Dead Is Dead

An armed robbery and a long-term dietary pattern involve different kinds of risk. Compare their frequency, consequences, and relevance to the person rather than treating a single meal as equivalent to an attack.

Firearms equipment and training address some threats of violence. Dietary advice and medical care address different risks. Neither substitutes for the other.

Forks over guns, then. Not because violent criminals are harmless, but because the fork, the cigarette, the car, and the hospital kill so many more of us, and for the person they take, and the loved ones they leave, the cause makes zero difference. Dead is dead.

A plane crash is a concentrated event. Dietary exposures accumulate across meals and years. Comparing their toll requires a common population, period, and measure of harm, not the vividness of the scene.

We're living through an epidemic of misallocated fear. Americans spend billions of dollars, millions of hours, and enormous amounts of emotional energy protecting themselves from dangers that are extremely unlikely to harm them, while the real leading causes of preventable death accumulate quietly, largely unaddressed, because they're not the right flavor of scary.

Heart disease kills more than 680,000 Americans a year. The vast majority of it tracks back to diet and lifestyle. Cancer kills more than 600,000, Americans every year, and most of those cases trace to causes we could control. Infections caught inside hospitals, the buildings we enter to get well, kill on the order of 100,000 a year, while other preventable hospital mistakes claim another 300,000 lives (See my Deadly Hospital Mistakes book.) Drug overdoses killed around 80,000 people last year, even after the steepest one-year drop ever recorded. We're a nation of people who buy guns and take firearms classes, just like the ones I teach, but have never asked their doctor if they washed their hands before entering your room.
Growing up, I lost count of the number of times my father said, or often yelled, "Pay Attention!" But not once did he turn that into useful advice, by naming that to which I was supposed to direct my attention.

Personal habits belong in a risk assessment alongside health conditions, exposures, and the systems responsible for safety.

I'm in favor of firearms training. I've spent more than three decades working with firearms as an armorer, stunt coordinator, instructor, and expert witness. That preparation does not address diet, cardiovascular disease, infection, or traffic risk.
I also value Gavin deBecker's The Gift of Fear and its emphasis on attending to warning signs. Immediate warnings and long-term health evidence belong in the same safety plan.

Dead is dead. That's the principle at the center of this book.

Prevention requires asking whether a danger was foreseeable, what was known at the time, and which measures could have changed the outcome. Those questions help organize an investigation, but legal responsibility also requires applying the relevant law to the evidence. Knowledge of a danger does not by itself establish that a particular death could have been prevented.

Attention is limited. Compare immediate concerns with the evidence about age, health, exposure, and local conditions before reallocating it. A falling national rate does not establish that a particular person's risk has fallen.

Mortality counts show the size of recorded categories. Determining what portion is preventable and which measures work requires additional evidence about causes, exposures, and interventions.

Forks Over Guns

Diet deserves attention alongside protection from violent crime. Compare diet-attributable harm with violence using consistent populations, periods, and outcome measures.

Murders account for roughly 16,000 to 19,000 American deaths in a typical recent year. Heart disease, stroke, diabetes, and cancer account for a much larger combined total, but that total cannot all be assigned to diet. A person's dietary risk and ability to change it depend on the person's health, access to food, and circumstances.

Deaths are only part of the harm from violence or chronic disease. A comparison that includes nonfatal injuries needs a defined outcome measure and time period for both. Mortality totals alone cannot establish which produces the greater combined toll.

A sudden attack is easier to picture than a disease process developing over years. That difference can affect attention, but an individual's risk estimate still requires health and exposure information.

Consider the people who plan for violence for a living. A police officer straps on a bulletproof vest at the start of every shift, guarding against a threat that, for almost every officer, never comes. The numbers are lopsided: an officer is roughly 25X more likely to die of heart disease than to be killed by a suspect, and the average officer who has a heart attack has it at forty-six, about two decades younger than the civilian average. The vest guards the chest from the outside while the food eaten inside it, shift after shift of drive-through meals in a parked cruiser, quietly does the killing from within. A department will buy that vest and replace it every five years, and never once screen the heart it's wrapped around. That's the entire argument of this book: we armor ourselves against the danger we can see, and feed the one we can't.

What actually kills Americans

Start with the whole picture. In a recent year the federal government recorded about three million deaths in the United States. Here are the leading causes, in order, drawn from final national mortality data.

Look at the top of that list and then look for the dangers that dominate the news. Homicide isn't in the top ten; it's the short red bar near the bottom. Plane crashes and terrorism are the two red slivers beneath it, a few hundred deaths a year, and a few dozen, set against leading causes in the hundreds of thousands. The dangers that fill our screens and our nightmares are, with few exceptions, statistical footnotes next to heart disease, cancer, and the slow accumulation of chronic illness.

Seen as a share of every death in a year, the gap is starker still. The leading causes together fill almost the entire circle. Murder is a sliver you have to be told is there.

Now hold two numbers side by side at their true proportion. This is the most important image in this book.

Every homicide represents a person and a family. A low national share of deaths does not establish low risk for every individual. Compare national totals with the person's age, circumstances, and exposure.

The evidence on diet isn't fringe

Research on diet and disease includes clinical studies and population studies. Evaluate each finding by its design, population, measurements, and limits. The weight of a body of research does not guarantee a courtroom outcome.

The most useful recent lens is the research on ultra-processed foods;  the industrially formulated products that make up a majority of the calories Americans eat. A 2024 umbrella review in a major medical journal pooled dozens of analyses and found strong evidence linking higher intake of these foods to higher death rates from cardiovascular disease, type 2 diabetes, and dozens of other adverse outcomes. A three-part Lancet series in 2025 reached compatible conclusions. Large cohort studies that follow people for decades find that the more of these foods you eat, the sooner you die.

Assess dietary evidence separately for each outcome. The strength of a cardiovascular finding does not automatically extend to every cancer or other disease. Use the evidence relevant to the condition when discussing changes with a clinician.

The same fork that kills us can also reverse the damage

Diet isn't only the leading cause of the disease that kills the most Americans; for many people it's also the treatment. The clearest demonstration comes from Dr. Caldwell Esselstyn, a surgeon at the Cleveland Clinic who, beginning in 1985, took a small group of patients with advanced coronary artery disease, who were un-helped by conventional care, and put them on a whole-food, plant-based diet with no added oil. Their cholesterol fell sharply, their chest pain eased, and follow-up imaging showed something cardiologists had been taught was nearly impossible: the disease in their arteries stopped advancing, and in several patients it had visibly regressed.

