Variabilities in X-ray diagnostic reference levels.

OBJECTIVES: To estimate the variability of X-ray diagnostic reference levels (DRLs) depending on the number of X-ray devices and data per device. METHODS: Dose-area products (DAP) were collected by the national nuclear control agency from the 590 devices installed in 345 medical centers in the country. From 2015 to 2017, the number of chest (postero-anterior (PA) view alone, and both postero-anterior and lateral views (PA/LAT)), abdomen, pelvis, and lumbar spine examinations collected in these centers ranged from 23,000 to 77,000. The impact of the number of devices and DAP data per device on DRLs' variabilities (95th confidence intervals divided by medians) is estimated using a bootstrapping method as a function of the number of devices and DAP per device. RESULTS: The DRLs' variabilities ranged from 30 to 200% depending on the number of devices and DAP data per device but stabilized at 30% when the number of devices was higher than 200 for chest PA and abdomen examinations, 300 for lumbar spine and pelvis examinations, and 400 for chest PA/LAT examinations, regardless of the number of DAP data per device. Extrapolations of our results suggest that thousands of devices are necessary to reduce DRLs' variabilities to 10%. CONCLUSION: DAP-related DRL variabilities are high but only moderately influenced by the number of DAP data per device and of devices provided this number is higher than 200 to 400 devices according to the type of examination. Harmonization of methods of data collection between the authorities of the EU states should be recommended. KEY POINTS: • DAP-related DRLs are not fixed values but ranges of values with at least 30% variability. • DAP-related DRLs strongly depend on the number of devices included when lower than 100. • If the number of devices included exceeds 200 to 400, the DRLs' variabilities do not depend on the number of DAP per device and should not exceed 30%.