How DOSE assists medical physicists in identifying and analysing significant events - part 1
July 22, 2021
By Dr. rer. nat. Hans Dieter Nagel
Part 1: Identification of significant events according to StrlSchV Annex 14 I and II (1)
The ability to identify and process significant events is one of the most important functions of a dose management system (DMS). In the German legislation, the criteria for this are speci-fied in Annex 14 of the Radiation Protection Ordinance . In relation to a group of persons (criteria I (1) and II (1)), a significant event is suspected as soon as the diagnostic reference value (DRL) for a certain type of examination is exceeded by more than 200% (i.e., by more than a factor of 3). In this case it has to be checked as to whether it is a one-time event or whether the dose values are permanently considerably above the DRL. The decisive factor is whether the mean value from this examination and the previous 19 examinations of the same type is more than 100% above the DRL (i.e., more than a factor of 2). Only then does the suspicion become a reportable significant event.
Identification and notification of more than threefold of a DRL overruns is a standard feature of most DMS. However, checking whether it is just an 'outlier' or a 'chronically' high DRL exceedance usually has to be done manually (i.e. downloading the data with subsequent sta-tistical analysis). DOSE offers the convenience of performing this check automatically. Fig. 1 shows a scheme of how this algorithm works.
These rules are stored accordingly in the notification centre (Fig. 2). The percentages (300% or 200%, i.e., factor 3 or 2) apply to the respective DRL. In the standard configuration, with which DOSE is delivered, the DRLs are entered in the limit value tables of the study groups in the column 'upper ACC'. The ‘upper acceptable’ limit, refers to the legal DRL.
How the events identified by the DOSE algorithm are colour-coded, with which message text they are provided and whether they are sent by email, can be configured individually ac-cording to user preferences (Fig. 3). Due to the additional criterion (mean value of this ex-amination and the 19 previous examinations of the same type > 2xDRL), actual reportable significant events are extremely rare and almost impossible in the CT application area. On the other hand, individual cases with more than triple DRL exceedance, occur occasionally to more frequently, depending on how far the device settings are from the DRL. The causes are mostly user or device-related errors. Cases of this kind should always be investigated and the error rectified.
Therefore, it makes sense to mark all exceedances identified with the algorithm in red and to get automatically notified, e.g. by e-mail. For each case, several warnings can be gener-ated at the same time (e.g. "200% DRL exceedance" plus "100% DRL exceedance of the mean value" plus "Significant occurrence Annex 14", for CT and for 2 dose parameters (CTDI, DLP) results in up to 3 x 2 = 6 warnings). The decisive factor for the obligation to re-port is whether there is a message underneath that textually indicates the fulfilment of both test criteria (e.g. 'Significant incident Annex 14').
The default configuration with which DOSE is delivered results in all studies with 'normal' DRL exceedances (i.e. by less than a factor of 3) being colour-coded red. Studies with 'nor-mal' and 'serious' DRL exceedances are thus visually indistinguishable (Fig. 4). However, the DRLs are not limit values, but rather 'serious guideline speeds'. Exceedances may occur, but they should not be the rule and should be limited to justifiable situations (e.g., obesity, ques-tion-related magnification of the examination region, foreign bodies in the exposed area etc.). In the case of devices with dose settings for the standard patient just below the DRL, there will be an apparent 'red alert' in more than 50% of all studies, which actually applies to only a fraction of the studies.
Fortunately, however, in DOSE there is the possibility of a second, lower threshold in the lim-it value tables, with the 'upper ACH' (= 'achievable') in addition to the 'upper ACC' (= 'ac-ceptable'). This is actually intended for in-house guideline values below the official DRL. Studies with dose values that lie between the ACH and the ACC are colour-coded orange. Therefore, the idea is to enter the DRL in the column 'upper ACH' instead and to use the col-umn 'upper ACC' for the triple DRL (see Fig. 5).
Studies with 'normal' DRL exceedances are then colour-coded orange and those with 'seri-ous' DRL exceedances are colour-coded red. As studies with dose values below the DRL are marked in green, a colour distinction is made, which is known as the traffic light system (see Fig. 6).
If this suggested configuration is followed, the advanced monitoring algorithm (Fig.1), which checks the exceedances of the 'upper ACC' field, is in a way 'levered out', provided that the other settings are left untouched. Since the column 'upper ACC' already contains three times the DRL, a significant event suspicion would only be identified if this value were exceeded by more than a factor of three, i.e., the algorithm would be activated and send a corresponding message only when the DRL is exceeded by a factor of nine.
Thanks to the high flexibility that characterizes DOSE by Qaelum, this dilemma can be solved with simple means. For this, only the criteria '300%' for the single deviation and '200%' for the group deviation, which apply to the originally intended DRL entries in the column 'upper ACC', must be adapted. Since the triple DRL is now written there (Fig. 5), the deviations are to be reduced by a factor of 3 accordingly, i.e., '300%' becomes '100%' and '200%' becomes '66%', as shown in Fig. 7. These changes must be made in the notification centre for each dose parameter used there.
In this way, the desired goal - intuitive distinguishability depending on the severity of the DRL exceedances as with the traffic light system - can be achieved without having to forego the ease of the advanced monitoring algorithm as to whether serious exceedances are simply a singular 'outlier'.
Overall, DOSE offers a high degree of flexibility to design the monitoring of significant events described in StrlSchV Annex 14 I+II (1) according to one's own ideas. However, the identification of such cases is only the first step in the processing of significant incidents. How DOSE assists in the application of the other criteria and in particular in the identifica-tion of the causes will be explained in future blogposts.
Stay tuned for some educational weeks!
1. Radiation Protection Ordinance, Federal Ministry for the Environment, Nature Con-servation and Nuclear Safety, Germany. https://www.bmu.de/en/law/radiation-protection-ordinance
|Dr. Nagel worked for many years as a clinical scientist for one of the leading medical imaging manufacturers. In 2009, he founded his own company focusing on MPE services in the application areas of CT and XA.|