He repeated the approach with a larger group. Of 198 patients with established cardiovascular disease who came to him for counseling, 177 stuck closely to the diet. Among those who adhered, over an average of nearly four years, the rate of further major cardiac events was about half of one percent. Ninety-nine of every hundred avoided another heart attack, stroke, stent, or bypass. Among the group who didn't stick with it, the majority had a new cardiac event. That's not a subtle signal.

These weren't large randomized trials with control groups. The patients volunteered and were highly motivated. A scientist doesn't call this proof that one diet reverses heart disease in everyone. What it is, is a striking clinical result that points in the same direction as a very large body of population research. Epidemiologist Dr. T. Colin Campbell laid out much of that population evidence in The China Study, and the documentary Forks Over Knives brought the whole argument, Esselstyn's and Campbell's together, to a wide audience. Dr. Esselstyn's son Rip, a former firefighter, has spent years showing through his Engine 2 and PLANTSTRONG work that ordinary people can actually LIVE this way, not just study it.

Over enormous numbers, the population data and the direction is consistent, and denied at your own peril: a diet built on whole plants protects the heart, and in many people it can undo damage already done. You don't have to accept every claim made by every advocate to act on that. The strongest, most defensible version of the finding is enough on its own: what's at the tip of your fork is loaded in both directions. It's the largest risk in most lives, and it's also one of the few risk factors you can reload as medicine. As Hippocrates stated, "Let food be thy medicine, and medicine be thy food."

The second killer hiding in plain sight

There's a second danger nearly as quiet as the fork, and it lives inside the buildings we trust most. Infections caught in the course of medical care kill on the order of 100,000 Americans a year. Sepsis, pneumonia, urinary tract infections, bloodstream infections from catheters, surgical site infections: these aren't exotic hazards. They're the predictable consequence of failures in basic hygiene and infection control. Handwashing. Sterile technique. Following the protocol every time, not just when it's convenient.

Ignaz Semmelweis figured out in 1847 that doctors who washed their hands before delivering babies dramatically reduced the number of mothers who died. He was right. He was also ridiculed by the medical establishment of his day and died in an asylum, because his colleagues couldn't accept that their own hands were carrying death from one patient to the next.

We know better now. The science of infection control is long settled. Yet studies consistently find that healthcare workers wash their hands correctly between 40 and 60 percent of the time. (Hand washing failures peaks between 2pm and 4pm) The gap between that compliance rate and the deaths it produces is filled, in large part, by people who came into the front door of a hospital for something entirely unrelated to germs, and left via the morgue in the basement.

Most Americans have never asked a medical provider if they have washed their hands. The question Semmelweis answered 175 years ago could still save your life.

Why we get risk assessment so wrong - it's not your fault

The answer is evolutionary, not stupidity. Human threat perception was calibrated for a world in which the dangers that most needed attention were visible, fast-moving, and socially communicated: the predator, the attacker, the contaminated water in front of you. Our nervous systems are tuned to respond to immediate, concrete, and imaginable threats. (30,000 year old cave paintings illustrate our ancestors drew training scenarios for hunting.) But our brains are dangerously bad at assessing statistical, distributed, slow-acting risks that play out over decades.

An immediate threat and a chronic disease process create different warning conditions. A person can have cardiovascular risk without a dramatic symptom. Seek assessment rather than waiting for one, and do not assume that prevention or treatment is futile after symptoms appear.

Recognizing a hazard is a first step. Risk reduction also requires selecting a suitable control, implementing it, and checking whether it works.

Why We Miss the Real Threats

Investigating an accident includes asking whether anyone noticed a warning, reported it, and had the authority or means to act. Do not assume that someone sensed the danger before every event or that an unspoken concern explains the outcome.

I said earlier that our misreading of danger is evolutionary, not stupidity. You can't correct a bias you don't understand, so here's what's happening under the hood, in our brains. The reason we fear the wrong things is that the brain doing the fearing was tuned for a world that no longer exists, by a process that never cared whether we were accurate: only whether we survived long enough to reproduce.

You have the brain of a hunted animal

Evolutionary accounts of threat perception offer explanations for rapid responses to immediate danger. They do not establish which individual ancestors survived because they reacted quickly or died because they deliberated.

A response to a sudden threat does not measure long-term health risk. Use relevant data to assess conditions such as cardiovascular disease rather than assuming that the intensity of fear tracks the probability of harm.

We're running prehistoric threat-detection software on a modern risk landscape, and the mismatch is killing us.

The smoke detector in your head

There's a specific reason our mental alarm is so easily tripped by the dramatic and so silent about the statistical. Psychiatrist Randolph Nesse called it the "smoke-detector principle." A smoke detector is deliberately set to over-react: it shrieks at burnt toast, because the cost of a false alarm is a few seconds of annoyance, while the cost of staying silent during a real fire is the life of everyone in the house. Evolution wired our fear the same way. Researchers Martie Haselton and David Buss formalized this as error management theory: when a danger is uncertain, the brain is biased toward whichever mistake was less costly to our ancestors. Mistaking the wind for a predator cost you a jolt of adrenaline. Mistaking a predator for the wind cost you everything. So we evolved to flinch at shadows.

This bias is a plus for survival, and a hindrance to accuracy. It means your fear system spends its energy generously on vivid, low-probability threats: the intruder, the attacker, the lurking stranger: and spends almost nothing on the high-probability, faceless ones. A cheeseburger never rustled the bushes. The unwashed hand never looked you in the eye. So the alarm stays quiet for the dangers most likely to take you out.

What's easy to picture feels likely

Kahneman and Tversky described the availability heuristic: an event can seem more likely when examples come readily to mind. A vivid report can affect that estimate. Check the estimate against frequency and exposure data rather than assuming that a memorable event is common.

