She nodded, with the calm authority of a tenured professor, and said: “Mhm, That’s right.”

That got a chuckle out of me. She is two, and does not know what a virus is. What had happened, is she had slid into the most universal human posture there is: appearing to already understand.

We never really grow out of that. And, perhaps, nowhere does it cost us more than in a doctor’s office.

Here’s the uncomfortable truth that you already know about your average medical appointment with your primary care doctor. You get something like twelve minutes.

Into those twelve minutes, you (and they) are supposed to fit your history, your symptoms, your questions, and a plan you’ll follow for the next year. Most of us (myself included) are at risk of walking out having used those minutes badly, not because the doctor failed and not because we’re dumb, but because of a few glitches in how we behave under time pressure and mild social anxiety.

The illusion you don’t know you’re under

Psychologists Leonid Rozenblit and Frank Keil ran a now-famous set of studies on what they called the illusion of explanatory depth. They’d ask people to rate how well they understood something ordinary, a zipper, a flush toilet, how a helicopter flies, and people would rate their understanding fairly high. Then, they’d ask them to actually explain it, step by step. The confidence collapsed.

Over and over, people discovered that what had felt like understanding was really just familiarity. They’d seen the thing a thousand times, so they assumed they understood it. When put to the test, they did not.

Be honest, how well do you think you could explain what happens inside the magic box of the zipper, and how it stays together so uniformly once zipped?

We fall under this illusion of explanatory depth constantly, and we cannot feel it… because that’s the whole nature of it.

In a doctor’s office it looks like this: your doctor says, “Your hemoglobin A1c is creeping up, so we’ll want to keep an eye on that.”

Perhaps you nod. It sounds reasonable. From the context, you intuit that a higher A1c is not good and you are perhaps near a border of some sort. But then, when you get to your car, you might realize that you can’t actually remember what your A1C value actually was, what it should be, what “creeping up” means over what time frame, or what “keep an eye on it” requires you to do, specifically, in the interim.

The fix is simple but hard to put into practice until you get past the second challenge.

Second Challenge: Stop trying to impress your doctor

There can be an almost involuntary pull in the clinic room, a feeling like you should be the kind of patient who has it together. You did some reading. You know a few terms. You don’t want to ask a question that might sound obvious, because some secret (perhaps subconscious) part of you wants the doctor to think: ah, finally, an informed patient who is serious about their health.

This instinct is natural, but is poison for your health.

The most valuable question in any appointment is probably one you’re slightly embarrassed to ask. “Sorry, could you explain that again, I don’t understand the part about ___?”

This is the path to break out of your own illusion of explanatory depth, and it can be uncomfortable.

The same pull to be the “serious, informed one” shows up, dangerously, around lifestyle. When talking to your doctor, it is tempting to sugarcoat the truth about stuff like how much you drink, how often you really exercise (vs. how much you know you should exercise or want to exercise), whether you actually take a medication every day or just most days. And I don’t say this as a criticism of character, because I think it is a natural thing to want to do - to round ourselves up in conversation.

But your doctor makes real decisions based on those inputs, and a flattering input produces a worse plan. My wife is a physician, as are many of our friends, and I can promise you that nobody is grading you on your responses. They just want to be able to provide you with the right plan, which requires accurate numbers (even if they feel like embarrassing numbers), rather than aspirational ones.

What about when you’re not being heard?

Sometimes the problem isn’t you at all. Sometimes, you gather the courage to raise a real concern and it gets waved off, “that’s probably nothing,” and the appointment moves on before you’re satisfied or ready.

You are allowed to pause and slow it down. The thing to say there is specific:

“I hear you that it’s probably fine. I’m still feeling a little unsure about this, and I’d like to understand more about what would change your mind, and what we’d watch for.”

That type of sentence is hard to dismiss, because it asks for the doctor’s actual reasoning and a tripwire, rather than more reassurance. Most good clinicians will engage with it.

And if you consistently feel unheard, after really trying, it is reasonable to find a different doctor. Naturally, you will find that you “click” with some people better than others. That is just as true in your relationship with your doctor as it is with anyone else you meet in your daily life.

Prepare before the visit, or it won’t happen

Another small behavioral truth to remind ourselves of: Psychologists talk about the channel factor: a tiny reduction in friction, made now, that opens a path for a behavior later that would otherwise never happen.

People who intend to do something healthy are far more likely to actually do it if they remove one small obstacle and make the next step concrete.

So, before the appointment, do two concrete things:

First, book it now. Not “soon.” Now, while you’re thinking about it, because the version of you reading this at this very moment is more motivated than the fictional version of you that will “do it soon.”

Second, write down your top questions and bring them on physical paper. A mental intention to “ask about a few things” won’t cut it here. I recommend that you write out the actual questions on actual paper. It is best to prioritize to your top few questions of most importance so that you spend your time on the things that matter most to you.

The appointment is the starting line

This is where a lot of value is won or lost.

In the next few days after your appointment, while it’s fresh, write the plan down in plain language: the target value you discussed with your doctor (e.g., perhaps blood pressure, cholesterol, etc), the timeline, the specific things you’re going to do between now and then, and the date you’ll recheck.

Do not trust your memory here; the illusion of explanatory depth will slowly rewrite your discussion with your doctor into something more vague and suspiciously more convenient.

Then, write down your plan to build the systems that make the plan happen, because having a plan and doing the plan are completely different skills. If your plan includes measuring something daily, attach it to a thing you already do consistently. For example, if you doctor asks you to record your blood pressure daily, instead of saying “I will take my readings in the morning,” try: “I will take my readings in the morning right after my morning coffee.”

If the plan includes a recheck in three months, that appointment should go on the calendar before you leave the parking lot. If the plan includes a hard change, think about how you can design your environment so the easy choice and the right choice are the same one.

A good appointment alone doesn’t change your health. The trajectory you set in the days and weeks and months after the appointment are what affect your health.

-Morgan

Morgan McSweeney, PhD (@dr.noc)

