Sunday, 27 September 2015

Corona counts continue to clarify...

From Mackay & Arden [1]
Click on image to enlarge
The Kingdom of Saudi Arabia's (KSA) Ministry of Health (MOH) has, over the past week, made a few changes to its Middle East respiratory syndrome coronavirus report layout. Good changes. Nice additions. And I say bravo. In fact - as we've said in the scientific literature [1] - the KSA data are far better than  those we have seen from other recent emerging infectious outbreaks. In particular, China's efforts for the avian influenza A(H7N9) virus cases over many months or Egypt's communication about its avian influenza A(H5N1) virus outbreak(s), have been relatively simplistic, uninformative and hidden from public view, by comparison.

Since the 17th of September, the reports have been undergoing constant change. These range from the subtle tweaks in colour, the addition of pie charts (I particularly like the strong colours in the latest!), inclusion of a marker to identify an asymptomatic MERS-CoV positive sample (are these now included as a "cases" by the MOH?) and the breakdown of how MERS-CoV positives were likely to have acquired their infection.  

Some things I'd still very much like to see added...

  1. In the 'Deaths' table, a column for the date that the deceased person was first described by the MOH - that is, the reporting date. Currently there is not always enough information to go back and say with conviction which e.g. "60 year old male Saudi from Riyadh" died when there may have been several cases with the same description, listed over the preceding days and weeks (even months as some MERS cases have been hospitalised for a very long time before eventually dying of complications)
  2. The number of samples tested each week, as listed by the Infection Prevention and Control-KSA webpage
  3. Details added retrospectively for those 'found' cases. Those data are clearly in the system somewhere as they were summarised back on the 3rd June 2014 [4] and the 18th September 2014. This would be a nice 'rounding off' of all the data to date.
  4. An explanation for what constitutes a "primary case". Currently, the suggestion is that over a third of cases are primary cases. To my mind, after excluding healthcare acquisition, this implies infection from animals. But only a touch over 7% (91 of over 1200 virus detections) of cases reported publicly have any animal contact described. So what does a primary case" mean to eh KSA MOH?  
  5. A single, regularly updated and downloadable line list of these deidentified detections, each with a unique identifier, for all 1200+ detections (akin to what the World Health Organization did for the South Korea cases...but better). This would be a resource that could be of use to future publications (which should cite the source appropriately), conversations whereby we could all discuss the same case using the same code, for student projects and for those number nerds the world over

Screen captures from the KSA MOH Command and Control Centre (CCC)[2]
Date of report is shown below each screen capture.
Compiled using Adobe Photoshop
Click on image to enlarge
But that little wishlist aside, it is fantastic to see this response from the MOH. I have no problem at all with the formats changing, even if it is day by day by day. I'm just happy to see a responsive data sharing outlook. And the willingness to put these changes up, and leave each version in public view is also a promising sign of openness from the MOH. There is nothing wrong with showing the process.

These website is becoming a great example of how we can and should communicate outbreak data in the future.


  1. The Middle East respiratory syndrome puzzle: A familiar virus, a familiar disease, but some assembly still required
    Mackay IM, Arden KE. J Infect Public Health. 2015 Sep-Oct;8(5):405-8. doi: 10.1016/j.jiph.2015.07.001.