For the clearest proof, look at which animals we fear, versus which ones actually kill us. Name a dangerous American animal and people say shark, bear, alligator, rattlesnake; things with teeth and venom and a starring role in our nightmares. Add up all deaths from these creatures in a year, and they kill a handful of people; sharks account for about one. The deadliest animal in the country, by a wide margin, is the white-tailed deer, which kills roughly 200 Americans a year, not by attacking anyone but by stepping into the road at dusk. Bambi outkills every shark, bear, and snake in the country combined, but triggers no fear. Our fear tracks the teeth, not the body count.

The dread you can feel and the risk you can't

Researcher Paul Slovic spent decades mapping which risks frighten people, and which don't, and found that the fear tracks a handful of feelings, none of them statistically correlated with actual danger. We dread risks that feel uncontrollable, catastrophic, involuntary, unfamiliar, and unfair, far more than risks that feel controllable, ordinary, and chosen; even when the ordinary ones kill vastly more people. A nuclear accident, a terrorist attack, a plane going down: all rate high on dread, and all are extremely rare. Driving your car, pouring a drink, lighting a cigarette, eating the way you always have: low dread, freely chosen, familiar, and among the deadliest things you do. The feeling of danger and the fact of danger are simply different measurements, and we keep reading the first when we should be reading the second.

When fear is high, we forget the odds

Legal scholar Cass Sunstein described a trap he called "probability neglect." If an outcome is scary enough, people stop weighing how likely it is, and fixate on how terrible it would be. Tell someone there's a tiny chance of a vivid horror and they will often treat it as if it were certain, because emotion crowds out the arithmetic. This is lever the fear economy of the next chapter pulls. Make the outcome dreadful enough and you won't have to lie about the odds; people will stop checking them on their own.

The numbers we can't feel

There's a final, fatal feature of our wiring. We're moved by the individual, and numb to the multitude. Slovic showed that a single identifiable victim, one named child in danger, mobilizes enormous concern, while the same risk multiplied across thousands of anonymous people produces almost no concern. He called it "psychic numbing": the more who die, the less we feel. This is why a single dramatic death becomes a national story and the hundreds of thousands lost each year to poor diet and to faceless infections vanish into statistics. Distributed harm isn't just easy to ignore; our wiring makes it nearly impossible to feel. And we don't act on what we don't feel.

It won't be me

For the slow, personal risks, one more bias does us in: people reliably believe they're less likely than average to suffer the bad outcomes of their own habits. A smoker who knows the statistics still feels that the statistics are about other smokers. This optimism about ourselves is comfortable, and is dead wrong about the outcomes we have the most power to change.

The deaths we deem "deserved"

There's one more reason the quiet killers slip past us, and it's not in the wiring of our brains. It's in the wiring of our morality. We don't grieve all deaths the same way, and we don't fear them the same way either, because we sort them, almost without noticing, into deaths that were done to a person and deaths a person brought on themselves. A plane falls out of the sky, a stranger opens fire, a drunk driver crosses the center line: these feel like violations, harms inflicted from outside, and they summon our fear and our sympathy in full. But forty years of cheeseburgers ending in a heart attack, or three decades of cigarettes ending in a tumor, we file somewhere else, under a quieter and harsher heading. We treat them as the bill coming due, as a choice catching up with the chooser, as, in a word we rarely say out loud, "deserved."

Blaming a person for a health risk can discourage sympathy or prevention. It does not establish that the harm was wholly chosen, and it should not determine whether assistance or prevention is warranted.

The heart of the matter is that our body doesn't know the difference. An artery doesn't care if the cause of its failure was a choice or a catastrophe, and neither does a grieving family. The moral sorting is real, it's human, and it's one of the most expensive mistakes we make, because it aims our fear and our prevention away from exactly the dangers that kill the most of us. Dead is dead. The slow death you decided was deserved kills exactly as completely as the dramatic one that you are sure was a tragedy, and it's far more likely to be the one that comes for you.

The cure is method, not willpower

Use an unexplained concern as a reason to investigate, not as a verdict. Compare the observation with the evidence and consider competing explanations. A method can check an initial impression without guaranteeing a correct answer.

When you notice fear rising, learn to ask: is this the smoke detector, or is this the data? Is this dread, or is this probability? Am I numb to a large number that should frighten me, or over-concered about a small one that shouldn't? You can't feel your way to an accurate map of danger, because feelings were calibrated for a different world. You can reason your way there. Naming and acknowledging the bias is the first move. Running the numbers is the second. Everything after this is how.

The Fear Economy

Our wiring supplies the vulnerability. An entire economy supplies the exploitation. Fear is a product. It's manufactured, packaged, and sold to audiences who consume it the way they consume any other commodity: with appetite, with habit, and with remarkably little scrutiny about where it came from or whether it's good for them.

The industries that sell fear aren't a conspiracy. They're a market. Media organizations learned long ago that threat generates more engagement than reassurance. Politicians learned that mobilizing fear is more reliable than mobilizing hope. The security industry profits when people feel unsafe. This is not a charge against any individual. There's a structural incentive that systematically distorts how we perceive risk.

Notice what that incentive selects for. A plane crash is vivid, rare, and photogenic, so it leads the broadcast, even though all the people killed in plane crashes in a year would never approach the number of people who die in a single day from diet induced heart disease. Terrorism kills a few dozen Americans in a typical year, yet sustains a security and surveillance apparatus that costs tens of billions of dollars a year. The mismatch isn't an oversight. It's the logic of the market made visible.

Compare spending with risk and the expected benefit of the proposed intervention, not visibility alone.

What gets amplified

A dramatic event supplies images, people, and a clear timeline for a news report. Coverage of chronic disease or cumulative hazards requires a different form of reporting. Compare the actual coverage rather than assuming those hazards receive none.

My brother Andrew Wolf spent years as chief videographer at CNN, and he put the rule to me plainly: "If it bleeds, it leads. An abandoned warehouse fire in the middle of nowhere beats out a medical budget cut that impacts millions of people. Stories are chosen for visual impact, not importance."

Compare coverage of concentrated events with coverage of traffic deaths, hospital harm, and chronic disease over the same period. The relative amount of coverage is an empirical question, not proof that a hazard never appears in the news or that nobody reads about it.

None of that is random. The dramatic, concentrated, visually legible event gets the coverage, the mundane and statistically enormous hazard doesn't, and you end up with a population whose sense of threat is unmoored from the numbers.