If you’re pregnant (or you know someone who is pregnant) and that unsettles you, here is what is actually in dispute, and what the data say.

~~~

If you are pregnant, you may have just learned that the largest group of OBGYNs in the country has broken apart from the CDC over which vaccines you should get. You are supposed to be able to trust that the official schedule is the correct one. Having two official-sounding sources giving you different answers is the last thing anyone wants in the middle of a pregnancy.

So, I will make the situation as clear as I can. The disagreement is narrower than the headlines suggest, two of the shots aimed at protecting your newborn are not in dispute at all, and the rest is something you can easily talk through in a short conversation with your own doctor.

Here is what happened and what the evidence actually shows.

What happened

This week, the American College of Obstetricians and Gynecologists, which represents most of the OB/GYNs practicing in the United States, did something it has never before done in its history. It published its own immunization schedule for pregnancy, independent of the federal government’s. Thirteen other medical organizations endorsed their decision, including the American Academy of Pediatrics, the American Academy of Family Physicians, and the American College of Nurse-Midwives.

“Changing national recommendations coupled with rampant vaccine misinformation are resulting in confusion for both patients and health care professionals,” ACOG president Dr. Camille Clare said. “It is incredibly important for the public to have access to reliable, evidence-based information on maternal immunizations from a trusted source.”

A medical society does not usually need its own vaccine schedule, because that is the CDC’s job. The reason this one now exists traces back to last year. In the spring of 2025, the Health Secretary, Robert F. Kennedy Jr., a lawyer and longtime critic of vaccines, announced that the CDC would stop recommending the COVID-19 vaccine for healthy pregnant women and healthy children.

To do so, he circumvented the agency’s usual data-driven process, which normally includes a public meeting where the evidence is laid out in the open and outside experts weigh in on benefits and risks. He then dismissed the CDC’s expert vaccine advisory panel and replaced it with members who shared his skepticism. Under the reconstituted committee, the standing recommendations that pregnant people receive flu and COVID shots were both withdrawn.

They have never presented good evidence to support those decisions.

ACOG, long an adviser to that panel, removed itself from the panel after these decisions, saying that the panel’s scientific integrity had been compromised.

ACOG’s own schedule, released yesterday, is the follow-up, an attempt to give pregnant patients a clear, evidence-based answer while the federal guidance is in muck.

The schedule ACOG recommends

ACOG recommends four vaccines in pregnancy.

1. Flu. A flu shot in any trimester, ideally in early fall before the season peaks. Pregnancy changes your immune system and puts extra strain on your heart and lungs, in ways that make catching influenza more dangerous than it would otherwise be, and getting the flu shot lowers the risk for you while also passing some protection to the baby.

A 2013 study showed that vaccination reduced the risk of flu-associated acute respiratory infection during pregnancy by up to half.

A 2018 study showed that getting a flu shot during pregnancy reduced the risk of being hospitalized with flu by an average of 40%.

2. Tdap (tetanus, diphtheria, and pertussis, the last of which is whooping cough). Once each pregnancy, ideally between 27 and 36 weeks. This is the shot that protects your newborn from whooping cough in the first months of life, before they are old enough for their own. Whooping cough can be deadly for a young infant, and the antibodies that the mom makes cross the placenta and go into the baby for them to take with them after birth, to help cover them until they start to generate their own protection from vaccination.

3. RSV. If you will be 32 to 36 weeks pregnant during RSV season (roughly Fall through Spring in the US), ACOG recommends the maternal RSV vaccine. RSV is the number-one cause of infant hospitalization in the country, and the first few months of life are the riskiest, when infants’ airways are smallest. Beyond maternal vaccination, there is a second way to protect the baby: a long-acting antibody shot that is given to the infant after birth. Usually it’s one or the other, rather than both. So in a later pregnancy, or if you miss the maternal-vaccine window, the baby can still be covered directly via the antibody shot. Note that the antibody shot is not technically a vaccine, because it’s not training the infant’s immune system against RSV. Instead, it directly provides the anti-RSV antibodies to give them short-term protection to get them through their first RSV season. Ask your OBGYN which path (maternal vaccine or infant antibody shot) fits your due date and the season.

4. COVID-19. ACOG recommends a COVID vaccine in pregnancy, at any point, and says the earliest opportunity is best. This is one of the two shots the CDC dropped from its pregnancy schedule (flu is the other), and it is where the dispute has been sharpest.

Where the CDC and ACOG differ

Precision matters here, because “the OB-GYNs broke with the CDC” can sound like everything is at odds. It is not. Both schedules agree on two of the vaccines: Tdap and RSV.

The disagreement is the other two: flu and COVID, which the CDC dropped from its pregnancy schedule and ACOG kept.

The flu piece is the more surprising one to me, because a flu shot in pregnancy is among the oldest and most studied recommendations on the whole list, backed by decades of data showing it protects both the mother and the baby without increasing risk for miscarriage or preterm birth.

The change that has drawn the most attention, though, and where the federal “reasoning” is most embarrasingly easy to fact check, is COVID.

So that is the one I want to walk through to give you a sense for why I personally trust ACOG more than the CDC.

What the evidence says about COVID in pregnancy

Two things are well established, across hundreds of thousands of pregnancies, over several years now.

First, COVID is more dangerous when you are pregnant. Pregnancy raises the risk of severe illness from a COVID infection, including the odds of ending up in intensive care, and infection is linked to higher rates of complications like preterm birth. That is the baseline risk the vaccine is meant to reduce.

Second, the vaccine has been studied extensively in pregnancy and found to be safe, with evidence that it lowers the risk of severe COVID for the mother and, through antibodies that cross the placenta, lowers the risk of COVID hospitalization for the baby in their vulnerable early months. This rests on a large and consistent body of work from independent researchers, health systems, and governments (around the world, not just the US).

So why did the federal recommendation change?

Before the announcement, RFK Jr.’s department circulated a document to members of Congress titled the “COVID Recommendation FAQ” laying out the case. The trouble came when fact-checkers and the studies’ own authors looked at the citations that “supported” their claims to Congress.

Several of the research studies simply did not say what the document claimed they did.

One clear example involves miscarriage. RFK Jr.’s document pointed to a large study to claim there were “higher rates of fetal loss” when the COVID vaccine was given before 20 weeks gestational age. That’s not what that study showed at all.

If you just read the title of the study, you might be tricked into thinking they found an association between miscarriage and COVID vaccines.

What that study title really is meant to show, is that it was investigating the topic of miscarriage and SARS-CoV-2 vaccination. The title tells you the topic of the paper, not the conclusion of that investigation.

If take the time to read the study, of more than 246,000 pregnancies in Canada, you quickly find that it actually concluded the opposite. Once the researchers controlled for the variables that are known to affect both miscarriage risk and who tends to get vaccinated, the vaccine showed no association with miscarriage at all.

The study’s lead author, Dr. Maria Velez, a reproductive epidemiologist at McGill University, told Politico that her work had been “misinterpreted” and did not find a link between vaccination and miscarriage.

Independent fact-checkers at FactCheck.org reviewed the full document and found that it had misrepresented that study. Further, they cited another paper for evidnece of “placental blood clotting,” but that other research paper never examined clotting at all!

(I personally believe that this is evidence of RFK Jr.’s team using AI to generate their “research report” that they gave to Congress to justify their vaccine changes. No human bothered to read past the titles of the papers they were citing to justify changes in care for pregnant women. If I was a pregnant woman, that level of casual disregard for my health would make me furious).

To this day, HHS has not backed down; a spokesperson said “the underlying data speaks for itself.”

And.. I guess the underlying data does speak for itself, it just happens to be saying the opposite of what the CDC is saying.

So what should you actually do?

I am not here to tell you what to put in your body. That is a decision for you and the doctor who knows your pregnancy and your history. Some of these choices carry nuance, and reasonable people weigh them differently.

Two of the vaccines, Tdap and RSV, are not in dispute. The two that are in “dispute”, flu and COVID, come down to a question of evidence. The ACOG has decided the CDC is wrong about the best available evidence, and so have the professional associations of OB-GYNs, the pediatricians, the family physicians, and the nurse-midwives, who have all lined up on one side, along with the researchers who actually study these safety and efficacy questions.

The public case for the federal changes, meanwhile, has leaned heavily on misrepresentations of that evidence.

So if you get two different answers, one from your OBGYN and one from the CDC, you should talk it through with the person who is actually responsible for your care. Bring questions to your next prenatal visit. Ask what they recommend for you specifically, and why.

-Morgan

Morgan McSweeney, PhD (@dr.noc

Most of the time, when a headline declares that something prevents dementia, the right response is… a shrug.

Many observational studies have the same hole in them: people who choose to get a vaccine, take a certain supplement, or eat a certain food, are likely to also do other pro-health things like exercise and go to the doctor.

