Outbreak occurring or under investigation
DougManuel opened this issue · 9 comments
Small area testing, such as a school, workplace, or long-term care home may have clinical cases reported or under investigation, but they are unable or do not have information about specific case counts.
A boolean flag for outbreak could be used in these situations.
It seems extreme to have outbreak as an attribute of a measure, because it would be blank/no most of the time. And while I think it could work as an attribute of a site (updated semi-regularly), I think that in that table it's too far for the measure to be useful. I know we talked about this before, but I can't seem to find my notes. Do we think it makes sense to have outbreak be a boolean measure? With the instructions specifying that it means that related case counts may be approximate/incomplete? @DougManuel @jeandavidt - what do you think? Does this match what was discussed?
I think outbreakStart and outbreakEnd could be site and population measures. They would be measures with boolean units, but we would only ever record measures when they are true 🤷
I think that's smart! Could you say more about your rationale for having start and end measures instead of a singular boolean measure?
To me, an outbreak is the span of time during which the cases "flare up" in a location, so I would want to record the entire period in which that flare up occurs. And I just figured that if we don't record the end of an outbreak explicitly, then we would have to create a new entry in measureReport on every single day of the outbreak. That would work too, but it might lead to more ambiguity - what happens if someone forgets to report that the outbreak is still ongoing on one of the days? By having a distinct measure for the beginning and the end, it would become obvious when an outbreak report is malformed (i.e., if it has a beginning and no end, or an end and no beginning)
I know teams that have an outbreak measure in their reports and data. I suggest that we include this measure.
The concept of outbreakStart and outbreakEnd is very much part of public health practice and has important meaning and practice implications. For this reason, I don't worry too much about ambiguity, but there are reporting concerns.
For version 2, I suggest that we include, but try to keep straightforward. I was wondering, for example, whether we want to have an outbreak ID (public health departments identify all give outbreaks an ID), but that and other concepts can be for a later version if there are requests.
What about having outbreak a measure. The values for the measure can be 'start', 'ongoing', and 'end'. Remember, each measure will have at least one date attribute.
We can suggest the use of 'notes' to record outbreak ID and other information.
That seems reasonable to me, we'll just need to add like a progressionCatSet with the start, on-going, and end values. And then capture the other aspects with documentation. Let me know if we're all good with this and I'll make the additions to the dictionary and close this issue.
Makes sense. I think outbreakCat aligns better with the naming approach for category labels - but I'll leave it to you.
As outlined above, we now have an outbreak measureID with the valid categories set to Outbreak start, Outbreak End, and Outbreak On-going to capture this information. Details in the instructions column indicate that this means that the numbers reported at the time of outbreak are to be treated with some caution.