The coverage gap

The distortion is measurable, and it isn't subtle. One year of New York Times reporting gave homicide about forty-three times more coverage than its share of deaths, and terrorism more than eighteen thousand times more. Heart disease and cancer together took well over half the deaths studied and about seven percent of the coverage. Rare, dramatic deaths filled more than half the reporting while accounting for less than one percent of the dying. Counted per body, heart disease earned about one article for every three hundred deaths, terrorism dozens for every single death. What people search for on their own tracks reality far more closely, so the distortion sits in the supply of news, not in our curiosity.

A concentrated event and a long-term disease process present different reporting challenges. The coverage comparison identifies a mismatch for the selected categories; it does not establish that every reader's risk perception is the inverse of reality.

The damage doesn't stop at feeling. Communications researcher George Gerbner spent decades studying what heavy television does to a person's picture of the world, and named the result the "mean world syndrome." The more dramatized danger people absorb, the more dangerous they believe the world is, and the more they overestimate their own odds of becoming a victim. The screen installs a worldview, and that worldview drives behavior, votes, and budgets.

Follow the money and the inversion turns concrete. After the attacks of September 2001, the United States spent on the order of two and a half trillion dollars on counterterrorism over the next decade and a half, well above a hundred billion a year, while terrorism killed a number of Americans you could seat in a single high-school gymnasium. Nothing close exists for the chronic diseases that kill hundreds of thousands every year, or for medical error, overdoses, car crashes, or the diet that does most of the quiet killing. The spending didn't follow the bodies. It followed the fear, and the fear followed the coverage.

The crime that fell while fear rose

Through the early 2020s, surveys kept finding that large majorities of Americans believed crime was rising nationally. Through much of that same period it was falling, and not slightly.

Murder surged during the pandemic, one of the largest one-year jumps on record. Then it came down. The most recent national figures show the fastest drop in murder ever recorded, with the violent crime rate and the property crime rate falling to lows not seen since the 1960s. The danger people named most often as rising was in historic retreat. That gap, between a real decline and a widespread belief in the opposite, is the fear economy working in plain sight.

Averages hide local realities. A falling national crime rate does not establish that every neighborhood became safer. A gap between survey responses and national data warrants investigation; it does not prove that someone profited from causing it.

The political dimension

The fear economy runs across the whole political spectrum. Different tribes keep different favorite dangers: crime, terrorism, environmental collapse, government overreach. The fears amplified in any given media ecosystem track those tribal preferences far more reliably than they track the evidence.

Here's a test you can apply for the rest of your life. When the fear you're being sold circulates mainly among people who share your politics, and it conveniently targets people or forces who are already your opponents, treat it with extra skepticism. Not because it's false; sometimes those fears are legitimate. Treat it that way because tribal fear amplification isn't evidence-based threat assessment, and the two reliably produce different outputs.

Political affiliation does not substitute for an assessment of cardiovascular risk.

Hospital infection and wildfire prevention require attention to exposure, vulnerability, and effective controls. Political attention to a hazard is not a measure of its medical or physical significance.

The safety industry's perverse incentives

The distortion reaches inside the safety world itself. After a dramatic accident, the institutional reflex is to bolt on visible countermeasures aimed at that exact kind of accident, no matter how rare it is, and no matter whether the new measures will stop the next one.

That's understandable as liability management. It's often counterproductive as safety. Money poured into the dramatic, recent, high-visibility incident is money not spent on the mundane, common causes of ongoing harm. Safety theater, the performance of safety without the substance, is the direct product of incentives that reward visibility over effectiveness. I've testified in cases where a facility had impressive binders full of procedures and a body on the floor, because the binders were written to be seen, not followed.

The alternative is unglamorous: sustained, evidence-based investment in the risks the data shows to be largest, whether or not they're in the news. That's how real safety cultures operate. It isn't how most organizations or political systems operate, because the rewards flow to the visible response, not the effective one. Spend your own attention the way a good safety culture spends its money: on what the evidence says, not on what the headline says.

You Can't Guard Against a Risk You Don't Recognize

Firearms instruction, CPR training, and advice about dietary risk address different hazards. Compare the preparation a household has with the risks relevant to its members. Class-enrollment assumptions do not establish where its largest gap lies.

That disproportion isn't an accident. It maps the risk landscape as people perceive it, not as it is. The space between those two maps is where preventable deaths live.

The visualization problem

The risks that hold your attention are the ones you can picture. The armed intruder. The plane falling out of the sky. The bomb. Those scenes arrive with immediate clarity, partly because they're dramatic and partly because news and entertainment feed them to you on a loop.

The risks more likely to kill you resist being pictured. You can't see plaque narrowing an artery. You can't see insulin resistance building across years of processed meals. You can't see a pathogen ride from an unwashed hand into a surgical incision. These things happen at a scale and a speed your visual imagination can't render, so your nervous system files them as not urgent. Being unable to picture a danger isn't evidence that the danger is small. It's a limit of the equipment, and it's one of the most expensive limits you carry.

What we prepare for

Thousands of shooting ranges, and an entire industry of equipment, training, media, and culture worth billions, all built around the possibility of armed confrontation. The person who practices monthly at the range is doing genuine preparation. The scenario is unlikely; the preparation is still rational, the way carrying a spare tire is rational.

But how many of those same people have had a serious conversation with a physician about their cardiovascular risk? How many have changed what they eat based on the evidence about which patterns drive the diseases that actually kill Americans? How many have looked up the infection record of the hospital where they're scheduled for elective surgery? On average, far fewer than the stakes justify. Not because people are foolish, but because the machinery that makes some kinds of preparation feel urgent and others feel optional has nothing to do with the underlying risk.

Expanding the threat model

In the security and intelligence world there's a concept called the threat model: the explicit identification of what a person or a system is defending against. A good one is built on evidence, weighs probability against severity, and updates as conditions change. A bad one is built on dramatic scenarios, and it overspends on low-probability, high-drama threats while underspending on high-probability, low-drama ones.

A personal risk assessment should include violent crime, diet, medical harm, traffic, falls, and alcohol where relevant. Compare the attention given to each with the person's exposure and the available means of reducing harm.

Updating the model isn't complicated. Look at the real data on what kills people like you, in your age range and health profile, and put your protective energy there. The data exists and it's public. What it takes is a willingness to let evidence override intuition when the two disagree. That's harder than it sounds. Intuition feels certain, and evidence has to be read.

