Radiographer targeted education to reduce repeat exposure in routine chest digital radiography

Abstract

Background: Repeat and reject images in routine chest digital radiography increase patient dose and workload. International benchmarks recommend very low repeat rates, still local audits report higher figures. Improving radiographer performance is a plausible lever for reduction. We aimed to evaluate whether radiographer targeted education and related workflow strategies reduce technically avoidable repeats in routine chest radiography. Methods: we conducted a PRISMA aligned systematic review of electronic databases (MEDLINE, PubMed, Embase, Scopus, Web of Science, and CINAHL, CENTRAL) from 2012 to 2025. Eligible designs assessed radiographer focused interventions (education modules, checklists, communication aids, QI bundles) against usual practice or pre intervention periods. Primary outcomes were repeat and reject rate or extra images; secondary outcomes included image acceptance and operational effects. Risk of bias was appraised with ROBINS-I, JBI, and NIH tools. Heterogeneity precluded meta-analysis; a structured narrative synthesis was performed. No prospective registration. Results: Six studies met inclusion criteria, including multi hospital quasi experimental programs, single site QI bundles, and an observer study. Education modules improved knowledge, motivation, and skills with significant, sustained reductions in repetition. A chest radiography bundle lowered extra frontal images from 4.6% to 3.3% (p=0.001). A multilingual instruction protocol improved inspiration quality, cutting poor inspiration rejects from 26% to 9% and increasing fully inspired images from 57.8% to 92.3%. An observer study showed radiologists accepted images more often than radiographers, indicating value in harmonized acceptance criteria and feedback. Conclusions: Radiographer targeted education, simple checklists, and language appropriate patient instructions reduce technically avoidable repeats in routine chest radiography.