demand evidence-based medicine. Aim: The primary aim was to assess the inter- and intraobserver agreement between surgeons in the classification and treatment of humerus fractures caused by gunshot trauma in a gunshot violence endemic area. The secondary aims were to analyse interobserver agreement with respect to debridement, removal of the bullet, the use of external fixators in patients with gunshot humerus fractures and to evaluate the effect of clinical scenarios surrounding surgical decision-making. Methods: This is an agreement study performed with a fixed panel of 32 observers who answered a set of 14 questions regarding classification and treatment by rating multiple X-ray views of a fixed set of 22 cases. The panel included junior registrars, senior registrars, orthopaedic trauma specialist and upper extremity specialists. Cases were extracted from the electronic Trauma Health Record between June 2014 and July 2016. Observers reviewed 16 midshaft and 6 proximal humerus fractures cases at 2 sessions with a 2-week interval. Descriptive statistics, Cohen's and Fleiss Kappa and rate of agreement were used to analyse data. Kappa was interpreted according to Landis and Koch guidelines. Results: There was slight yet significant overall interobserver agreement on the AO classification (k=0.20); the highest interobserver agreement ('fair') was achieved by the upper extremity specialists and senior registrars (k=0.28, 0.27). Overall interobserver reliability of agreement on preferred treatment was similar to classification agreement (k=0.18). Only trauma specialists achieved fair agreement with a significant difference compared to senior registrars and upper extremity specialists (k=0.26, 95%CI 0.21-0.32). Overall intraobserver reliability was fair for classification and moderate for treatment (k=0.39, 0.42). There was fair overall agreement on debridement of the wound (k=0.26) and removal of the bullet (k=0.31) and close to poor agreement for the use of temporary external fixators (k=0.03). Vascular injury was rated as influential factor on decision-making by the majority of observers (53.7%), followed by bilateral (37.1%) and other fractures (26.8%). Conclusions: This is the first intra- and interobserver agreement study that evaluated classification and treatment of gunshot humerus fractures in the light of a broader spectrum of patient- and fracture-related factors. Consistent with previous studies, there was low interobserver agreement for the classification and treatment of proximal humerus fractures, thereby contributing to the field of knowledge with specific evidence regarding gunshot trauma. Future research should further assess predictive factors in surgical decision-making and analyse global preferences in order to develop evidence-based classification and treatment guidelines for the management of patients with humerus fractures.
Description

Reference:

Collections