Statistics

Breaking: Wind From Tropical Cyclones Cure Hemorrhoids

Lost count of the number of “studies” that feel—not think—that they must invoke terror to justify their undertaking. Take this peer-reviewed gem from Nature Communications by Robbie M. Parks and a slew of others: “Tropical cyclone exposure is associated with increased hospitalization rates in older adults“.

“The intensity of tropical cyclones is predicted to…” Predicted to. How about some Reality instead of predictions?

Such as the number of actual hurricanes hitting land:

“Say, Briggs, why don’t you point to an official source, like NOAA?”

Because that bureaucracy has so far only counted up to 2004. We wouldn’t want to rush them.

Never mind all that. Let’s focus on this paper, for it is an excellent example of the depressing practice of releasing models fits, and the sad reliance on p-values. Except for the opening, it’s a mostly harmless paper, too. It concludes people are more likely to be injured during and right after a storm than before it, which the authors say is good for hospital planning. Both statements are surely true; neither is a surprise.

They counted Medicare hospitalizations, and only those, for places that had “at least one tropical cyclone during our 16-year study period”. They looked from 0 days (the day storms hit) and up to 7 and not 8 days after, because why not.

Here’s their big finding:

In Fig. 4, we present average relative (percentage) changes in hospitalization rates across the eight examined lag days across the 13 causes in the main analysis, as well as for sub-causes with at least 50,000 hospitalizations during our study period. The sub-causes are linked to the 13 main causes in Supplementary Table 2. Respiratory diseases exhibited the largest average increase in hospitalizations (14.2%; 95% CI: 10.9, 17.9%). We observed the largest decreases in hospitalization rates for cancers (4.4%; 95% CI: 2.9, 5.8%).

All right. Here’s a closer picture of some of those sub-causes:

The vertical dashed line is 0% change: to the right are increases in hospitalizations in the 0-7 days after a cyclone hits, and to the left are DECREASES. In other words, if tropical cyclones can cause increases in hospitalizations, they can also cause decreases.

We’re doing science here. Because it is science, the authors have proved, using wee p-values and a convoluted model, tropical cyclones cure “Anal and rectal conditions” and “Bowel obstructions”. Also “Hemorrhoids.”

Must be those fresh breezes.

Look. If you take any event, cyclone or cheerleading practice dates or whatever, and look before and after that date, and then order hospitalizations by increases and decreases about that event date, a certain order in diseases will always appear. After all, some disease has to come first and some last. Because you have an n of 70 million, you are guaranteed to get wee p-values, or, their equivalent, small confidence intervals.

You will certainly get an order of diseases, from decreases to increases, but this is no proof, no proof whatsoever, that the event you picked caused the decreases and increases. There is no proof the event caused some diseases to grow worse and some better, so that hospitalizations would decrease or increase.

This is obvious, no?

If it isn’t obvious, you are left with bizarre interpretation that tropical cyclones cure hemorrhoids. That cyclones can exacerbate asthma (another result) is plausible, but then we didn’t need a “study” to prove it. But if cyclones can cause asthma, then it must also be true that they alleviate breast cancer. Both “results” are reported here.

Or how about this, a limitation which the authors later acknowledge. During storms people don’t go to the doctors unless it’s an emergency, like for injuries—and we see injuries are on the increases side of the plots. On the other hand, during hurricanes and the like people stay home and wait for the hospitals to clear out to have their hemorrhoids lanced and prostates fingered—and these decreases are also “discovered”.

But you’d get the same kind of results for any date you picked with an n this size. It might not be the same order, but there’d be some order, and you’d have the overwhelming temptation to conclude the event you picked caused the order. That’s what they did for cyclones. If it’s a legal statistical move for that event, it is for any.

Which is backhanded proof (as if we needed more) that probability models can’t prove cause. The authors would have been better off releasing a predictive model saying “Here’s how many vacation days proctologists can take when a storm is expected” and the like.

Add to all this the epidemiologist fallacy: we never know the actual cyclone “exposure” of anybody. It’s all a guess, which the authors admit. But they still claim “Any resulting bias” is in their favor.

Don’t use p-values or confidence intervals. Report predictive models instead.

