Details about Statistical Rereads
CAUTION
The statistical reread strategy is only available when the “Absorbance Count” for a given calibration is set to “1” (1 dosimeter per pouch). This is because a method for handling one dosimeter out of a package of two total dosimeters using this strategy has not been determined.
A statistical re-read strategy is typically used to handle a situation where the original dosimeter reading does not conform to an expectation but there is no visible dosimeter defect and we are certain that the measurement procedure was followed correctly.
The Admin wants to configure re-reads for verification that the original measurement is statistically reproducible. This strategy will require the technician to make three (3) measurements of the dosimeter in question and, if desired, allows a statistical evaluation of the results to determine if the reread data supports accepting the original measurement or supports changing the original measurement. You can also choose to include the original measurement in the statistical evaluation.
In this example, we use the statistical reread strategy reason and name it “Dose Confirmation”.
For a statistical re-read, the Admin will need to consider the following:
1 Include original measurement in calculations? By selecting this box, the system will include the original measurement in the evaluation of the average dose of 3 additional rereads taken in succession. Leaving it unselected configures the system to average only the 3 rereads together.
Be careful about including the original measurement. If the original measurement was extremely bad (user zero’d with the dosimeter in the holder prior to measurement, the dosimeter was very dirty, etc.), and this original measurement is included in the statistical evaluation “pass/fail” C.V. for the reread session, the session may fall outside the accepted C.V. because of the original measurement even if the reread measurements were good.
2 Maximum coefficient of variation (C.V.): The coefficient of variation entered here will be used as ‘pass/fail’ criteria for the reread session. The doses of each of the measurements will be averaged and a coefficient of variation for those doses will be calculated and compared against the value we enter here. Enter a CV value here not a CV% value – the system defaults to 0.03 (3%) and it is recommended that the value should lie between 0.003 and 0.03 (0.3% to 3.0%) depending on how aggressive the user wants to be and for what reason.
If the calculated CV of the rereads is lower than the maximum allowed that we have configured, the reread session passes, and the average dose becomes the new dose for that dosimeter ID in the dosimeter report. If it fails the CV test, the reread session fails and the original dose is retained as the measured value for the dosimeter ID, and the reread measurements are rejected.
3 Number of allowed rereads when using the statistical strategy: Your value must allow for 3 rereads in a single session. The “Number of allowed rereads” must be a factor of 3 for a statistical reread. For example, if you want to allow two statistical reread sessions then set the number of allowed rereads to six (6), or in multiples of 3 for any total number of sessions.