The European Union standard has set a minimum detection rate of 1.5 times the underlying incidence at second and subsequent screens (The European Commission, 2001).
Some organisations stipulate an absolute minimum detection rate, others a minimum rate proportional to the underlying incidence (i.e., the incidence one would expect if screening were not taking place).
In the UK, the NHS breast screening programme stipulates a minimum detection rate of invasive tumours of 3.6 per 1000 at prevalence (first) screen and 4 per 1000 at incidence (second or subsequent) screen (NHS Breast Screening Review, 2003).
We also suggest that three figures might be stipulated in monitoring of screening programmes: a minimum standard, corresponding to our minimum detection rate, a desirable standard to aim at, corresponding to our expected rate, and a maximum, above which further investigation would take place.
Table 3 shows the underlying incidences per 100 000 based on the estimated rates, and the corresponding expected, minimum and maximum absolute detection rates per 1000, for prevalence and incidence screens.
Finally, we suggest new minimum and maximum standards for detection rates in the UK programme.
Considering that the performance at zero Hamming distance threshold indeed pinpoints the minimum FAR and detection rate value on the receiver operating characteristic curve (ROC), optimizing such point in (15) essentially provides a maximum lower bound for the ROC curve.
We use two well-known metrics in evaluating the performance, namely, equal error rate (EER) and minimum detection cost (MinDCF).
Selected types of OBT samples such as rice, corn, rapeseed, fresh lettuce and pork were analyzed for method validation of recovery rate reproducibility, the minimum detection concentration, and the uncertainty for typical low level environmental sample was evaluated.