Such healthy habits are known to “cluster” together, and many of them help protect the brain. This is sometimes called the “healthy vaccinee” effect.

That’s why I did a double take at the data from the shingles vaccine. It would appear that shingles vaccination is significantly associated with decreased risk of dementia. The signal is hard to explain away, and the reason is almost funny. It all started with an accidental natural experiment.

A vaccine rationing rule in Wales set up the kind of natural experiment we are not otherwise allowed to run on people (ethics boards take a dim view of randomly assigning someone to the “no-vaccine” group when a vaccine has established net benefits).

Let me walk you through why this association held up when I poked at it, and then where to stay skeptical (because there is plenty of room for that).

An accident in Wales

When Wales rolled out the shingles vaccine in 2013, there were not enough doses to go around, so the health service drew a line by birthday. The program started on September 1, 2013. If you were 79 that day, you were eligible. If you had already turned 80, you were not (and you never would be).

Here is the part that some clever researchers used to make the experiment work. They realized that two large groups of people, born a week apart, just under the cutoff vs. just over, could be the same in many of the other ways that are known to matter for dementia risk… and yet one of them could get the vaccine while the other was locked out, purely because of which side of an arbitrary date they happened to land on during rationing.

That is about as close to a coin flip as you get without running a trial. Researchers at Stanford, led by Pascal Geldsetzer, used this quirk to compare more than 280,000 adults ages 71 to 88, none of them diagnosed with dementia when the program began. About half of the eligible group got the shot, and almost none of the ineligible group could.

First, they checked whether, in their dataset, vaccine receipt decreased shingles incidence. They knew it should, based on prior randomized controlled trials. Indeed, the vaccine cut shingles itself by about 37% in this data, right in line with the prior trials, so the natural experiment appeared to be behaving the way it should, representative of the randomized controlled data.

Importantly, the two groups looked the same on many of the other factors that usually muddy these studies: education, receipt of other vaccinations, even their rates of comorbidities like diabetes, heart disease, and cancer.

However, one place the vaccinated and non-vaccinated groups did diverge was in their rates of dementia.

What they found

Seven years later, by 2020, about one in eight of those adults had been diagnosed with dementia. The ones who got the shingles vaccine were about 20% less likely to be among them, relative to those who did not get the shingles vaccine.

Let’s keep that number in proportion. “20% off” a one-in-eight risk is called the relative risk reduction. The “absolute risk reduction” was about 3.5 percentage points.

That is a dent, not a cure, and plenty of vaccinated people still developed dementia. But still, a 3.5% absolute risk reduction, from a single shot, against a disease where few other interventions move the needle, is the kind of result that is worth celebrating. Geldsetzer’s own reaction, from the Stanford writeup: “It was a really striking finding. This huge protective signal was there, any which way you looked at the data.”

And the interesting bit is that it was not a one-country fluke. The same pattern has since turned up in records from England, Australia, New Zealand, and Canada, which staggered their rollouts in similar ways. One country could be a fluke, but the same pattern across five is hard to dismiss.

There are a few things the headlines almost never mention, though. For example, you have to dig into the primary paper to learn who this result is actually about… because (spoiler!) it did NOT benefit women and men equally.

Here is what the data looked like when all participants were analyzed together. The vertical dotted red line represents the “cutoff” date for receipt of shingles vaccination, and the black lines represent the probability of a dementia diagnosis for people who were too old to be eligible (left side) vs. those who were eligible for the vaccine (right side).

If the vaccine had no effect on dementia odds, you would expect the black lines to be continuous, right across the dotted vertical line. Instead, you see that the dementia risk “jumps” downward right at the age of shingles vaccine eligibility. The magnitude of that reduction in risk of dementia diagnosis is what allowed the researchers to estimate the risk reduction due to vaccine eligibility (or receipt, calculated separately).

However, things got funky when they separated out this analysis by sex:

Read more

Measles, a disease we had all but eliminated, is back in American schools. “I did my own research” has become both a punchline and a battle cry. But the interesting thing is that none of this is very new. Three hundred years ago, the same types of tensions looked like this:

COTTON MATHER, You Dog, Dam You; I’l inoculate you with this, with a Pox to you.

Late on the night of November 14, 1721, someone threw a grenade through the window of Cotton Mather’s house in Boston. The fuse was lit, with a note tied to it threatening his life. By luck, the fuse came loose as the grenade crashed through and hit the floor, and the bomb never went off.

What does a man have to be doing, in colonial Boston, to get a grenade thrown through his bedroom window? He was a clergyman. His offense this time around, which nearly got him killed, was promoting an untested medical procedure.

The procedure in question was inoculation, a crude precursor to vaccination. In 1721 it sounded completely insane… partly because it kind of was. Smallpox was tearing through Boston. The town held about eleven thousand people, and more than half of them would eventually catch it. The basis of inoculation was to take pus from the pox of an infected person and work it into a small cut on a healthy person to deliberately give them a mild case so they would have better odds of surviving the real thing later… hopefully…

Imagine hearing that proposal, with no concept of what a virus even is, no concept of a vaccine. And this is coming from a local reverend. Of course people were horrified. I would have had some serious questions, too.

This usually gets told as a triumph, but I am not exactly celebrating what Mather and the local physician Zabdiel Boylston did. Deliberately giving smallpox to hundreds of people, with no data and no real idea whether it would save them or kill them, is outlandish. In many ways, they got lucky that the benefits outweighed the risks. It could have broken the other way, and then they would be remembered as the men who poured fuel on an epidemic.

Here’s a detail that is too-often left out. The inoculation technique came to Mather from Onesimus, an enslaved African man in his household, who described how it had been done for generations in Africa, and how he had received the same procedure - he showed Mather the old scar on his arm. The practice was ancient across parts of Africa, the Ottoman world, and elsewhere. Mather took a piece of what would turn out to be life-saving knowledge seriously from Onesimus - a man his society was built to ignore. Mather had not gone looking for this and made some grand discovery. Boston finally caught up in 2016, naming Onesimus one of the “Best Bostonians of All Time.”

(By the way, for what it’s worth, this is the same Cotton Mather whose writings had helped feed the Salem Witch panic decades prior...)

When Mather and a physician named Zabdiel Boylston started actually inoculating people, Boston came apart. Other doctors denounced them. The newspaper run by James Franklin, (Ben’s older brother), went after them in print. They were among the most hated men in the city. Then came the grenade that failed to explode.

In the fire of the smallpox outbreak, they never did convince the city that their approach was a good idea. And, to be completely clear, they didn’t have any data to show that it was. Opponents suggested that intentionally inoculating healthy individuals may actually hasten the spread of smallpox through the city. However, Boylston kept records. He wrote down every single person he inoculated and exactly what happened to them.

Those records proved critical. Boylston inoculated around 287 people. Six died, and even that number flatters the critics, because it appears that several of those people who died had actually already caught smallpox (the ordinary way) before inoculation. Call it a 2% death rate (6/287).

Out in the city, meanwhile, smallpox was killing roughly 14% (1 in 7) of everyone it touched. By the time the epidemic burned through the population, 844 Bostonians were dead. Considering the city started out with 11,000, that is a devastating toll.

As it turned out, the people who took the terrifying, unknown risk of inoculation turned out to be about six times less likely to die.

But they got lucky. Nobody knew, going in, which way it would break.

I think it’s easy to take the wrong lesson from all this. Sometimes it is told as “the smart, brave people were right and the mob was a pack of anti-science fools.” But the mob wasn’t stupid and inoculation really did kill some people. Being told to deliberately infect your own child with smallpox is an absurd suggestion, in the absence of data.

The important thing is to note how the truth eventually came out. It came down to three things (and a better insult was not one of them). The first was intellectual bravery: on the part of Onesimus, for proposing the exact solution that he had seen used back home, despite knowing that he was not exactly in a position of scientific authority as an enslaved man in 1721 Boston. On Mather’s part, open mindedness, for taking what must have sounded like a wild proposal seriously. The last was patience and diligence on the part of Boylston: keeping careful count of outcomes among inoculated patients to provide hard data after things settled down.

I think about this when someone says something about their health that I feel sure is wrong. Someone who doesn’t want a particular vaccine. The friend halfway down a rabbit hole. To this day, my first instinct is to “win.” To line up the facts that should end the conversation. Although that is my first instinct, that is never what I do, because I have learned that type of combative approach is just a slower grenade. Even if your intentions are positive - trying to look out for their health - people feel it coming through the window, they duck, and they ultimately come to trust you a little less than they did before.