The security researcher Bruce Schneier put his finger on the whole problem years ago: the feeling of security and the reality of security aren't the same thing. You can feel safe when you're not, and be safe without feeling it. A weapon in the nightstand delivers the feeling in abundance while doing almost nothing about the risks most likely to end your life. Real security comes from looking at where the danger is and spending your attention there, even when it gives you no comfortable feeling at all.

The doctor who won't wash their hands

Ask health-care staff to clean their hands before touching the patient, and raise a concern if hygiene is inadequate. This addresses a specific infection-control step, not every route of infection or every major risk in the patient's life.

Guns, Cameras, and the Gap Between Perception and Reality

Nowhere is the gap between the danger we picture and the danger we run wider than with the object at the center of the American imagination. I've been a firearms instructor, a stunt coordinator, and an expert witness in shooting cases. I've handled more firearms than most people will see in a lifetime.

A firearm-incident investigation should examine loading checks, muzzle direction, trigger handling, storage, transfer, supervision, and the firearm itself. Do not assume that ignorance or carelessness explains every event before testing the evidence and alternative failure mechanisms.

The safety ecosystem

Training, disciplined handling, supervision, and suitable equipment are separate safeguards. Their presence reduces specific hazards but does not establish that a firearm operation is risk-free. Their absence warrants correction, not a prediction of the date of an accident.

I put it this way for my concealed-carry students: owning a handgun no more makes you able to defend yourself than owning a guitar makes you a musician. The object isn't the skill. Training, practice, and discipline are the skill, and they live in the ecosystem, not in the steel. A gun in a drawer is potential, nothing more, and potential cuts both ways.

None of that means the preparation is irrational. I teach people to carry, and I stand by it, because a firearm has something in common with a fire extinguisher. You have to own it before you need it, you have to reach it immediately, and you have to know how to use it at once, because if you have to wait, it's not much use. Nobody calls you paranoid for keeping a fire extinguisher in the kitchen, and nobody should call you paranoid for preparing to defend your life. The mistake isn't preparing. The mistake is letting the vivid, rehearsed danger crowd out the quiet ones far more likely to be the thing that kills you. You can carry a gun and still understand that your fork, your car, and your habits are the larger threats.

What the firearm numbers actually say

In the most recent year of complete federal data, about 44,000 Americans died from gunshot wounds. The popular image, the stranger with a gun, the robbery, the random violence, describes a minority of those deaths. Most were suicides.

In the federal totals summarized here, suicides account for about six in ten firearm deaths and homicides for roughly a third. Accidental deaths are a smaller share. Homicide totals do not identify how many victims were killed by strangers. Keep those categories separate when interpreting the data.

The large suicide share warrants attention to crisis support and reducing immediate access to lethal means. A category's share of deaths does not by itself rank the effectiveness, reach, or cost of competing interventions.

When that ecosystem collapses on a film set, the failure is public. The shooting on the set of Rust, a matter in which I was retained as an expert, is the clearest case I can point to, and Chapter 7 takes it apart failure by failure.

The camera effect

There's a phenomenon I think of as the camera effect. Cameras and production pressure change behavior, and not always toward caution. People perform for cameras. They cut corners when the shot is ready, the director is waiting, and slowing down feels like an overreaction. The armorer who decides this isn't the moment for a complete weapons check, because everything has been fine all week, is making a choice the camera made easier.

Schedule pressure can conflict with time needed for checks in a hospital, construction site, warehouse, or production. Examine the actual staffing, incentives, supervision, and procedures before concluding that pressure caused a particular shortcut or injury.

The disproportion

Compare firearm mortality with other causes using the same time period and scope. Accidental firearm deaths are only one subset of firearm deaths. Using that subset to assess all firearm-safety attention would omit the larger suicide and homicide categories.

The value of mosquito precautions or firearm preparation depends on the person's exposure to the hazards each addresses.

This isn't an argument against firearms safety. It's an argument for proportion. The energy you spend defending against a danger should track the harm that danger actually does. Right now it doesn't, and people are dying in the mismatch, mostly from causes no one is arguing about.

Fire, Wildfire, and the Danger We Keep Ignoring

Wildfire preparation requires attention to fuel, weather, ignition sources, structures, evacuation, and available response resources. Examine the gaps in each rather than assuming that knowledge is complete and only political attention is missing.

The accumulation problem

Fuel accumulation can increase wildfire hazards. The consequences also depend on ignition, weather, terrain, structures, and response conditions. Assess those conditions together rather than treating accumulated fuel as a sufficient explanation of catastrophe.

Since 2020, my home state of Colorado alone has logged several of its largest and most destructive wildfires on record. Across the American West, fire seasons run longer and hotter than at any point in the record. The conditions are understood. The gap isn't knowledge. It's will, resources, and attention that lasts.

The technology gap

I work on wildfire defense because of the distance between what we know how to do and what we deploy in the field. I hold eleven patents; several relate directly to wildfire defense. When I describe to people in the field what these technologies can do, the response is almost never disbelief. It's recognition: yes, that would work, so why don't we have it at scale?

Deployment requires evaluating technical performance, operating limits, cost, organization, and political support.

What good preparation looks like

Wildfire preparation includes fuel management, detection, structural hardening, evacuation planning, and land-use decisions. Each requires site-specific design, resources, maintenance, and evaluation. Attention is necessary to sustain that work, but attention alone does not explain or solve the constraints.

Controlled fire and uncontrolled fire

I spent twenty years using fire professionally, in stunts, in special effects, in demonstrations. Controlled fire is one of the most powerful tools a competent professional has. Uncontrolled fire is one of the most destructive forces on earth. The difference isn't the fire.

The fire already in your house

When I say fire, you probably pictured a wildfire on a ridge, or the fireworks I'm coming to. The fire most likely to kill you is duller than either, and closer. It's in your living room, and the time you'd have to escape it has collapsed inside a single generation.