Subscribe or donate to support this site and its wholly independent host using credit card or PayPal click here

Categories: Statistics

18 replies »

  1. Not everything that counts can be counted, and not everything that can be counted counts.

  2. Reality is BORING. Predictions are fun. That’s why all the predictions. Really, you had to know that.

    Does Preparation H know about this? Plus, this should upset those brave souls that surgically “cure” hemorrhoids, or use electrocution, etc to “cure” them. Maybe they get vacation days, but it seems from your comment of “how many vacation days a protologist can take” that this is as temporary as any OTC treatment so nothing to worry about? And much more brutal.

    Remember when science wasn’t silly?

  3. We’re told here, that during the week after a “tropical cyclone” hits, local hospitals see an up-tick in cyclone-related visits, while at the same time, there’s a down-turn in non-cyclone related hospital visits. Well duh.

    This revelation of the seemingly obvious and predictable, still requires graphs, and so Briggs, in his genius way, concludes that since their graphs show hospitalizations for hemorrhoids go down right after cyclones, the cyclones must have cured them. So funny.

    We’re also warned in this study, about the “health impacts years after tropical cyclone exposure”, because how can they implement “disaster risk reduction measures”, without wild conjecture about possible future cataclysmic events?

    These go together like love and marriage, or horse and carriage, or “The Population Bomb” and mass starvation, or rising CO2 levels and zero ice on the north pole by 2013.

    “There is some limited evidence to suggest that there are measurable long-term impacts on health in the years after a disaster47,48. There are plausible causal links between health outcomes and tropical cyclone exposure for many of the associations here6,28,29,30,31,32,33, but more work needs to be done to identify and formalize these pathways.” Indeed.

    “Identifying and formalizing pathways” to conclusions, admittedly founded on “limited evidence”, nevertheless made “plausible” with graphs and the like, so they can implement “disaster-risk reduction measures”. (With plausible deniability when the whole thing goes south.)

    Right away in their “Introduction”, these authors sing the first “O-Antiphon” of the Useful Idiots Psalmody, which is “Anthropogenic Climate Change”, and the whole reason for the “peer-reviewed study” becomes as clear as CO2, in case we had any doubts. (We didn’t.)

    “Tropical cyclones, such as hurricanes and tropical storms, have a devastating impact on the economy1,2,3, environment4,5, and human health6,7,8. Exposure to such events is an important public health concern and one of the key drivers for seeking disaster risk reduction measures9. The intensity of tropical cyclones is predicted to change due to anthropogenic climate change10,11,12,13,14,15,16,17,18,19. “

    “Anthropomorphic Climate Change” and “Systematic Racism”, may soon officially join “Covid-19 Cases”, as public health emergencies subject to gubernatorial decrees.

    And as “cases” decline, The Lancet Psychiatry Journal comes to the rescue (with another “peer-reviewed study”!) postulating that Covid-19 causes psychiatric disorders in fully a third of its victims, even long after they’ve seemingly recovered:

    “One in three people who overcome COVID-19 suffer from a neurological or psychiatric diagnosis six months on, according to the largest study so far published on the mental toll that long-COVID takes on survivors.

    “The most common conditions were anxiety (17 percent of patients) and mood disorders (14 percent).”

    https://www.sciencealert.com/study-suggests-1-in-3-covid-19-survivors-suffer-mental-or-neurological-problems

    So we see that this upturn in “anxiety and mood disorders” is NOT caused by 24/7 Fear Porn, nor lockdowns, nor closed schools, nor lost jobs and businesses, nor forced isolation and masking, nor the current disastrous administration and its policies, nor the coming financial meltdown.

    Our anxiety and/or “mood disorders” aren’t caused by Mad Scientists and their billionaire donors thinking the planet’s temperature must always be what it was in, say, 1972, nor is our anxiety caused by the prospect of forced “vaccines” and annual “boosters”.

    No, any anxiety and/or “mood disorders” we’re experiencing are the result of brain damage (“six months on”!) from Post-Covid Syndrome, or else justifiable worry about the long-term impacts of health outcomes after tropical cyclone exposure due to anthropomorphic climate change.

    In other words, it’s our own fault.

    We may be depressed, anxious and moody, but we’re no longer living in the dark, thanks to the GASLIGHT from all of these “peer-reviewed studies.”

  4. Briggs, you’re farting around with that funny humorrhoid headline. Reminds me of that butterfly flapping its wings in Texas that cured hemorrhoids in China. Because chaos. And science.

