Inter-Group Agreement in Chest Radiographic Quality Assessment: Comparing Final-Year Students and Experienced Radiographers Across Three Quality Assessment Domains During the COVID-19 Pandemic.

Abstract

Introduction: Reproducible image quality assessment (QA) by radiographers is vital to minimise inter-observer variation and reduce unnecessary repeat exposures. This study compared the consistency of chest X-ray (CXR) image quality assessments between final-year student radiographers and experienced radiographers across three distinct QA domains (positioning, sharpness and visibility) to observe skill consistency across different pathology categories.

Methods: This retrospective observational study analysed 70 de-identified CXRs acquired between May and September 2021 from patients presenting with symptoms of pulmonary infection during the COVID-19 pandemic and retrieved from the PACS archive of a tertiary teaching hospital in Sydney, Australia. Two experienced radiologists independently reviewed each case and discussed their interpretations to reach a consensus classification based on the available clinical information.

Results: No significant differences were found between groups in any domain across the different pathology categories (p > 0.05). Bland-Altman analysis showed minimal mean variability, with students rating marginally lower and narrow 95% limits of agreement across all domains. Overall ICC values indicated good-to-excellent reproducibility (ICC = 0.90, 95% CI 0.85-0.94). These single-institution findings should be interpreted within the dataset's context.

Conclusion: Final-year students and experienced radiographers applied QA criteria consistently, demonstrating robust skills across different pathology categories. The small negative bias observed may reflect a more cautious interpretive approach rather than meaningful disagreement. These findings indicate that final-year radiography students demonstrate agreement with experienced radiographers in the application of structured image-quality criteria under controlled assessment conditions.