Here’s what I have found does work (it’s slow and a little maddening until you’ve done it enough times to trust the process). Start by asking questions and actually listening to the answers. Underneath almost every bad health claim is a fear that makes sense: a side effect they watched a family member go through, or a doctor who brushed off their questions and left them feeling like nobody was being straight with them. That part is real, and saying so (out loud) keeps the conversation alive and healthy. Resist the urge to pile on every fact you’ve got. One good question often lands harder than ten carefully sourced counterarguments. So gently ask what would actually change their mind, and answer the same for yourself, out loud: tell them where you’re not certain and what would move you on the topic. And then let it go. You are not going to win it at the dinner table, or in any single conversation, and trying to will only make them dig in.

Give them one specific thing they can check for themselves later, when no one is watching and there is no face to lose, and make sure they know they can come back to you without having to admit they were wrong.

Have patience, curiosity, and a goal of understanding, not “winning”.

-Morgan

That is the story you have probably seen cross your headlines. It is accurate, and the WHO has made a point of saying that there is no general public health threat and no need for travel restrictions. I agree with that.

But the most important outstanding factor in my view depends on a question that is being decided in a sequencing lab right now.

The question is which hantavirus this is.

Hantaviruses are a family, not a virus

When you read “hantavirus” in a news story, it is easy to picture a single pathogen the way you might picture SARS-CoV-2 or measles. That is not how hantaviruses work. The name refers to a whole family of related viruses, each carried by a particular rodent species, each living in a specific geographic range, and each behaving differently inside human beings. The hantaviruses that circulate in mice in the American Southwest are not the same as the ones that circulate in Asia or in South America. The clinical syndromes they cause are not the same. And the rules of how they spread are not the same.

The species that matter most for understanding the cruise ship are the New World hantaviruses, which can cause hantavirus pulmonary syndrome. This is a respiratory illness in which the small blood vessels in the lungs become leaky, fluid pools in the air spaces, and oxygen exchange fails. There is no specific antiviral treatment. Care is supportive, and even with intensive care, somewhere around a third of patients with the pulmonary form of the disease do not survive. That mortality figure, which the CDC has reported for years, is the reason a confirmed hantavirus case in any setting commands serious attention.

Almost every documented hantavirus case follows the same basic pattern. A person inhales aerosolized particles from the dried urine, droppings, or saliva of an infected rodent (e.g., while cleaning a barn, a cabin, an attic, or a basement that mice have been living in). The virus does not pass between people. You catch it from the rodent reservoir.

However, there is one exception, and it is the reason this cruise ship story has the attention of every infectious disease specialist who has seen the headlines.

The Andes virus exception

Andes virus is a type of hantavirus that circulates in a particular rodent (primarily the the long-tailed pygmy rice rat) in southern Argentina and Chile, in the same general region where the MV Hondius began its voyage. It is the only hantavirus for which person-to-person transmission has been clearly documented in the medical literature. The documented chains are small and the events are uncommon, but they are real, and they are the reason South American public health agencies treat suspected Andes virus cases differently than the rest of the hantavirus family.

So when a ship that boarded passengers in Ushuaia arrives in port six weeks later with multiple sick passengers and three deaths, two very different scenarios are possible.

In the first scenario, the ship picked up rodents somewhere in its supply chain or in port, the rodents were carrying a hantavirus, and several passengers were independently exposed to rodent contamination on the ship. This would be unusual but quite believable.

In the second scenario, one passenger picked up Andes virus before boarding or in an early port stop, and the virus then moved between people on the ship. This would be a more unusual and more concerning story. It would be one of the largest documented person-to-person hantavirus clusters on record, and it would have some small impacts on how cruise medicine, expedition travel, and outbreak surveillance are conducted in the parts of the world where Andes virus circulates.

The WHO statement specifically mentioned that sequencing of the virus is ongoing. That sequencing is what will tell whether it’s Andes or another hantavirus. Andes virus has a recognizable genomic signature, so it’s just a matter of time before those tests are complete.

Any headlines you read that compare this to the “early days” of the COVID-19 pandemic are vastly overblowing the situation. This won’t send us into a pandemic.

A few caveats

As of this writing, only one of the six affected individuals has had laboratory confirmation of hantavirus. The others are suspected based on clinical presentation and the epidemiological link, which is reasonable but of course not the same thing as confirmed. It is also possible that the final picture will include a coincidental second pathogen (responsible for the other deaths/illnesses), or a misclassified case, or revised details about the ship environment that change the interpretation.

It is also worth saying that hantaviruses are not new and they are not rare in absolute terms. There are 100k+ hantavirus infections identified globally each year, mostly in Asia and mostly with the renal rather than pulmonary syndrome. The cruise ship cluster in the news right now is unusual because of the setting and the geography, not because the virus itself is new.

What to do with this

For almost everyone (if not everyone) reading this, the practical takeaway is small. There is no reason to change your travel plans, and there is nothing you should be doing differently today because of the situation on a cruise ship far away. Of course, if you ever find yourself cleaning out a shed, an attic, or a cabin that has had mice or other rodents in it, you should probably wet the area down with a bleach solution before you sweep, wear gloves, and try to avoid stirring up and breathing in that dust. But that’s been true since long before this week.

When you read the next round of stories about this outbreak, the line to look for is whether sequencing has identified what type of hantavirus is in play. If the answer is Andes virus, the story will deepen and the conversation around expedition travel medicine may develop. If the answer is one of the other hantaviruses, it is a contained, contaminated-vessel investigation that will be worked through without much further public attention. The tricky part is that the headlines will use the word hantavirus either way, even though those are two very different trajectories.

Best,
Morgan

Morgan McSweeney, PhD (@dr.noc)

Two numbers are important to understand here. Colorectal cancer incidence has been rising about 3% per year in adults aged 20 to 49, while it is dropping 2.5% per year in adults 65 and older, according to the same American Cancer Society report. The same disease, in the same country, in the same year, moving in opposite directions in two age bands. Hm.

The second number tracks the shift over time. In 2022, 22% of all colorectal cancers diagnosed in the United States were in people younger than 55, double the 11% share in 1995, even though that age group has shrunk as a fraction of the US population. Trends look similar in men and women, with the steepest rises in both sexes occurring in the rectum and the distal colon, the segments closest to the rectum.

What has changed?

Last spring, an international team led by Marcos Díaz-Gay at the Spanish National Cancer Research Center and Ludmil Alexandrov at UC San Diego published an important finding in Nature. They sequenced 981 colorectal cancer genomes from patients in 11 countries and looked for chemical fingerprints, the specific patterns of DNA damage that different toxins and biological processes leave behind in tumor cells.

Two of those fingerprints, called SBS88 and ID18, are produced by a substance called colibactin, which is made by certain strains of Escherichia coli already living in many people’s colons. Those colibactin fingerprints were 3.3 times more common in cancers diagnosed before age 40 than in cancers diagnosed after 70.

The signatures were imprinted early in tumor development. Earlier work had already dated colibactin-driven mutations to childhood, with the strongest evidence pointing to the first decade of life. Which means the cells that eventually became these cancers may have started accumulating their first hits while their owners were still in elementary school. The same study found that colibactin signatures were tightly linked to mutations in APC, the gene whose job is to STOP colon cells from dividing uncontrollably. ID18 alone accounted for roughly a quarter of APC driver insertions and deletions in colibactin-positive tumors. The toxin appears to hit one of the master switches in colon biology before its target has graduated from high school.

Now I must be boring for one moment. The above links are suggestive, but causality has not yet been established. Most people with colibactin-producing E. coli in their gut as children will never develop colorectal cancer, and the global signature data is correlation across populations. The hypothesis fits the geography, with colibactin signatures heaviest in countries where early-onset cancer is rising fastest. It fits the timing, with rates climbing in birth cohorts born after roughly 1950, when antibiotic exposure during early childhood, formula feeding patterns, and ultraprocessed-food intake during the first years of life all changed substantially. But none of that is proof. The hypothesis is still being investigated.

Our daughter is a toddler, and my wife Zina sees children every week in clinic. Reading this paper changed how I think about the microbiome during the first few years of life. I’ve been paying closer attention to choices that I already knew were important.

For an adult reading this now, the practical news is mostly good. The biology that colibactin exploits is all the same biology that lifestyle choices in adulthood also can affect. A baseline that was set by your microbiome in childhood is not a fatalistic sentence. The boring levers in front of you still work to reduce risk.

