Date_confirmation

[sadiekelly]

Consider addition of test result Clinically assessed positive | Lab confirmed | Lab negative | Not tested. These response options are consistent with those stated in the ISARIC Core CRF. Also consider addition of Inconclusive to both G.h schema and ISARIC Core CRF.
Date_confirmation captures date case was confirmed, and method captures the test used for the confirmation, but it could be possible that the test result was negative however this currently cannot be recorded

[sadiekelly]

Relates to the day0 schema variable case_status.
Existing responses for case_status are confirmed | probable | suspected | discarded | omit_error.
Suggest replacing response option 'discarded' with 'negative' to capture negative results clearly.

[lauramerson]

Agree with discarded = negative for clarity. As confirmed/probable/suspect case definitions can vary between outbreaks and location, I support the inclusion of test results as above.

[sadiekelly]

@aimeehan1 @JacqSauer @kelseytoups

[aimeehan1]

@julianasopko

[aimeehan1]

For mpox, we had early discussions around the use of 'Negative' vs 'Discarded' for case status. I think we were originally using Negative and then updated to Discarded to match WHO's interim guidance.

I don't know that I can agree with Discarded = Negative.

'Discarded' can include both a negative test case and a discarded case; however, 'Negative' status would not always capture a discarded case.

Here's an example of the language from WHO:
Discarded case:
A suspected or probable case for which laboratory testing of lesion fluid, skin specimens or crusts by PCR and/or
sequencing is negative for MPXV. Conversely, for example, a retrospectively detected probable case for which lesion
testing can no longer be adequately performed (i.e. after the crusts fall off) would remain classified as a probable
case.

https://iris.who.int/bitstream/handle/10665/354486/WHO-MPX-Surveillance-2022.1-eng.pdf?sequence=1&isAllowed=y

[aimeehan1]

Changing case definitions, and variation by country (or sometimes even within a country!) presented a lot of challenges for recording case status.

Our mpox gap analysis provides a helpful summary of the challenges for case definitions:
https://docs.google.com/document/d/1m9-GFh3xE87CdqBkdjaemDWZNkV2dXsJKOAGGwqLZZs/edit?usp=sharing

Case definitions to describe Case_status [confirmed | probable | suspected | omit_error] were not available at the beginning of the outbreak. Individual countries began to develop their own case definitions - some were published and publicly available, others were not. Case definitions varied by country/ territory, which made assigning Case_status difficult and contributed to a lack of standardized data. Some countries included probable cases in their confirmed counts (i.e. U.S. CDC, Belgium, Australia), while others kept this data separate (i.e. Canada). Some countries, like the U.K., created their own terminology for categories [e.g. possible | probable | highly probable | confirmed]. It was very confusing to keep track of. Due to unstandardized case definitions, the growing number of cases, and our limited resources, we stopped tracking probable and suspected cases for almost all countries/ territories and only focused on confirmed counts. However, we published case definitions where available openly on our Github for transparency.

As another example, later in the outbreak, countries including Italy started segregating their case counts as ‘active’ versus ‘inactive/ recovered’, which added further complication for curators as they had to ensure they were tracking cumulative totals (active + inactive/ recovered) and not just those reported as ‘active’ status. Because of the large number of cases, aggregate presentation of data, and limited team resources, there was no way for us to track Outcome. Case definitions also included categorization of cases based on travel and testing. Little was officially documented around this topic, so we had to create our own guidelines and apply them as consistently as possible. For example, in June, a resident of Sweden was diagnosed with mpox in Denmark; since the case was detected in their healthcare system, Denmark acknowledged the case and some mpox case count trackers included it in total counts for Denmark. Our curation team decided to count it as a Swedish case because that’s where it was officially counted, which meant that our counts for Sweden and Denmark differed from some other trackers (i.e. BNO news + Wikipedia). Similar situations arose involving a case that was a resident of one country but diagnosed in another; curators had to pay attention to where this case was officially counted and discuss as a team how best to log the case to avoid it being double-counted.

[aimeehan1]

For dates, there was also a lot of variation in how cases were presented in open source media.

Date_confirmation captures date case was confirmed. However, confirmation details were rarely provided by source. Therefore, the date of the report was used as the date of confirmation.

[sadiekelly]

@aimeehan1 thanks!
To prevent confusion between a test result (positive or negative or otherwise) and a confirmed case (clinically or lab confirmed), I think we should capture test results separately to case status. The test results would be in the laboratory section of the schema. A test result that is positive would have case_status = confirmed, and a negative result would be case_status = discarded (to keep WHO terminology). Case_status = discarded could include cases discarded for any other reasons not just a negative test.

[sadiekelly]

Regarding dates I think it is important to have clear meanings to the dates captured in the schema so they can be used appropriately in analysis. If some dates within a single variable have a slightly different meaning this might affect calculated transmission rate estimates. All reports would have a report_date. If a case is clinically or laboratory confirmed the date of confirmation can be added. Additionally for lab samples we can collect the date the sample was collected in the laboratory section.

[sadiekelly]

Agreed to capture test result in the laboratory section.

Date of confirmation is a required field and therefore must be populated with either confirmation date or report date.
Date of report may also be used for death date when no explicit date is given.