Establishing Diagnostic Reference Levels for CT Through a Provincial Medical Informatics Metadata Repository in Ontario.

PURPOSE: To determine whether computed tomography radiation dose data could be captured electronically across hospitals to derive regional diagnostic reference levels for quality improvement. METHODS: Data on consecutive computed tomography examinations from 8 hospitals were collected automatically in a central database (Repository) from April 2017 to September 2017. The most frequently performed examinations were used to determine the standard protocols for each hospital. Diagnostic reference levels across hospitals were derived using statistical distribution for 2 radiation dose metrics. These values were compared between hospitals, within and between hospitals by scanner and against national Health Canada achievable doses and diagnostic reference levels. RESULTS: Three master protocol groups, Head, Abdomen-Pelvis, and Chest-Abdomen-Pelvis, accounted for 43% of all valid studies (N = 40 277). For the Repository, 11 of 12 mean values and 75th percentile diagnostic reference levels were below the Health Canada mean and 75th percentile values, and one was the same as the Health Canada value. Mean radiation dose by protocol varied by as much as 97% between hospitals. There was no consistent pattern in the difference between mean doses between large and small hospitals. CONCLUSION: This electronic data acquisition process could be used to continually update achievable doses for frequently used computed tomography examinations in Ontario and eliminate the need for nationwide manual surveys. Results compared across institutions will allow hospitals to maintain achievable doses and lower patient exposure.

Tracking and Resolving CT Dose Metric Outliers Using Root-Cause Analysis.

PURPOSE: The aim of this study was to examine the frequency and type of outlier dose metrics for three common CT examination types on the basis of a root-cause analysis (RCA) approach. METHODS: Institutional review board approval was obtained for this retrospective observational study. The requirement to obtain informed consent was waived. Between January 2010 and December 2013, radiation dose metric data from 34,615 CT examinations, including 26,878 routine noncontrast CT head, 2,992 CT pulmonary angiographic (CTPA), and 4,745 renal colic examinations, were extracted from a radiation dose index monitoring database and manually cleaned. Dose outliers were identified on the basis of the statistical distribution of volumetric CT dose index and dose-length product for each examination type; values higher than the 99th percentile and less than the 1st percentile were flagged for RCA. RESULTS: There were 397 noncontrast CT head, 52 CTPA, and 80 renal colic outliers. Root causes for high-outlier examinations included repeat examinations due to patient motion (n = 122 [31%]), modified protocols mislabeled as "routine" (n = 69 [18%]), higher dose examinations for patients with large body habitus (n = 27 [7%]), repeat examinations due to technical artifacts (n = 20 [5%]), and repeat examinations due to suboptimal contrast timing (CTPA examinations) (n = 18 [5%]). Root causes for low-outlier examinations included low-dose protocols (n = 112 [29%]) and aborted examinations (n = 8 [2%]). On the basis of examination frequency over a 3-month period, the 90th and 10th percentile values were set in the radiation dose index monitoring database as thresholds for sending notifications to staff members responsible for outlier investigations. CONCLUSIONS: Systematic RCA of dose outliers identifies sources of variation and dose excess and pinpoints specific protocol and technical shortcomings for corrective action.