Which brings us to the part that I think most health newsletters write badly. The standard prevention list (eat more dietary fiber, drink less alcohol, exercise more, lose weight if indicated), lands as generic noise because it’s not often tied to why these factors matter for this cancer specifically.

Each item below acts on the same biology described above. They are independent levers on the same machinery, and the hierarchy comes from the World Cancer Research Fund and American Institute for Cancer Research’s 2018 expert report, which ranks the evidence as convincing, probable, or limited based on a global systematic review. Where I cite specific effect sizes, the numbers come from the underlying meta-analysis of 111 cohort studies (Vieira et al., Annals of Oncology, 2017).

Processed and red meats. Each 100 grams per day of red and processed meat combined is associated with a roughly 12% higher colorectal cancer risk (95% Confidence Interval (CI), 4 to 21%). The signal is much stronger and cleaner for processed meat (hot dogs, bacon, deli slices, sausage) than for unprocessed red meat, with the mechanism most likely involving heme iron, nitrites, and the formation of N-nitroso compounds in the gut.

Alcohol is a convincing cause for men and a probable one for women based on current data. Each additional 10 grams of ethanol per day (a bit less than one standard drink) is associated with a 7% higher risk (95% CI, 5 to 9%). The dose-response is steady, with no apparent threshold below which alcohol stops contributing.

Whole grains are a convincing protective factor, in part because of fiber which gets turned into short-chain fatty acids when colon bacteria ferment that fiber. Each 90 grams per day of whole grains, about three servings of brown rice, oats, or whole wheat bread, is associated with a 17% lower risk (95% CI, 11 to 21%).

Excess body fat has also been observed to be linked, both in terms of general adiposity and abdominal adiposity. The mechanism runs through insulin resistance, IGF-1 signaling, and chronic inflammation, all of which feed the same pathways colibactin damage feeds.

Physical activity is a convincing protective factor for colon cancer, with a more uncertain signal for rectal cancer. The estimate is roughly a 20% lower colon cancer risk for the most active adults compared with the least active. Of course, even setting cancer aside, getting more physical activity per week carries some of the largest beneficial effect sizes in all of medicine for cardiovascular disease, type 2 diabetes, and all-cause mortality.

The World Cancer Research Fund estimates that adherence to its full set of recommendations could prevent roughly half of colorectal cancer cases in countries like the United States. That is a population estimate, not a personal guarantee.

What would be some symptoms that could be concerning? A 2024 systematic review and meta-analysis in JAMA Network Open pooled 81 studies and roughly 24.9 million patients to identify the signs that should never be brushed off in someone under 50. Rectal bleeding, including blood streaks on toilet paper or visible blood mixed in stool, was associated with at least a 5-fold higher risk of early-onset colorectal cancer and was present in nearly half of cases at the time of presentation. Abdominal pain showed up in roughly the same proportion. A persistent change in bowel habits, meaning stool that is consistently narrower than usual, more frequent, looser, or harder (over weeks rather than days), showed up in about a quarter. Iron deficiency anemia in someone not menstruating heavily or losing blood elsewhere is its own red flag. The standard worth taking from this paper is simple - any of these symptoms persisting more than two or three weeks, with no clear alternative explanation, deserves escalation. Rectal bleeding in particular gets attributed to hemorrhoids in young adults far too often.

What about the system that catches these cancers? (or fails to?) The same paper found that the average time from symptom presentation to diagnosis was 6.4 months, with a median of 4 months, and that delays in younger patients ran up to 40% longer than in older ones. Some of that gap may be due to factors including feelings of embarrassment, the assumption that 32-year-olds are not likely to get cancer, and the cost of taking time off work.

A practical script for the next appointment, if you have symptoms that have already been dismissed once but have not resolved:

“I have had [symptom] for [duration]. I understand the most likely cause is benign. But I’m still feeling concerned, and I would like to document today whether colorectal cancer has been considered, and if not, what the plan is to rule it out.”

If you have had the above symptoms for a while and feel like you haven’t been taken seriously, this approach is entirely reasonable.

For anyone reading this without a family history of colorectal cancer, average-risk screening in the United States is now recommended to begin at age 45. The standard intervals depend on the test: colonoscopy every 10 years if results are normal, Fecal Immunochemical Test (FIT) every year, and stool DNA tests (such as Cologuard) every 3 years.

The US Preventive Services Task Force lowered its starting age from 50 to 45 in May 2021, and the American Cancer Society had already made that shift in its own guideline in 2018.

If you are 45 or older and have not been screened, the standard advice is unambiguous - it’s time to get checked. The ACS screening guideline is a good one to read after you finish this article.

For anyone with a first-degree relative (parent, sibling, or child) who had colorectal cancer or an advanced precancerous polyp before age 60, or with two first-degree relatives diagnosed at any age, the rule shifts earlier. The American College of Gastroenterology and the US Multi-Society Task Force on Colorectal Cancer recommend beginning screening at age 40, or 10 years younger than the relative’s age at diagnosis, whichever comes first. Most people who qualify for this earlier start do not realize the rule applies to them. If you have a parent who had a polyp removed at colonoscopy, the relevant follow-up question is what type of polyp it was, because the family-history rule is about advanced adenomas and certain serrated lesions, not on every polyp ever removed.

For anyone over 45 who has made the personal decision that they will not get a colonoscopy, an annual fecal immunochemical test (FIT) is a low-friction stool-based alternative. Of course, the tradeoff is that FIT misses some advanced adenomas that colonoscopy would catch. But a stool test that gets done regularly still beats a colonoscopy that does not.

Some of the microbiome dice may have been cast decades before any of us were old enough to vote. But the good news is that our biology still responds to fiber, to movement, to fewer drinks, to less processed meat, etc., maintained over years.

If you feel this article has been helpful, please forward it on to someone important in your life.

Best,
Morgan

Morgan McSweeney, PhD

If you have not encountered it yet, the premise is that some men, most of them not actively trying to have a child, are ordering at-home sperm tests, logging their sperm concentration and motility on apps, and running multi-week training reg…

Read more

If you are trying to decide whether to get a COVID booster next fall, the CDC has a study with a data p…

Read more

The first story is about cancer immunology. A team led by Jing Chen at the University of Chicago screened a library of compounds normally found in human blood, looking for any that might enhance the killing power of a class of immune cells call…

Read more

Read more

But the air inside your home has its own pollutant profile, its own chemistry, and - this is the part that changes the calculation - its own measurable effects on your cardiovascular system, your brain, and your immune system.

This article is about fine particulate matter (PM2.5), a consequential indoor air pollutant for your long-term health.

I’ll walk through what it does to your body, how to measure PM2.5 in your home, and what actually works to bring levels down.

What PM2.5 does to your body

PM2.5 - particulate matter smaller than 2.5 micrometers - are particles small enough to deposit deep in your airways and cause significant inflammation and oxidative stress. The Global Burden of Disease 2021 study attributed 7.8 million deaths worldwide to PM2.5 exposure.

This is an important topic to understand.

Most of those PM2.5-attributed deaths are cardiovascular, not respiratory. A meta-analysis of 42 cohort studies published in the Journal of the American Heart Association (Alexeeff et al., 2021) found that each 10 µg/m³ increase in long-term PM2.5 exposure was associated with a 23% increase in ischemic heart disease mortality and a 24% increase in cerebrovascular (stroke) mortality.

For context, the difference between a well-controlled indoor air profile and a casual one (perhaps where someone cooks regularly on a gas stove without exhaust or burns candles and incense) can easily be 10–20 µg/m³.

So, that “10 µg/m³ increase” in PM2.5 exposure that is measurably associated with cardiovascular risk is not some crazy level of exposure.

The neurological data is also striking. A nationwide U.S. cohort study published in PNAS followed the entire Medicare population from 2000 to 2017 - tens of millions of adults - and found that each interquartile range increase in long-term PM2.5 exposure was associated with a 6–7% increase in incident dementia and a 9% increase in incident Alzheimer’s disease, with no apparent safe threshold. A 2025 meta-analysis of 28 longitudinal cohort studies in Nature Aging identified at least 14% increased dementia risk when comparing people in the 15th and 85th percentile of PM2.5 exposure.

The 2024 Lancet Commission on dementia prevention now lists PM2.5 air pollution as one of 14 modifiable risk factors for dementia - placing it alongside hypertension, smoking, and physical inactivity.

This is a topic that I take seriously for my family’s health, and I think you should be aware of it. This is one of the few environmental exposures I’ve actually changed things in my own home for. And unlike many environmental risks, this is one where your exposure is happening for hours every day, inside a space that you largely control.

Today, we will discuss the interventions that actually work to improve indoor air quality in your home, how to think about filtration and ventilation, and what to do about volatile organic compounds that HEPA filters don’t touch.

Read more

A cruel kind of neurological weather. A storm that appears late in life without warning and sweeps away pieces of a person one memory at a time.

But the science tells a very different story.

In recent years, researchers have started to map the long runway that leads to dementia. Not months. Risk factors that often stretch over thirty or forty years. A slow accumulation of small biological changes in the brain and blood vessels that eventually reach a tipping point.