  5. Einstein

    I was curious so I looked up your quote
    https://quoteinvestigator.com/2010/05/26/everything-counts-einstein/

    It’s an interesting article that touches on many of the topics that Briggs blogs about

    It goes through many variations of the quote and suggests your quote more properly ascribed to a Professor of Sociology named William Bruce Cameron.

    Imagine a sociologist being that self aware

    I highly recommend the article

  6. See how JOhnb makes his comment, which is a “correction’ look like a compliment

    That’s what I mean about pretension
    Johynb’s way is the wise way
    Don’t tell the whole truth!

  7. I love rabbit holes and wanted to share

    Cameron sounds like someone Briggs could appreciate … check out this book title (AND what it’s about)

    “The Elements of Statistical Confusion Or: What Does the Mean Mean?”
    Cameron discussed the difficulty of performing appropriate statistical measurements
    “Counting sounds easy until we actually attempt it, and then we quickly discover that often we cannot
    recognize what we ought to count. Numbers are no substitute for clear definitions”

    They acknowledge that Einstein has been credited, but they favor Cameron

  8. Joy

    No worries! … and … Thanks! I get into more trouble when I’m not self aware.
    (I actually wanted to “rename” myself William Bruce Cameron for the comment, that would’ve been wrong)

    I just found it fascinating that there was a Professor of Sociology that seemed to share Briggs’ philosophy of statistics back when Briggs was but a lad. Worth further inquiry

  9. I think Briggs has more in common with a lot more people than he realises if he were only to open his eys

    I’m no fan of sociology, quite the opposite. I’d put a freeze on all sociology degrees from today if I were in charge.

    Just because a sociologist says something it doesn’t make it false.

    “As Churchill said, many stumble upon the truth, but most manage to pick themselves up and carry on as if nothing happened. ”

    (It’s not the flavour of truth they like)

    Maybe I’ll check the link now!

  10. The details matter.

    Recently, there was a nine year period (2006-2014) where the continental USA had ZERO landfalling major hurricanes (CAT 3 or higher) and within this a five year period of ZERO landfalling hurricances. Meterologists will dispute my second assertion. In three of those years a CAT 1 Hurricane made landfall after being downgraded to a Tropical Storm. Also, another CAT 2 Hurricane that never made landfall, but a few remote weather stations on barrier islands recorded CAT 1 sustained winds. Intelligent people can debate whether those four should count or not.

    Both these periods included 2012 and the fact Superstorm Sandy was not considered a Hurricane.

    Another tidbit that is kind of fuzzy, but I still think is accurate:
    Is that after the USA finally experienced a CAT 3 Hurricane, the AGW crowd starting crowing about how their models predicted this. And then the climatologists started putting public pressure on the meteorologists to re-classify Sandy as a hurricane. Two reasons, first not having a CAT 3 Hurricane landfall on the USA for so many years was impossible according to their models. And, including the damage caused by Sandy to their “cost to society” for not enacting their draconian job-killing wealth-confiscating policies would be music to the ears of their publicists, i.e. the mainstream media. I am pretty sure I got this nugget from a friend who used to be an AGW true believer, but left the cult when he could no longer ignore the brazen manipulation of data and the sweeping under the rug of model results that contradicted the narrative.

  11. For those who enjoy audio or video, Briggs talk on tropical storms was one, if not the best he ever did. Mr Briggs said it was dull!

    Highly recommend people who are, like me, NOT statisticians, or haven’t heard much of the argument about “climate change”, listen.

    It should be on the telly but the’d never let it past the doors of the BBC or SKY over here.

  12. Breaking Wind From Tropical Cyclones Cure Hemorrhoids

    Hmmm … I would think that could be a cause of hemorrhoids.

  13. My college stats professor used to cite education journals for this type of nonsense. Now it’s found in a Nature publication offshoot. Most memorable was the “study” demonstrating shoe size predicted academic achievement (or intelligence?) in elementary school students…

  14. I used to get in arguments with a physicist because I would say things like “statistics are lies” and he would counter that math doesn’t lie.

    Since I found your site, I don’t even bother talking to him about statistics anymore. It may be that math doesn’t lie, but physicist that use hyperbole to shut down a valid points are sure guilty of something.

Leave a Reply

Your email address will not be published.