Discrepancy for Spain confirmed cases

[CSSEGISandData]

We are issuing this announcement due to the frequent messages we are receiving regarding our reported case counts for Spain. These messages are focused on one of two issues:

1. The drop in our case count from April 23rd to April 24th
2. The difference between our reported case totals and other data aggregation sites or press releases

Both issues are directly related to the current and historical approach taken by Spain to report cases, and highlight the difficulties with aggregating this data. Spain reports its total cases as an aggregation of the total positive tests reported at provincial health levels. For an unknown period of time, some provinces were reported both PCR+ and Ab+ tests as positive cases. Beginning April 16th, some provinces began reporting the breakdown of how the cases were being identified (either PCR, ELISA, or other). By April 20th, this data was available for all provinces, except for some irregularities with Gallicia. On April 24th, the Spanish government began to only report the PCR positive cases. Using Ab tests to confirm cases for a time series is problematic, as these represent historical infections and cannot be used to identify those that are actively infected with SARS-CoV-2. To stay consistent with other countries, we are only interested in PCR+ tests.

1. In regards to the drop in case count, the lack of the full breakdown of PCR+ vs Ab+ test results will result in there being a drop in the confirmed cases at a single point in time - we simply have the ability to choose which day that will be. We have chosen to have the drop occur between April 23rd and April 24th as this is the first day where we are confident that the test results only represent PCR+ results.

2. Other data aggregation sites and news reports are reporting a case burden ~40,000 cases higher than our own. This discrepancy is due to their inclusion of the Ab+ test results within their case total.

Ultimately, we are doing our best to only include data for cases identified through positive PCR tests. Ab tests demonstrate historical infection but cannot be used to assess temporality of infection. The dashboard and repository are particularly useful for interpreting the spread of the virus in space and time, and including cases that do not represent active infections at that point in time reduce the usefulness of the dataset for that purpose.

CSSE