Which means something unsettling, and also strangely hopeful:

The factors that shape that trajectory are far more familiar (and more within your control) than most people realize.

I’ve spent a large amount of time in the primary literature on this question - not as a physician, but as a scientist with a PhD in Immunology and Pharmaceutical Sciences who reads clinical evidence for a living. I don’t diagnose or prescribe.

But I can help translate the data into takeaways that will feel useful for your own health planning purposes. And what the data says right now is more hopeful than may expect.

The Number That Changes the Conversation

In July 2024, the Lancet Commission on dementia prevention - one of the most comprehensive and authoritative bodies working on this problem - published its third major report. The Commission identified 14 modifiable risk factors that, taken together, account for roughly 45% of dementia cases worldwide.

Take a moment to think about that number -45%.

Nearly HALF of all dementia cases are linked to factors that can, in principle, be changed.

This is a large, well-supported estimate. The Commission’s work draws on decades of meta-analyses, randomized trials, and longitudinal cohort studies involving millions of participants across dozens of countries. And the 2024 update actually increased the estimate from their earlier reports - up from 40% in 2020 - after adding two newly identified risk factors (untreated vision loss and high LDL cholesterol) to the existing twelve.

I want to be precise about what this does and does not mean. It does NOT mean that 45% of dementia is fully preventable.

Some of these risk factors are only partially modifiable, not entirely eliminable. Some of the underlying relationships involve confounding variables that are difficult to untangle.

Like always, even if you do everything “perfect” and you have zero modifiable risk factors, that does not guarantee you will avoid dementia - genetics, age, and factors researchers haven’t yet identified all play roles.

But it does mean that our odds are influenced - significantly - by things that are, to some extent, within reach. And many of those things are ones that most people have never been told to think about in the context of cognitive health.

Here are three that I think deserve your attention.

1. Hearing Loss

Hearing loss is tied for first place (with LDL cholesterol) as the largest modifiable risk factor for dementia in midlife. I will explain why.

The Lancet Commission’s meta-analysis estimated that people with untreated hearing loss have a 37% higher risk of developing dementia compared to those without hearing loss. The Commission estimates that hearing loss accounts for 7% of all dementia cases worldwide - substantially more than hypertension, obesity, or diabetes.

The relationship is dose-dependent: every 10-decibel decrease in hearing ability is associated with an additional increase in dementia risk, with studies reporting widely ranging estimates, anywhere from a 4% to 24% increase per 10 dB, depending on the population studied.

The proposed mechanisms are specific and seem plausible. When hearing diminishes, the brain has to work harder to process incoming sound - a constant increase in cognitive load that diverts resources from other functions like memory and executive processing. At the same time, progressive hearing loss may reduce social engagement (it’s exhausting to follow conversations you can only half-hear), and that social withdrawal compounds the problem through a separate pathway I’ll discuss in a moment.

Here’s what makes this actionable: there is now randomized trial evidence suggesting that treating hearing loss may slow cognitive decline in some people. The ACHIEVE trial, published in The Lancet in 2023, randomized 977 older adults with untreated hearing loss to either a hearing intervention (hearing aids plus audiological counseling) or a health education control.

Here’s the nuance: in the full study population, the hearing intervention did not significantly reduce cognitive decline over three years. But, in a prespecified subgroup analysis of participants who had higher baseline risk for cognitive decline - older, more cardiovascular risk factors, lower baseline cognition - the hearing intervention did appear to reduce the rate of cognitive decline by 48% over just three years.

That subgroup result needs to be interpreted carefully. Subgroup analyses are hypothesis-generating, and the overall trial result was negative. But the finding is consistent with the broader observational evidence, and the Lancet Commission concluded that the evidence for hearing aids reducing dementia risk is “consistent and supportive.”

Longer-term follow-up of the ACHIEVE cohort is ongoing.

The practical upshot: if you are over 50 and have not had your hearing tested, you are flying blind on one of the most modifiable risk factors for cognitive decline. Hearing loss is gradual. Many people accommodate it for years before acknowledging it. By the time you think you might have a problem, the problem has been compounding silently.

But here is the problem: telling someone “get your hearing tested” is both good advice and completely insufficient.

Most people who read that sentence will agree that it is a good idea and do nothing right away. The behavioral science on this is clear - the barrier is friction more than it is awareness.

Psychologist Kurt Lewin called it the “channel factor”: a tiny situational detail that opens a channel for behavior that would otherwise get stuck. You can think of it like “making the next step easier.”

If you find your nearest audiology clinic right now and go online to book an appointment and put it on your calendar, you have opened a channel for action. That single act of reducing friction - turning a general intention into a specific commitment - is worth more than a year of meaning to get around to it.

I bring this up here because this pattern - knowing what to do but not having a system for actually doing it - applies to basically every one of the 14 risk factors on this list. In this letter, my goal isn’t just to inform you, but to help close the gap between knowing what you should do and what you do.

2. Depression

This one is especially important for younger readers.

The 2024 Lancet Commission conducted a new meta-analysis on depression and dementia and found that people diagnosed with depression had more than double the rate of developing dementia 10 to 14 years later, compared to people without depression (relative risk 2.25). In absolute terms, dementia remains uncommon even among people with depression - but that relative increase is striking. A separate Danish cohort study of over 246,000 adults with depression found an elevated risk across all ages of onset - with the association present whether depression first appeared in early adulthood, midlife, or later life.

The Commission classified depression as a midlife risk factor for dementia, and they were explicit about why: the relationship between late-life depression and dementia is complicated by the possibility of reverse causation (depression at older ages may, in some cases, be an early symptom of dementia rather than a cause of it). But the midlife evidence is more clear. Depression in your thirties, forties, and fifties is clearly preceding any likely dementia by decades, and the association remains strong even with long follow-up periods.

If you are 35 years old and managing a depressive episode, or if you went through a difficult period in your late twenties and never sought treatment, this data reframes what that experience means for your long-term brain health. Treating depression is not only about how you feel right now. The long-term association with cognitive health gives additional reason to take it seriously.

For readers over 65, the picture is also important but requires more nuance. Late-life depression is massively underdiagnosed in older adults - often dismissed as grief, loneliness, or “just how it is” after retirement or the death of a spouse. The Lancet data suggests that depression at any age increases dementia risk, and treatment should not be deferred at any age.

I want to take a second to be careful here. The research establishes an association, and the Commission considered the evidence strong enough to classify depression as a modifiable risk factor. However, whether treating depression directly reduces dementia risk - as opposed to treating depression being independently good for you (while also being correlated with lower dementia risk) - is a question that trials have not yet fully answered.

The data says: depression and dementia appear to be linked. Treating depression is worth doing for many reasons, and the potential cognitive benefit is one of them.

3. Social Isolation

Social isolation is classified by the Lancet Commission as a late-life risk factor for dementia, accounting for approximately 5% of all cases. A systematic review of longitudinal studies found that people with less frequent social contact had a 57% higher risk of developing dementia - a population-level risk greater than that of physical inactivity, hypertension, diabetes, or obesity.

That comparison bears repeating: at the population level, being socially isolated accounts for more dementia cases than many other conditions that receive far more publicity and attention.

The mechanisms are not fully understood, but the leading hypotheses center on cognitive stimulation. Social interaction is one of the most complex cognitive tasks we perform - it requires language processing, emotional regulation, memory retrieval, perspective-taking, and real-time adaptation, all simultaneously. When that stimulation is reduced, the brain loses a primary source of the kind of demanding, varied engagement that may help maintain cognitive function.

This risk factor cuts across age groups, but it manifests differently. For a 35-year-old, social isolation might look like remote work, a shrinking friend group after having kids, or withdrawal that often accompanies depression (note how these risk factors cluster and reinforce each other). For an 85-year-old, it might look like mobility limitations, or - and here the hearing loss connection returns - avoiding social situations because following conversations has become exhausting.

That theme of linkages among risk factors is critical to understand.

The Risk Factors Don’t Act Alone

In the research, factors like hearing loss, depression, and social isolation are treated as separate estimates because that’s how the evidence is organized. But that’s not how they always work in real life.

