Hurricane Maria’s Death Toll: A Statistics Story

I spend my working life dealing with data and statistics, and somehow I do not get enough of that to avoid writing and thinking about it in other contexts besides that of work [1].  Today I would like to talk about data as it relates to a very serious but somewhat under-discussed issue, and that is the problem of natural disasters.  Now, I would like to caution at the outset that I believe that a great deal of the blame for natural disasters is overblown [2], and so this piece is not designed to be a cheap hit for political perspectives of the kind that most people are used to reading.  I do not blame former President Bush, for example, for the problems that resulted from Hurricane Katrina, although certainly it was not FEMA’s greatest moments, nor a noble moment for the citizens of New Orleans, many of whom became unwelcome and troublesome refugees in other cities like Houston for some time after the storm.  It would be fatuous to claim that there are no political implications to problems of data and statistics, but I wish to make it plain that I do not wish to turn a discussion of these issues into merely a pretext for scoring cheap political points even if there are undeniable political implications to these matters.

My interest in this matter was piqued by a recent study concerning supposed “excess deaths” in Puerto Rico.  Part of the problem here, though, relates to the lack of credibility of the sources talking about it.  Try as I might, I could not get any sources better than Vox or the New York Times, which are pretty useless sources when it comes to politically charged matters like this one.  I could find no remotely unbiased or fair-minded source, nor one that could give an explanation of why the death count of Hurricane Maria varies so widely.  You would think that making a death count of a hurricane was a straightforward manner, but the fact that Puerto Rico’s infrastructure was overwhelmed by the storm meant that essential services and worthwhile data collection was not done in the aftermath of the storm when there was political wrangling over the response to the storm and the slow regaining of power and rebuilding of infrastructure in the storm’s aftermath.  It should also be noted that the official Puerto Rican statement of the death toll of Hurricane Maria is also currently undergoing a revision, and this suggests that there will be a smaller disparity between the current official death count of 64 and estimates of 4600 or more deaths as a result of the hurricane.

How does one count a death?  Again, we might think this is a straightforward matter, but it is not straightforward when one is using statistical models of excess deaths based on ordinary trends and patterns.  A death because one is dragged out to sea by storm surge and drowned or because one is impaled by projectiles launched by high winds is fairly obviously a death from a hurricane.  But a hurricane can kill indirectly as well as directly.  If someone died because of mold-related issues in the aftermath of the hurricane because of storm damage in the weeks and months after a storm, does that count as a death from the hurricane as well?  The question I see, and it is not one answered by the reports, is how the excess deaths of 4600 were counted, and what grounds were counted.  Again, we are dealing with untrustworthy news sources here and the question of what data was used and how wide of a net was cast in terms of counting a death as belonging to the hurricane, much less determining the blame for such deaths, as blame is inevitably affixed to people in matters like this one, is not an easy question to answer.

At least the foreign sources, if hardly reliable ones themselves, gave some idea of what excess deaths were counted.  “Disruption to health care was a “growing contributor to both morbidity and mortality” in natural disasters, they said, because growing numbers of patients had chronic diseases and used sophisticated equipment that relied on electricity,” wrote the BBC, for example.  What this suggests is that the general poor health of many people in Puerto Rico meant that they were unable to handle the disruptions in electricity that resulted from the storm.  As I would have expected, the storm killed indirectly and Puerto Ricans themselves were part of the problem, a similar issue to what happened in Katrina when the faults of New Orleans being cited as it is and the poor response of the citizens of the city themselves contributed to a great degree to their own suffering.  This suggests that statistics has something to say about our suffering, and that this suffering is complicated.  We suffer not only from the events in our lives, but also if we have a lack of resources to deal with those events.  We suffer because of our own decisions that make us more vulnerable and also because government is not always reliable in providing safety and assistance in times of need (and to rely on government in general is unwise).

Obviously, there are political repercussions for this, but it seems unjust to make this purely a matter of blaming the federal government.  After all, the low death count of 64 comes from the Puerto Rican government itself.  Likewise, Puerto Rico is a rather remote possession of the United States, and is not well-connected to the rest of the infrastructure of the United States.  Had Puerto Rico been a state, it is likely that the Federal government would have done at least a good a job sending generators and trailers to Puerto Rico by ship as the government did in the aftermath of Katrina, a similarly powerful hurricane.  The fact that Puerto Rico is a bankrupt territory instead of a state does make the matter more complicated.  What sort of response will there be to this storm and its aftermath.  Will Puerto Ricans desire even more to be deeply integrated into the United States so they are not isolated, or will this trigger a desire to go it alone despite the fact that they appear not to be able to handle their own business?  Much remains to be seen, and although I have readers in Puerto Rico, I have never gotten comments from them.  This is surely a time when I wish that someone from the area itself would comment on such matters rather than let the issue become part of the general political malaise of our contemporary culture.

[1] See, for example:

https://edgeinducedcohesion.blog/2015/04/09/im-just-here-because-i-love-data/

https://edgeinducedcohesion.blog/2015/04/10/data-humanism/

https://edgeinducedcohesion.blog/2015/08/07/book-review-big-data-work/

https://edgeinducedcohesion.blog/2017/02/01/big-data-small-world/

https://edgeinducedcohesion.blog/2014/10/24/lord-of-the-data/

https://edgeinducedcohesion.blog/2014/09/03/your-database-sucks/

https://edgeinducedcohesion.blog/2018/04/17/we-get-that-all-the-time/

About nathanalbright

I'm a person with diverse interests who loves to read. If you want to know something about me, just ask.
This entry was posted in History, Musings and tagged , , , . Bookmark the permalink.

One Response to Hurricane Maria’s Death Toll: A Statistics Story

  1. Pingback: Maybe We Don’t Want Our Towns To Be Strong | Edge Induced Cohesion

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s