For some CT and conventional radiology procedures, additional information collected during surveys allowed to differentiate between single-shot and multiphasic procedures, which was a first step towards considering the clinical indication. In particular, DRLs for CTDIvol were computed only from monophasic examinations while DRLs for DLP were computed for monophasic examinations (the dose for the “main” scan) and non-differentiated procedures (cumulative DLP).

DRLs and other percentiles of the distributions were continuously decreasing despite the relatively short period between surveys, but progressively stagnated for CT and mammography. Between the last surveys the decrease of DRLs for conventional radiography ranged between 17% and 39% for adults. Factors influencing the evolution of DRLs were not investigated. Last results were in accordance with European values [2].

For children, DRLs were established per age and weight category in the last few years but no real evolution was observed taking into account the poor number of data. Results were in accordance with European values from the PiDRL project [3].

For mammography, thorough optimisation, mostly due to the use of this technique in screening programs, has offered well optimised protocols on most devices. This results in relatively small deviations on the curves. Using the 75th percentile as DRL could lead to inappropriate dose reductions, hence the competent authority FANC decided to use the 95th percentile as DRL. 

For mammography and conventional radiology, splitting the data per type of detector allowed to set the DRL for devices with DR detectors respectively 17-21% and 52-60% lower than that with CR or film detectors. Between iterations 2014 and 2017, the number of devices with DR detectors increased by 30% for mammography and doubled for conventional radiology. 

No DRLs were established for interventional procedures due to lack of data and large discrepancies between individual procedures with the same name. However, the Belgian competent authority recommended strongly the active use of trigger levels for skin effects, as a first parameter to be aware and to follow up of the doses that are used [4].

For PTCA procedures, a substantial part of data were observed to be above these threshold levels. A decrease in procedures exceeding the trigger levels was observed between the last surveys.

Thresholds on fluoroscopy time related to the trigger levels were estimated for PTCA procedures. For instance, the risk of exceeding trigger level is expected to be very low for a fluoroscopy time lower than 5 minutes while it can be very high for a fluoroscopy time greater than 25 minutes. 

Numerical values illustrated in figures and leading to all the observations and considerations above are available on www.fanc.fgov.be [5].