A generation ago, if a fire started in your home, you had about seventeen minutes to get out before the room reached flashover, the instant everything in it ignites at once and the space becomes unsurvivable. Today you have about three. Fire didn't get faster and firefighters didn't get slower. What's in the room changed. The cotton, wool, and solid wood that used to fill a house have been replaced by polyurethane foam, polyester, and plastic, which are solid petroleum. They catch faster, burn hotter, and reach flashover roughly seven times sooner. Underwriters Laboratories proved it, furnishing two identical rooms, one the old way and one the modern way, and lighting them side by side. The legacy room took about twenty-nine minutes to flash over. The modern room took three and a half.

Nobody voted for this. Nothing in your house got more dangerous on purpose. The threat grew because the couch changed, and the couch changed for reasons that had nothing to do with fire: cheaper, lighter, easier to clean. The whole argument of this book is sitting in your living room. A real danger got worse while everyone's attention was pointed somewhere louder.

Install and maintain smoke alarms, plan exits and an outdoor meeting place, and rehearse escape. Bedroom doors closed at night can provide a barrier to heat and smoke. Escape still depends on detection, the fire's location and growth, mobility, and usable routes.

The danger you light yourself

You don't have to wait for a wildfire to watch this play out in miniature. Every Fourth of July, Americans run the experiment on themselves. A firework, as I put it in my work on these cases, is a small machine engineered to fail in a beautiful, controlled way, and everything after ignition is physics. Every fuse has a specified burn time. Every shell has a known burst radius. The profession exists because the hazards are fully characterized and fully manageable. Injury begins the instant the failure stops being controlled, and on the Fourth that happens about 14,700 times a year, the number of Americans treated in emergency rooms for fireworks injuries in 2024, roughly two-thirds of them in the weeks around the holiday. Most are burns, and most of the burns are on the hands and the face, the parts of the body closest to the device when it fires. Even the sparkler handed to a toddler burns at around two thousand degrees Fahrenheit.

Now set that next to the danger people rehearse against at a summer gathering. Ask someone what they're wary of in a holiday crowd and you'll hear about the stranger, the bad actor, the rare violent moment. The numbers don't support that fear. On the Fourth, the things most likely to hurt you are the ones you brought to the party: the firework in your hand, the alcohol in the cooler, the water everyone is swimming in, and above all the drive home. The Fourth is one of the deadliest days of the year on American roads. Federal data counted 2,719 traffic deaths over the Fourth of July holiday periods from 2020 through 2024, with thirty-eight percent of the drivers killed legally drunk, and the first week of July is the single deadliest week of the year for drunk driving.

The danger at the barbecue is the barbecue. The threat is the celebration itself, administered by you.

A holiday gathering can involve fireworks, alcohol, water, and road travel. Plan controls for each relevant exposure rather than assuming that the most conspicuous hazard is the largest or that every danger is under the host's control.

On-Set Safety and What Hollywood Keeps Getting Wrong

On-set safety requires suitable equipment, technical controls, procedures, and the authority to use them under schedule and budget pressure. Investigate those elements together rather than ruling out a technology problem in advance.

The economics of risk

Film production runs on schedule and budget. Delay costs money, sometimes staggering amounts. Any protocol that slows the day creates pressure to shortcut it. And the people most responsible for safety, the stunt coordinator, the armorer, the special effects supervisor, are usually freelancers working at the pleasure of a production that can replace them the moment they become inconvenient. Safety independence is compromised before anyone steps on set.

People responsible for safety need authority to stop work and support when they use it. Review whether the rules are adequate, whether staff can follow them, and whether violations receive a proportionate response. Accountability does not substitute for resources or technical controls.

What actually causes accidents

Every accident I've investigated in thirty-five years involved a cascade of small failures, not one dramatic one. Training slightly inadequate. A protocol slightly unclear. A supervision gap slightly wider than it should have been. An accountability structure slightly too lenient. Alone, none of those had to produce harm. Together they built a tragedy.

Ask of any incident not just what failed, but how long it had been failing, quietly, before anyone noticed.

Normalization of deviance describes departures from procedure becoming accepted after repeated operations without apparent harm. Investigators should establish which departures occurred, how long they continued, and how they affected the incident. A history of uneventful operation does not prove a shortcut was safe.

Rust, and the cascade that caused it

The clearest case is the one the whole country already knows. On October 21, 2021, on the set of a Western called Rust at a ranch outside Santa Fe, New Mexico, a live round fired from a revolver in Alec Baldwin's hand killed the cinematographer Halyna Hutchins and wounded the director, Joel Souza. Hutchins was forty-two, and by every account exceptional at her work. Her death should never have happened.

It was a cascade. Live ammunition, which has no legitimate place on a film set, was present anyway, and roughly five hundred rounds were later recovered. The person responsible for the weapons was young and reportedly stretched across more than one job, and nobody ran the checks meant to catch exactly this error. New Mexico's workplace safety regulators issued their most serious category of citation, a willful violation, along with the maximum fine the state allows. Normalization of deviance, written in the worst possible ink. The protocols that would have caught it have existed, unchanged in their essentials, for decades.

Three deaths, and the meal nobody mentioned

Here the case stops being about Hollywood. The death of Halyna Hutchins set off one of the largest safety responses in the history of the industry. Thousands of hours of broadcast coverage. Hundreds of thousands of pages of reporting. Open letters signed by hundreds of cinematographers vowing never again to work alongside a functioning firearm. Productions that swore off real guns entirely, switching to plastic airsoft replicas with the muzzle flash painted in afterward by computer. Insurers rewrote how they cover any film that wants a real gun, and some stopped covering it at all. California passed the first state law in the country governing safety on a film set. The machinery of an entire industry turned, hard, in a single direction.

Now look at the number that machinery was answering. In the entire history of American film and television, deaths caused by a firearm discharging on set total three. Jon-Erik Hexum, in 1984, killed by a blank fired at close range. Brandon Lee, in 1993, killed by a fragment left in a barrel. Halyna Hutchins, in 2021. Three lives across roughly a century of production, each one a real loss, and not one of them evidence of an epidemic. The response was enormous. The hazard it answered, measured in lives, was among the smallest a person on that set faced all day.

Food, working hours, and long-term health deserve attention on a production alongside immediate stunt and firearm hazards. Comparing them requires evidence about crew exposures and outcomes over a defined period. A career-long health risk does not make a preventable acute hazard unimportant.

Review both the controls for on-set firearm use and the conditions affecting crew health over a career.

