Hurricanes Cause Millions More Deaths Than Reported | Hacker News Reader

Hurricanes Cause Millions More Deaths Than Reported

scitechdaily.com

12 pointsgeox22 comments

"...people are dying earlier than they would have if the storm hadn’t hit their community”

Which is a fundamentally different statistic. I get why they are trying to draw attention to the impact, and it's important, but when you go this far off the deep end you do your cause huge damage. When the "official" number is 10,000 and yours is a wide range of 3.6 to 5.2 MILLION, I will dismiss you out of hand with only a brief review at the methodology.

"The researchers estimate 25% of infant deaths and 15% of deaths among people aged 1 to 44 in the U.S. are related to tropical cyclones."

It's a weird stretch to go from the highly believable long-term economic impact of a storm, to the reasonable uneven burden across socioeconomic groups, to estimating 1 in 4 infants die related to a tropical storm. That word, "related" is doing a lot of heavy lifting.

here's a map that shows the rough distribution of tropical storms: https://geography4u.com/wp-content/uploads/2020/04/Distribut...

somehow a country where the majority of the population does not experience direct impact of tropical storms, still loses more people to them than any other cause?

> “more than all deaths nationwide from motor vehicle accidents, infectious diseases, or battle deaths in wars during the same period.”

The researchers put big numbers on the indirect impact but they were precise about what they meant. But then the journalist (or maybe college PR department?) grabbed the biggest total they could find, turned around, and compared that total-life-course indirect count to acute direct causes of mortality… sigh.

Where are those EA types when we need them, with their fast-and-loose treatment of “quality-adjusted life years”?

"10,000 and yours is a wide range of 3.6 to 5.2 MILLION, I will dismiss you out of hand"

Except the article did NOT say that.

You are taking a miss-reading and spreading your own misinformation.

It said "7,000 to 11,000 excess deaths" Per Storm.

Across ALL storms from 1930 , 3.6-5.2 Million.

For Impact. Maybe to also reach, since you are ok with reaching, what if one of the reasons the South East, AL, Miss, Florida, GA are 3rd world dumps, is because of constant bombardment of storms and it has had an economic impact.

Your correction is incorrect; the gp commenter cited TFA correctly. Perhaps it’s worth a closer read:

> All told, they estimate tropical storms since 1930 have contributed to between 3.6 million and 5.2 million deaths in the U.S. – more than all deaths nationwide from motor vehicle accidents, infectious diseases, or battle deaths in wars during the same period. Official government statistics put the total death toll from these storms at about 10,000 people.

I’d gently suggest that it doesn’t add much to your argument to refer to entire regions of the richest nation in history as “3rd world dumps.” Things are far from evenly-distributed, but the problems facing the Southeast US are mainly pretty different from those facing most of what you’re referring to as the “3rd world.” It’s not really a very reasonable characterization, and it comes off as antagonistic and unsympathetic.

I know.

Have you been to Mississippi?

Perhaps a study indicating that maybe repeat storms is a factor, would help understand why they have a difficult time, and what to solve for. That would help explain why the richest country in the world could have pockets of complete backward destitution.

> Florida > 3rd world

Florida has more than double my state’s GDP. If we counted each state on the global scale then FL would be more than the Netherlands. And that’s ignoring the fact that 3rd world is an inaccurate misused term to begin with.

Not sure having Northern Retiree's bring in capital is a great business plan for other states or countries.

Florida has nice weather and beaches, not some great Silver City on a Hill.

I wish it was more common in public discussion to speak in terms of Quality-adjust life years (QALYs) or other relevant metrics.

https://en.wikipedia.org/wiki/Quality-adjusted_life_year

And micro-morts to compare risk across various activities

When people’s lives are uprooted, they eventually (up to 15 years later, in this study’s counting) die earlier than they would have if their lives hadn’t been uprooted. And people who are barely holding on before the storm do worse than people with the means to just up and start a new life.

That feels like an important thing to understand as a public policy matter, but it’s carrying things so far down the causal chain that saying the bad things “cause deaths” is just silly. And comparing these cases to motor vehicle deaths and direct war fatalities… eesh. Where’s the comparison to total excess mortality amongst all veterans and civilians exposed to war?

(Edit: would it make sense to link to the paper itself? It’s an important idea, and the paper is rather less silly than its corresponding PR piece.

https://www.nature.com/articles/s41586-024-07945-5 )

Title is clickbait; study is somewhat interesting and seems to concern the long-term externalities of hurricanes. But the "millions" pertains to a cumulative total over the past ~90 yrs:

This is a great illustration of why taking about deaths is not really informative, and they should be talking about life-years lost.

Why not billions? If you’re going to lie with statistics, might as well lie big.

"If we include people in their 20's, teen deaths skyrocket!"

You sound like someone that says Chernobyl only killed a few people, the few workers at the site. And ignore the thousands that died later of radiation poisoning, cancer that occurred years later.

People argue about the Chernobyl statistics also, but ask yourself who/why people want to spread uncertainty. To make the risk appear smaller.

that's not what these researchers are saying, and the warped coverage is making worse. If you where a worker at Chernobyl uneffected by the accident, but lost your job, and then couldn't afford dental care, and your child got an abscessed tooth that went septic and they died, this methodology would count it as an "infant mortality caused by nuclear accidents". Their (several levels of) indirect grouping to make an impactful headline is what I object to; which is too bad because the real story here is the more subtle but important and very real uneven and long-term impact of huricanes is under measured.

This would be like saying that millions died a few months before they normally would 30-40 years later because they had higher levels of stress due to the accident. The statistical (let alone causal!) connection in such things is so tenuous, it’s utterly meaningless. Control groups or randomized testing is impossible even in principle. Meanwhile, we can’t even reliably determine if eating eggs increases cholesterol.

"saying that millions died a few months before they normally would 30-40 years later because they had higher levels of stress due to the accident."

Isn't this exactly the type of reasoning they are using? You seem to be using this as an example to refute the study, when that is exactly the point.

I just think the whole "statistics is magic so we should not trust any science and go back to living in trees" trope is getting a bit tired. It is 21st century, because people have learned their statistics from the internet doesn't mean studies are invalid. The whole point is that there are methods to determine if something is " utterly meaningless" beyond a knee jerk reaction.

As long as you can accept that kind of statistical argument for things that you did not agree with a priori.

Well. They didn't consult with me on this study to see if I agreed a priori.

But I am able to understand what they are attempting to quantify. I'm just not in the camp of something being difficult means it is garbage.

It isn't that hard to conceptualize:

-> traumatic event happens,

-> there are follow on ramifications,

-> some of them shorten life spans.

-> And a shortened life span means death.

As others in the thread said. Maybe something like " Quality-adjust life years (QALYs)" would have been more generalizable metric.

I understand the logic it just seems like using that chain of consequences seems to produce weird results i.e. Item A kills 1% of the population a year, item B kill 1% a year, Item C kills 1% a year but somehow only 1% of the population dies every year.

Though QALYs may get at that.

Ok. So kind of like problem with Drake Equation. Depending on made up % for each variable, can have wild swings.

I don't think that is the case here.

The 'science' around counting deaths is older and more fleshed out than you are giving credit. People have been studying the impact of different types of events and co-factors on counting deaths for many decades. This isn't just winging it with some new made up statistics.

Reply on news.ycombinator.com