Tracing from units with false-positive detections

Category:bug report
Priority:not yet set
Status:ready to commit

According to our model specification (version 1.2.1):

If the test result [of a diagnostic test] is positive, the unit is considered to have been "detected" as diseased (see Section A5). Detection either by clinical signs or diagnostic testing may trigger subsequent control actions, including further traces.

In this way, NAADSM 3.2 and higher can produce false-positive detections, which should be treated like true-positive detections.

It is not possible, however, to conduct tracing from a false-positive unit. The model only simulates contacts from truly infected units. Since the model has not simulated any contacts from uninfected units which come up as false positives, no tracing can occur.

It has already been recognized that the amount of effort expended to carry out tracing is underestimated by the model, due to situations like the one described above. (Another situation that involves underestimating of tracing effort is tracing back to uninfected units. The same reasoning applies: no contacts were recorded in the model from uninfected units that would have to be traced back to in reality, so the model cannot simulate this activity.) Should these limitations be more explicitly documented?




Status:active» ready to commit