The response to an on-set death should address the failures supported by the investigation. Long-term health protections deserve their own assessment. The two priorities need not be treated as competing explanations of the same death or as mutually exclusive investments.

The Safety Ecosystem

Every case so far points at the same idea about what safety is. The conventional model is object-focused: dangerous things are dangerous, and we make the world safer by controlling access to them. That isn't wrong. It's radically incomplete. The same object can be safe or lethal depending on the system of people, training, protocols, and accountability around its use.

Assess equipment characteristics and the people, procedures, and conditions governing its use.

The five elements

Training. People who work with dangerous equipment or in dangerous places must be trained continuously, and checked to confirm the training was retained and applied. Training that isn't practiced decays. Training that isn't tested hides its own gaps.

Protocols. Concrete procedures for safe operation must exist, be communicated clearly, and be followed every time. A protocol's value isn't in existing. It's in being applied. A protocol followed only when things are easy is a liability document, not a safety system.

Supervision. Protocols require enforcement. In any high-risk environment, someone has to own the safety standard, and that person needs the authority to enforce it under pressure. An armorer with no authority to stop production isn't a safety officer. They're a prop.

Accountability. Address a protocol violation even when it causes no injury. Determine why it occurred, correct the condition, and apply proportionate consequences. An unaddressed violation warrants action without a prediction that tragedy is inevitable.

Culture. Staff need a shared expectation that they can report hazards, stop unsafe work, and obtain a response. Procedures cannot anticipate every condition, so people also need the training, resources, and authority to address unfamiliar situations.

The operator's responsibility

Safe handling is the operator's responsibility. That's the core principle of my work, and it's routinely misunderstood.

It doesn't mean the operator is responsible for everything. The ecosystem has many layers, and a failure at any layer can create conditions no one person can fully compensate for. It means the operator is the last line of defense, and that job can't be held passively. Every legitimate safety protocol I know builds the operator in as an active participant, not a passive recipient of protection. The operator verifies, checks, and refuses when conditions aren't right. The armorer checks the weapon before handing it over. The surgeon confirms the count before closing. The pilot runs the checklist before takeoff. Not because the systems before them have certainly failed, but because they know those systems might have.

Operators should verify the checks assigned to them rather than assume someone else completed them. That responsibility complements equipment controls, supervision, and the checks of other staff. In personal care, ask questions and report concerns without assuming that the patient can compensate for every system failure.

The Preventable Dozen

The following categories include opportunities for prevention. Their totals do not show how many deaths a particular intervention would prevent. Assess exposure, evidence of benefit, feasibility, and residual risk before choosing priorities.

This isn't the leading-causes table from Chapter 1. That table lists what people die of in clinical terms: heart disease, cancer, stroke. This one names the upstream causes, the behaviors, exposures, and system failures that feed those diagnoses. Diet runs through heart disease, stroke, diabetes, and several cancers, and it has its own chapter. Set it aside and look at the discrete killers that follow.

Read that chart like a triage board. The biggest bars are where the lives are. Notice what's missing: the dangers that dominate our fear have no bar big enough to see. The work is in the quiet rows.

1. Tobacco: the largest preventable killer of all

Cigarette smoking, including secondhand exposure, causes more than 480,000 American deaths a year, close to one death in five. It's the single largest preventable cause of death in the country. What's encouraging is how far it's already moved: adult smoking has fallen from more than 40 percent in the 1960s to roughly one in nine today. A deadly, embedded behavior can be turned around at population scale with sustained, evidence-based effort. That's the clearest success prevention has, and the template for the rest of this list.

2. Excessive alcohol

Excessive drinking is tied to about 178,000 American deaths a year, and the number is rising, up nearly a third in a few years. Two-thirds come from conditions that build over time: several cancers, liver disease, heart disease, and alcohol use disorder. The other third come from a single episode: crashes, poisonings, alcohol-involved overdoses. Alcohol takes more than two decades off the lives it ends.

It's one of the most under-recognized items here, because alcohol is so woven into ordinary social life that naming it a leading preventable killer feels rude. The data doesn't care.

3. Harm inside the hospital

Ray Chapman was one of the finest shooters who ever lived. He won more than 250 international competitions, including a world championship, and he taught survival shooting to soldiers, police officers, and special operations teams: men who walked into the most dangerous places on earth and walked back out because of what he taught them. He was my friend. What no enemy on any battlefield ever managed to do, an ordinary hospital did. After a minor fall in his garden, Ray was admitted for what should have been a brief visit. A lapse in procedure let an infection take hold, and a clerical error changed his medication dose. An afternoon became months of terminal suffering. Dead is dead, regardless of the cause. Ray is the reason the phrase at the center of this book has never been an abstraction to me.

Take this one seriously, because it's the danger you walk into voluntarily, trusting you're safe. The World Health Organization states the global picture plainly: about one patient in ten is harmed during health care, more than half of that harm is preventable, and medication mistakes account for roughly half of the avoidable harm. In American hospitals, federal data find that on any given day about one patient in 31 carries an infection they caught there. A recent study of inpatient care found an adverse event in nearly a quarter of admissions, and a preventable one in about one in fifteen.

Patient-safety rankings vary with the method and year. Preventable harm remains a serious concern. Patients can contribute by asking questions, checking information, and reporting discrepancies, while clinicians and institutions remain responsible for the care they provide.

This danger is different from every other one in this book: you're not a bystander. You're in the bed, or standing next to it. That's leverage you don't have against a wildfire or a drunk driver. I wrote a whole book about how to use it, called Deadly Hospital Mistakes, and its central rule is four words long: never be hospitalized alone.

If there's any safe and lawful alternative, don't face a hospital stay without an advocate. An advocate who knows the patient can watch, keep a log, and ask questions, complementing clinical monitoring. When the patient cannot participate, the advocate can communicate concerns. The job is to observe, verify, clarify, and escalate respectfully, not practice medicine.

The posture that works is firm, calm, and collaborative. A line from my hospital book that readers tell me they've used at the bedside: I know everyone is busy, and I appreciate what you're doing; I need to verify this before it happens. Said without apology, that sentence can save a life. So can a few specific questions, asked at the right moment.

