Saturday, 12 March 2016

Where's the Zika, Brazil?

These data, which seem to represent the sum total of Zika virus (ZIKV) epidemiology from Brazil in the public space - please correct me if you know of better public sources.

Interestingly, these numbers are not hosted by the Brazil Ministry of Health (MOH) either, but via an excellent Zika virus data visualization page maintained by the Pan American Health Organization (PAHO) regional office of the World Health Organization (WHO).[1,2]

Zika virus suspected and confirmed infections from Brazil (left) and Colombia (right).
Screen capture from [1]. Images combined and edited using Adobe Photoshop.
Click on image for an enlarged version
Things to note from these data...

  • Apart from a little bit of data in week 19 and 27 of 2015, most Brazilian numbers are from just one week - #5 in 2016. 
  • The PAHO report notes that the suspected cases are from media monitoring - not from official numbers. Ouch, Brazil MOH.
  • Keep in mind that the Public Health Emergency of International Concern (PHEIC) was called to work out what, if any, link exists between ZIKV infection and microcephaly.[4] I hope there are more data somewhere otherwise we're starting this investigation using only ZIKV data from well after October 2015, when officials in Brazil publicly raised concerns over a health emergency due to the rising number of microcephaly cases occurring in Brazil.[5]
  • There were - according to the PAHO report - 526 confirmed ZIKV infections in week #6 of 2016. That's nothing to sneeze at and clearly ZIKV is still raging away in Brazil.
  • 27% (526/[1,451+526]) of 1,977 suspected ZIKV infections have been laboratory confirmed in Week 6. About one quarter, meaning about three quarters (73%) have not been confirmed.
  • How many suspected infection samples were discarded because they tested negative? Without a denominator, it's always hard to know whether there is a true rise in infections/cases, or just a rise in sample numbers being tested (thanks @kat_arden). 
  • If we accept the assumption that ZIKV is new to Brazil, then we can at least say that lots of positives means lots of infections where there were none previously. But even that is made difficult by (a) basically no weekly ZIKV data and (b) any way of knowing when ZIKV actually arrived in Brazil vs when testing for it first stumbled across it. 
  • Colombia has much better data coverage - spanning a greater time period. But if we take Week 6 and look closely we see that only 2.8% (108/[3,756+108]) of 3,864 suspected cases have been laboratory confirmed. That means nearly all cases are being diagnosed based on clinical features of rash, fever, joint pain etc. 
  • ZIKVsigns and symptoms are highly non-specific features. If this was the only disease in town that caused these signs and symptoms - great. Perhaps we could extrapolate from clinical diagnoses alone then, but because of the next point - we cannot use clicnial features as a reliable sample to be extrapolating from. I also don't know the case definition being used by Colombia for suspect, probable and confirmed [likely lab positive probables] ZIKV disease. 
  • We know that both Dengue virus and Chikungunya virus are concurrently circulating in Colombia.[3] These can both cause similar signs and symptoms of disease so we can't just say that the way a patient looks and feels means they have this, that or another virus. We need that lab testing data.

For those who wonder why I'm still not "convinced" that ZIKV virus is definitely the main (?sole) cause of an increase in microcephaly cases in Brazil - the data gaps outlined here are just one aspect for my unease. Of course, I can believe what I like, but beliefs are not data. I want more empirical evidence to validate my and other's beliefs, more data to justify those beliefs. I don't want motivated reasoning to be the only driver of associating a virus and a disease.[6] There have without doubt been a few very interesting studies of late, but they each have significant gaps-as noted recently (see the quote card below).

I'm not jumping on the bandwagon with some media - and scientists - who seem hell-bent on singing "I believe Zika virus is definitely, 100%, no doubt at all, the cause". I'll wait a while longer for the studies to bear fruit that is a little more ripe. 

While we were far, far too late to Ebola virus disease in West Africa, we can't make up for those mistakes by creating new ones, and jumping the gun on Zika virus and microcephaly in Brazil. Let's not lose our heads and chase only some of the leads. Or the wrong one. 

References...

  1. http://ais.paho.org/phip/viz/ed_zika_epicurve.asp
  2. http://ais.paho.org/phip/viz/ed_zika_cases.asp
  3. http://www.ins.gov.co/boletin-epidemiologico/Boletn%20Epidemiolgico/2016%20Bolet%C3%ADn%20epidemiol%C3%B3gico%20semana%208.pdf
  4. http://www.who.int/mediacentre/news/statements/2016/1st-emergency-committee-zika/en/
  5. http://www.cdc.gov/mmwr/volumes/65/wr/mm6509e2.htm?s_cid=mm6509e2_w
  6. http://www.scientificamerican.com/article/why-we-should-choose-science-over-beliefs/