Consider a common sequence. You’re 68. Your hearing has been slipping for a few years, but not enough to do anything about - you just turn up the TV and ask people to repeat themselves. Dinner parties become tiring. You start declining invitations, not because you don’t want to go, but because the effort of following conversation in a noisy room leaves you feeling drained. Your social world starts to subtly contract. You spend more evenings alone. Over months, the isolation compounds. Your mood drops. You lose motivation for the things that used to get you out of the house. Now you’re less likely to book that audiology appointment, less likely to call a friend, less likely to get out and exercise. Each factor feeds the next, and the whole system drifts in the wrong direction - not because of any single dramatic event, but because of a slow, invisible accumulation of headwinds.

The Lancet Commission identifies 14 risk factors and presents them individually. But the real insight is in how they cluster and reinforce one another. Vascular factors (we will discuss these below) group together - hypertension, high cholesterol, diabetes, obesity, and physical inactivity are rarely isolated problems. Depression reduces the self-care behaviors that manage every other factor on the list. Smoking and excessive alcohol increase vascular risk while also correlating with depression and social patterns. Education shapes occupational complexity, which shapes cognitive reserve.

This is why addressing even one factor can have outsized effects.

When you treat hearing loss, you don’t just fix hearing - you remove a barrier to social engagement, which protects against isolation, which buffers against depression.

When you manage blood pressure, you don’t just reduce stroke risk - you protect the small cerebral vessels that supply the brain structures responsible for memory and executive function. The factors are interconnected, and the interventions are, too.

The paid section walks through all 14 factors with this “systems” view in mind - not just what each factor does in isolation, but how it interacts with the others and where a single intervention can interrupt multiple pathways at once.

A Quick Self-Check

Before you read further, try this. Answer honestly - not for anyone else, just for yourself.

Have you had your hearing formally tested in the last three years? Not a screening questionnaire - a pure-tone audiometry test administered by an audiologist. If the answer is no, or if people close to you have commented on your hearing even though you think it’s fine, it may be time.

Have you experienced persistent low mood, loss of interest, or withdrawal from activities you used to enjoy - lasting more than a few weeks - at any point in the last few years? If so, have you been evaluated?

Has your social world contracted significantly in the last five years? Through retirement, loss of a partner, friends moving, reduced mobility?

If any of them gave you a moment of pause, that’s useful information. These cover three of the fourteen dimensions. The full personal risk assessment in the paid section covers the other eleven of fourteen factors - sensory health, vascular and metabolic health, mental health, social engagement, lifestyle, and environmental exposure - so you can see the complete picture and identify which clusters may apply to you specifically.

The patterns that emerge when you see all fourteen together are where the real benefits pop up, because that’s where you can see which single actions might interrupt multiple risk pathways at once, given your personal circumstances.

Subscribe now

What Comes Next

I’ve spent dozens of hours researching and writing the version of this article that I would want to use to guide my own thinking about my risk for dementia. It covers:

• All 14 dementia risk factors

• Behavioral strategies to help interrupt risk patterns

• A “personal risk” question set to help you think critically about your circumstances

• Suggested questions to bring to your next doctor’s appointment

Every factor gets the same treatment you saw above for hearing loss: the mechanism, the evidence, the action, and the specific behavioral technique that makes the action stick.

I wrote this because I kept finding that the information existed but it was never written out into a plan that felt easy to apply. I found most resources to be either too basic (just a list of risk factors), or too technical (aimed at dementia researchers).

I wanted something organized enough, specific enough, and honest enough about the gap between knowing and doing to actually help change behavior to reduce risk. That’s what’s below - factor by factor, mechanism by mechanism, action by action.

Remember, I am a scientist, not a physician. I hold a PhD in Immunology and Pharmaceutical Sciences, and I do not diagnose or prescribe.

My goal with this letter is to help you show up to your next doctor’s appointment with better questions - and in this case, better questions might mean a meaningfully different future.

The 14 Factors

As we discussed in the free section above, the 2024 Lancet Commission identifies 14 modifiable risk factors that collectively account for roughly 45% of dementia cases worldwide. I covered three of them above - hearing loss, depression, and social isolation. Here’s the full list.

Read more

By Morgan McSweeney, PhD (@dr.noc)

Most people think of GLP-1 drugs as weight loss medications. That’s not wrong. But it’s increasingly incomplete - and the part that’s missing might matter more than the part everyone’s talking about.

In 2023, a large randomized trial found that semaglutide - a drug most people associate with “weight loss” - reduced the r…

Read more

By Morgan McSweeney, PhD

A man I’ll call David - a composite drawn from patterns I’ve seen repeatedly in the literature and in life - turned 61 last spring. He walked three or four mornings a week, kept his weight within a reasonable range, and hadn’t smoked in over 20 years. His doctor had mentioned that his blood pressure was “a touch high.” His LDL cholesterol had been in the 140s for several years. His fasting blood sugar levels had crept up over time, though no one had used the word “diabetes.”

He felt fine. He looked fine. His wife said he seemed healthy.

In January, David had a heart attack while shoveling snow.

His family described it the way families often do: it came out of nowhere.

But it almost certainly didn’t.

The signals were there. They were visible on routine lab work. They were present in blood pressure readings taken in a doctor’s office. They accumulated in the background, year after year, while David - quite reasonably - assumed that “borderline” meant “not yet a problem.”

This article is about those signals - what they mean, how easy it is to dismiss them, and what the long-term data actually shows about the years between “a touch high” and a clinical event.

I’m writing this after getting back from a two-mile jog with our dog, Franklin, which is a sentence that matters more to me than it might seem. I have a two year old daughter, and the calculation is simple: the decisions I make now about my blood pressure, activity, and metabolic health determine how strong, active, and independent I will be when she’s twenty, thirty, forty, and fifty. That’s the timeline I’m optimizing for. And everything in this article comes from that same orientation - not fear of disease, but a commitment to being functional and present for the people who matter most, for as long as I can.

I first understood this in an undergraduate physiology course, watching a professor walk us through the cellular mechanics of atherosclerosis - endothelial damage, lipid infiltration, foam cell formation, the slow layering of plaque inside an arterial wall.

That was the first time it clicked for me that cardiovascular disease isn’t something that happens to you on a single bad day. It builds up, layer by layer, over years and decades. That lecture changed my behavior, and those changes have lasted to this day.

I’m a scientist, not a physician - I hold a PhD in Immunology and Pharmaceutical Sciences, and I don’t diagnose or prescribe. But I can read clinical evidence carefully, and I can help you show up to your next doctor’s appointment with better questions. That is the goal of this letter.

How Cardiovascular Risk Actually Builds

One of the most important - and most misunderstood - things about heart disease is its timeline. Atherosclerosis, the progressive buildup of plaque in arterial walls, does not begin at the moment of a heart attack. It begins decades earlier, sometimes before a person’s thirties, and advances at a pace shaped by a handful of risk factors sustained over time.

The analogy that comes closest is compound interest. A blood pressure of 134/86 does not, on any given Tuesday, feel dangerous. An LDL cholesterol of 145 mg/dL on a Thursday morning does not produce symptoms. A fasting glucose of 105 doesn’t cause pain. But these values, maintained over ten or fifteen or twenty years, exert a cumulative biological cost - arterial stiffening, endothelial damage, inflammatory signaling - that adds up in ways a single snapshot cannot capture.

This is why the American Heart Association and the ACC/AHA guidelines increasingly emphasize lifetime risk, not just current risk.

Here’s a detail that changed how I think about this. Your blood vessels, laid end to end, would stretch roughly 60,000 miles - enough to wrap around the earth more than twice. And the surface area of the inner lining of those vessels, the endothelium, is a single-cell-thick layer.

Every square inch of that lining is exposed to your blood pressure, every heartbeat, every minute of every day. When your blood pressure is “a touch high,” what that means in biological terms is that this enormous, delicate organ is experiencing slightly more mechanical stress than it should be - continuously - for years. That’s the machinery behind the compound interest of cardiovascular risk.

Two Signals That Deserve More Attention Than They Get

Cardiovascular risk shows up first on paper, in numbers that are easy to wave off because they fall near - but not dramatically outside - the normal range. I want to walk through two of the most commonly underestimated patterns now, and we will cover three more further below.

Blood Pressure in the 130s

For years, a systolic blood pressure below 140 was considered acceptable for most adults. That’s probably what you remember learning. In 2017, the ACC/AHA updated their hypertension guidelines to reclassify 130–139/80–89 as Stage 1 hypertension. This change was driven by a large body of evidence showing that cardiovascular event rates begin to rise meaningfully well before the 140 threshold.

The data here is strong: a sustained reduction of just 10 mmHg in systolic blood pressure is associated with approximately a 27% decrease in risk of stroke, 28% decrease in risk of heart failure, and 17% decrease in risk of coronary heart disease. That number comes from a meta-analysis of 123 studies including over 613,000 participants, published in The Lancet.

