Not just about machines: Analysis of MRI quality determinants in the Turkish health system context.

This study aims to investigate the determinants of the quality MRI in the Turkish healthcare system. The analysis is done by analysing the referred cases to a major university radiology department in Turkey, and matching the hospital and MRI use characteristics of the source institutions, where the original MRI was taken. Quality of MRI was measured by specialist radiologists. The resulting quality was analysed by gender and imaging area characteristics, source institutional quality, MRI use statistics in source institution and MRI machine use inclination of the source institution. Chi-square and logistic regression were conducted, with regional fixed effects. In the largest dataset, the highest quality institutions have significantly higher average expected MRI quality compared to one level beneath them (0.74 vs. 0.63) (P = 0.02), there is also a significant MRI quality difference between the second highest level of institution, and the third and the fourth (0.63-0.54). Smaller (<0.1) but significant quality difference (P = 0.05) exists for institutions with the lowest two quality levels. In the smaller dataset, with data only from the lowest two institutional quality groups, with a finer institutional quality grading, differences in institutional quality is again found to be a significant driver of MRI quality (P = 0.035).