Capturing data on CT radiation doses from RIS/PACS databases was more accurate and cost effective than self-reported survey data.
Capturing data on computed tomography radiation doses from RIS/PACS databases was more accurate and cost effective than self-reported survey data, according to the results of a single-center study published in the European Journal of Radiology.
Radiation dose information is currently captured using self-reported surveys, which have inherent bias because they do not capture a random sample of data and rely on participation by providers, which may not represent the full story of the community, researchers said.
“Nowadays the majority of radiographic examinations are captured digitally and for CT along with the images, information about the scanning parameters and radiation dose given to the patient are also available,” study author Rachel Moorin, MD, of the Centre for Population Health Research at Curtin University, Perth, Wash., told Diagnostic Imaging. “This represents an opportunity to collect radiation dose data in a more systematic and comprehensive manner. However it requires that providers participate by allowing their data to be captured for audit purposes.”
In their study, Moorin and colleagues compared data from self-reported surveys and that taken from the RIS/PACS for a variety of CT examination types to determine the reliability of the self-reported information. All of the data in the study was from one large metropolitan teaching hospital in Western Australia that used the same type of 64 slice scanner.
The researchers found that across six different clinical scenarios the number of sequences, CT dose index and dose length product varied according to the data collection method. Although a single sequence was reported for all scenarios on the self-reported survey, database information showed that only head and abdomen/pelvis CT had a median number of sequences of one.
Additionally, data from the RIS/PACS showed that the average CT dose index across cases was significantly lower than the average indicated on self-reported surveys for three of the clinical scenarios. For example, when looking at chest CT for obstructive pulmonary disease the self-reported data showed a CT dose index three times greater than that indicated with the RIS/PACS data.
When looking at organ and effective dose, results from the study showed that the disparity between collection methods was greater and more variable for organ dose than effective dose. PACS data had a substantially lower effective dose for three clinical scenarios and a substantially higher dose for two scenarios. Data from PACS showed higher doses in the organs at the limits of the scanning field compared with self-reported survey data, “most likely resulting from the reliance of the survey data on ‘standardized’ anatomical start and stop limits compared with actual start and stop limits in the PACS data,” the researchers wrote.
“This study has confirmed data extracted from RIS/PACS is superior to self-reported survey data and has shown survey data contains both proportional and systematic bias not consistent across CT examinations,” the researchers wrote. “We recommend national and local databases that are established to routinely capture aggregated and anonymous CT dose data for the development and monitoring of diagnostic reference levels and surveillance of population radiation dose.”