What makes this signal easy to dismiss is that blood pressure in the 130s produces no symptoms. You feel the same at 133 as you do at 119. But your arteries do not experience those numbers the same way, and over decades, the difference in cumulative vascular stress is substantial.

If you take nothing else from this article, take this: if your systolic blood pressure has been sitting in the 130s and no one has had a serious conversation with you about what that means over a ten- or twenty-year horizon, it is time to have that conversation.

And if you don’t have any idea where your blood pressure numbers currently fall, it’s time to find out. More on how to do that at home, later.

An A1C That’s Higher Than It Used to Be

Your average blood sugar levels are measured with a value called hemoglobin A1C. Since your red blood cells circulate for an average of 2 to 3 months, the test measures the sugar content on your red blood cells to indicate an average of your blood sugar levels over that timeframe. Typically, an A1C of 5.7–6.4% is considered pre-Type 2 diabetes, and 6.5% or higher is Type 2 Diabetes.

Most clinicians won’t flag an A1C below 5.7%, the standard threshold for prediabetes. But, it is worth pausing to note that having an A1C of 5.6% is not the same as an A1C of 5.0% - even though neither number tends to generate an urgent conversation.

The ADA itself describes diabetes risk as “continuous and curvilinear” - meaning it doesn’t switch on at a threshold. Risk builds across a spectrum. A 2024 population-based study of over 608,000 adults without diabetes or cardiovascular disease at baseline, published in the Journal of the American Heart Association, found that even A1C levels in the 5.5–5.9% range were associated with increased cardiovascular hospitalization risk in men - despite being well below the standard diabetes cutoff. The difference in risk in that A1C range was shy of statistical significance in women, but the overall trend in the data was apparent for both sexes at higher A1C levels, with a continuous increase in heart risk as A1C went up.

Here’s where this matters a lot. Consider someone whose A1C was 5.1% at age 45, 5.7% at 52, and 5.9% at 58. At each visit, the number was noted. At 5.7%, someone may have mentioned prediabetes. But at 5.9% - still well below the Type 2 diabetes threshold of 6.5% - it probably still was not treated as a crisis. And that’s true; it’s not a crisis in the immediate sense. The problem isn’t the number at any single visit. Risk grows over time. If that person spent the better part of a decade in the part of the glycemic spectrum where insulin resistance was slowly progressing, then their cardiovascular risk was also measurably elevating and the biological environment was accelerating atherosclerosis. Each additional year spent at 5.8% or 6.2% isn’t standing still. It’s compounding.

What this means in practical terms: if your A1C has moved from the low 5s into the high 5s or past 5.7%, you haven’t necessarily developed a disease. But you have moved to a meaningfully different part of the risk spectrum - and more importantly, the time you spend there matters. Ten years at 5.5% and ten years at 6.0% are not the same thing, even though neither number, by itself, would alarm most doctors on a given visit.

To Do: Ask your doctor what your A1C was three or five years ago, and compare it to your most recent result. If it’s been stable and low, that’s reassuring. If it’s moved meaningfully higher - that’s still not a reason for panic. But I would encourage a specific conversation with your doctor about where you sit on the risk spectrum and what you could or should do about it.

Understanding these signals - blood pressure and A1C - and how they impact your long-term health is useful whether you’re 30 or 70.

But they’re only part of the picture. There are three more patterns that matter just as much - and beyond recognizing the signals, the real question is what to actually do about them.

Subscribe now

Paid subscribers get the rest of today’s article, including:

• The three other signals people routinely underestimate - LDL elevation, sedentary time (even with intermittent exercise), and why family history changes what “borderline” actually means.

• 5 specific changes with the strongest evidence behind them - including one that will challenge something you’ve believed about heart health for decades.

• Four numbers that predict your cardiovascular trajectory. (Most people can’t answer all four.)

• A Doctor Visit Prep checklist - sample questions to ask and a list of what to bring so your next appointment feels more productive.

I wrote this article because I believe that the gap between “probably fine” and actually knowing where you stand is the source for many preventable cardiovascular events.

The information above the paywall helps you understand the problem. Everything below it is about solving it. This information can be the difference between a forgettable checkup with your doctor and a productive conversation that may save your life.

Read more

By Morgan McSweeney, PhD (@dr.noc)

As of February 19, 2026, the Centers for Disease Control and Prevention (CDC) has confirmed 982 measles cases across 26 states since the start of the year. That figure is more than four times the case count at the same point in 2025, a year that ended with over 2,200 confirmed cases and the highest annual total the coun…

Read more

There is a habit of reasoning that can trick you about your health because it feels like common sense.

It’s the same subtle factor that resulted in a surprising increase in lung and prostate cancer in randomized controlled trials of high-dose antioxidant supplements.

I think of it as the direction bias - the habit of extending a true effect beyond its tes…

Read more

Public health only works when evidence, process, and trust move together. When one breaks, the damage can quickly cascade into a crisis.

We are now exactly one year into Robert F. Kennedy Jr.’s tenure as Secretary of Health and Human Services. Decisions that once moved through transparent advisory processes are now issued behind closed doors. Longstandin…

Read more

There is a class of words that show up again and again in health, food, and wellness conversations that perform a very specific job in the background of our thinking.

These words do not explain how something works. I think of them as placeholder words.

“Natural” is the most successful example.

When people encounter the word on a label or in advice, it rare…

Read more

There’s a reason so much modern health advice comes packaged with the same little badge of legitimacy.

“Supported by studies.”

Studies sound responsible. They sound scientific. They sound like the speaker is doing evidence-based medicine rather than internet-based vibes.

And of course, the obvious critique is true: not all studies are created equal. A mouse study, a tiny pilot trial, a correlation in a big cohort, and a well-powered randomized controlled trial can all be called “a study,” even though they deserve different levels of confidence.

If you’re reading this, you already know that. Even if you’ve never put words to the intuition, you recognize that a supplement label that says “clinically studied” without providing further detail may have included 12 college students and a survey.

The more interesting problem occurs even when the methods and study design are high quality and no one is being misleading

Here’s the important part: you don’t need to look up the study to spot the problem we are discussing today.

Most of the time, the giveaway is already in the claim.

This letter is about learning to hear when a claim has grown larger and broader than the evidence. Larger than any reasonable study could generate.

Once you notice that mismatch, you’ll start hearing it everywhere.

Why “supported by studies” works so well

From a marketing perspective, “supported by studies” is brilliant.

It implies that the argument has already been settled somewhere else, by people in lab coats, and that you’re now allowed to skip to: why don’t I give it a try?

It also politely suggests that asking for details would be inelegant, the equivalent of requesting the raw footage after someone says a movie “won an award.”

You will never be able to chase down the citations for every health claim that passes through your feed. So, the phrase signals a mental shortcut: studies exist, therefore the claim is grounded.

Here’s the thing - science is a constraint system. It narrows what you can say, under what conditions, in who, and with what confidence.

“Supported by studies” is frequently used to do the opposite. It widens what a study can be made to imply. That’s convenient if you’re looking to sell a product or an idea, but deeply misleading if you’re looking to act in the best interest of your health.

Translation Creep: where the slippage happens

The pattern I see most often is translation creep.

By that, I mean that a study measured one thing, in one population, over one timeframe.

Then, outside the paper, that result gets extrapolated into a larger promise.

• A change in some specific biomarker broadens into “improves health.”

• A benefit observed in a specific population becomes “everyone should do this.”

• A mechanistic finding becomes “prevents disease.”

Research studies ask narrow questions and get narrow answers. Marketing claims invent much bigger questions and borrow those narrow answers.

A high-quality research study measures one tree very carefully. Marketing then uses that measurement to make claims about the entire forest.

This is why “study quality” is sometimes a distraction

There’s an irony I want to drive home: even high-quality evidence can be misused if it’s being cited to support a different claim than the one it actually tested.

A well-designed randomized trial can be turned into a weak claim if the result is generalized beyond the population studied, beyond the timeframe studied, or beyond the outcome studied.

This is why you can have two people on the internet cite “studies” to support opposite conclusions. They’re not necessarily fighting over whether the cited studies exist. They’re fighting over how far the studies can be stretched.

In the paid section of this letter, I lay out the simplest framework I know for catching this kind of claim drift in real time (without having to pull up the research article…) and for explaining it to other people in one or two sentences.

This is something you can use for yourself while scrolling or mid-conversation, and without any notes or prep.

Last, I present a few practice claims for you to train on (some suspiciously broad, some appropriately narrow).

A few recent comments from paid subscribers, shown as they appear on Substack:

Read more

That’s when doubt starts to creep in.

They’ve been exercising more consistently. They’ve made a real change to how they eat. They’re paying attention to sleep in a way they weren’…

Read more