Exploring Image Quality Judgements in Radiography

Background. Radiographers decide routinely if a radiograph is fit for reporting or should be repeated. This has direct implications for the patient radiation dose and the reporting by the radiologist and potentially diagnostic accuracy. Current literature describes decision making of radiographers in a broad way without describing the role of key factors (such as contrast, position, dose) in influencing these decisions. This study aims to investigate how radiographers incorporate different factors, from technical to cultural, into their reasoning and to explore the importance of image quality judgements. Methods: A mixed method approach was taken and four major studies conducted. In the first study ratings of various factors that might contribute to image quality judgments were correlated with overall perceptions of image acceptability by radiographers and radiologists, and the results compared, to determine which factors were most associated with image acceptability in each profession. These results were enriched with data from cognitive interviews exploring the reasoning of radiographers concerning the rejection or acceptance of a radiograph for diagnosis. Additionally, the correlation of this final decision to reject or accept was correlated with radiologists’ diagnostic accuracy. All methods above had an intercultural element, defined as the comparison between countries or the inclusion of observers from different counties. In a qualitative follow up study radiological reports were explored for evidence of feedback on image quality. Results: Radiographers and radiologists do correlate for the assessment of the visualisation of anatomical structures, with. radiographers demonstrating a more critical attitude. Comparable results were demonstrated in the different countries and in the different studies of this thesis. The interviews demonstrated the inclusion of context and patient in the reasoning of radiographers. The results indicate that for radiographers factors, such as the patient's ability to maintain a certain positioning, into the final decision making to repeat a radiograph or not. No evidence was found to support that the judgement of clinical usability by radiographers correlates with the diagnostic expertise or confidence of the radiologist. Neither did the assessment of the radiological reports highlight feedback concerning image quality. Conclusion: The goal for this research was set to explore image quality judgement in radiography. From a quantitative perspective, this first study highlighted the absence of a correlation between the classification of radiographs for the clinical usability, defined as RadLex and assessment of anatomical structures. Additionally, the study also found a predictive value for the perceived dose of the radiograph. A qualitative approach in the second study delivered more insights into the reasoning of radiographers. The study demonstrated how radiographs include context and patient when deciding to accept or reject the radiograph. Following up with a quantitative approach to link the assessment of radiographers (in anatomical structures and categorisation for clinical usability) to the diagnostic accuracy of radiologists did not demonstrate a correlation. The final study, a mixed approach based on the assessment of reports and the classification of radiographs, did neither demonstrate a correlation to link the assessment by radiographers for clinical usability to report. A further understanding of the radiographer's decision making is needed as a possible path in reducing the number of unneeded rejects. For education and the introduction of innovative technologies, such as AI, the role of reflection on the imaging strategy is crucial in the formation of correct patient templates.