The Use of Hip Radiographs in Primary Care: The Inter-Observer Agreement of Reporting Native Hip Pathology

Abstract

Plain radiographs are often the first line investigation in identifying the cause of hip pain, though they do not differentiate all morphologically normal and abnormal hips. Interpretation of radiographs is subjective, depending on the patient history and physical signs. A radiological report can be misleading and may lead to, unnecessary radiation exposure, a delay in referral and a delay to definitive treatment. This study was to investigate the inter-observer reliability in the reporting of plain radiographs. Sixty-one consecutive antero-posterior (AP) radiographs of native hips were identified that had been referred to one of the senior authors from the community with "hip pain". Images were anonymised within PACS (picture archiving and communication system) and reviewed by a consultant orthopaedic surgeon, two consultant musculoskeletal radiologists, one senior orthopaedic registrar and one senior radiology registrar. Each reviewer was given a pre-constructed proforma, and asked to report the radiographs. There is a varying degree of 'agreement' for many of the commonly used terms on a hip radiograph. Agreement between all observers varied, by description, between 23.3% (CAM Lesion) to 93.3% (AVN). Multi-rater agreement showed Kappa values ranging from 0.12 (poor) to 0.6 (moderate). Overall there were widespread inconsistencies, even amongst the most widely used terms. There is variable agreement amongst musculoskeletal radiology and orthopaedic experts when analysing plain radiographs of the native hip. This has implications for utility of reporting and therefore treatment in the community setting.