4. Drug overdose

Overdose killed about 80,000 Americans in the most recent year, and here there's real news: that figure fell about 27 percent in a single year, the steepest decline ever recorded, after a brutal decade-long climb. Overdose is still the leading cause of death for adults under forty-five, and the supply is still dominated by illicit fentanyl, potent enough to kill in tiny quantities. But the drop shows that even an entrenched epidemic bends when prevention, treatment, and a lifesaving antidote reach people at scale.

5. Motor vehicle crashes

Crashes kill around 39,000 Americans a year. The toll has been drifting down, but it's higher than a decade ago and high against peer nations. The leading contributors haven't changed in a generation: speeding, impairment, distraction, and no seatbelt. These are among the most studied and most solvable risks in this book, and the most reliable life-saver is the cheapest device in the car. Wear it.

6. Falls

Falls killed more than 43,000 older Americans in the most recent year, and they're the leading cause of injury death for adults sixty-five and older. Most people file them under the ordinary wear of getting old. They aren't.

7. Firearm death

About 44,000 Americans die of gunshot wounds in a year. Chapter 5 took that number apart, and the shape of it is the reason it belongs on this list. About six in ten are suicides. Homicides, the category that drives nearly all the public fear, are roughly a third, and falling. Accidents are about one percent.

For firearm suicide prevention, address access to a loaded weapon during a crisis and access to help. Storage decisions and conversations are part of that work, not substitutes for clinical care or a reason to exclude policy measures.

8. Blood pressure nobody is watching

High blood pressure is the most common modifiable driver of cardiovascular death, and it has no symptoms until it does. It won't get its own line in the mortality tables, because it kills through the heart disease and stroke already at the top of them. That doesn't make it smaller. It makes it invisible, which is worse.

This is the purest example of the argument of this book. A danger with no face, no sound, no moment, and no story, sitting inside more than 680,000 heart disease deaths a year. The measurement takes thirty seconds and costs nothing.

9. Sitting still

Physical inactivity is the other killer with no line of its own. It works through the heart, the blood vessels, the metabolism, and the brain, so its toll is already counted inside the diseases at the top of the table, attributed to them rather than to the years of not moving that helped build them.

Regular activity can support cardiovascular, metabolic, and mental health. Choose an activity and intensity suited to current ability and medical advice rather than assuming that everyone can start without a plan.

10. The cancer that was catchable

Cancer takes more than 600,000 American lives a year. Prevention and appropriate screening can address some of that risk, but an outcome also depends on the cancer, its biology, test performance, access to care, and treatment. A missed appointment alone does not explain most differences in survival.

Ask a clinician which screening, preparation, follow-up, and repeat testing fit the patient's risk.

11. Drowning

Drowning takes about 4,500 lives a year and is the leading accidental killer of children between one and four in the data discussed here. Prevention warrants attention when children or adults have water exposure, regardless of its ranking against other national causes.

Drowning can occur without a loud call for help or conspicuous thrashing. Watch people in the water directly rather than relying on noise to signal an emergency.

12. Isolation

Social isolation belongs in a health assessment. It requires attention to relationships, access to help, and the person's circumstances rather than a universal comparison with smoking or every device and supplement.

A person without regular contact may have less help noticing a crisis, attending an appointment, or obtaining care. Assess those specific needs rather than assigning isolation a role in every other cause of death.

The one that isn't a cause: the bystander gap

Roughly 356,000 Americans suffer cardiac arrest outside a hospital each year in the estimates discussed here, and survival is low. Immediate bystander CPR can improve survival, yet not everyone receives it before responders arrive. That gap does not establish that every untreated patient could have been saved.

Use the relevant categories to check for gaps in prevention. Their importance to a particular household depends on exposure and the benefit of available measures.

Danger Is Local

Death rates vary across countries. Those differences can help identify prevention questions, but they do not by themselves distinguish the effects of policy, health, age, exposure, reporting, and other population differences.

The same body, different odds

An American and a Northern European share a species, a physiology, and the same medical textbooks. What they don't share is life expectancy. In 1980 the United States lived about as long as its peer nations. Then the lines split. Life expectancy kept climbing in comparable countries and stalled here, and the average American can now expect several fewer years of life than the average resident of Japan, Switzerland, Italy, or Australia. We spend far more on health care than any of them and we die sooner. That's not a medical mystery. It's a sum of preventable parts.

Researchers who line the United States up against wealthy peer nations take the survival gap apart cause by cause, and the same categories surface every time. They're not exotic. They're the quiet, preventable killers this book keeps returning to.

The chart compares U.S. rates with selected peer-country rates for several causes. A rate ratio is not a death count. Determining which causes contribute most to a life-expectancy gap requires comparable populations, periods, and an analysis of that gap, not addition of the plotted ratios.

This is the book's argument written in the language of geography. Even the one dramatic danger where America genuinely is more dangerous sits surrounded by quiet ones that are larger in total and far more fixable. Peer nations aren't shielded from heart disease and diabetes by magic. They have different diets, different built environments, different default behaviors, and systems that catch people earlier. None of it is beyond reach. Every one of those multiples is a number another country has already driven down.

The danger moves when you do

Road design, hospital care, and access to overdose treatment are relevant local conditions. A lower observed death rate does not prove that a particular intervention caused it. Examine the population, exposures, reporting, and implemented measures before transferring a practice.

Geographic variation can identify questions worth studying. A uniform risk is not necessarily unavoidable, and a variable risk is not necessarily controllable by an individual. Assess the cause of the difference and the evidence for the proposed intervention.

Why we can't feel the gap

Chapter 2 laid out why distributed harm never registers. What's new at this scale is the clock. If American life expectancy dropped several years overnight, it would be the only story on every screen. It didn't drop. It opened slowly, across four decades, paid one quiet death at a time among hundreds of millions of people, and a drift that slow never trips the alarm.

Cross-country comparisons can identify large differences that deserve explanation. They do not prove that public fear is misdirected or that the same measures will produce identical results in another population.

Investigate why a risk changes across borders before choosing measures to address it locally.

Danger Has a Season

The odds change when you cross a border. They also change when you cross a birthday, and when the calendar turns. The same person, in the same town, carries a different risk at twenty-five than at seventy-five, and a different one in July than in January. That doesn't overturn the argument of this book. It sharpens it. Knowing which danger is rising toward you points your attention more precisely, as long as you keep the sizes honest. Most of what they move is small. A few things aren't.

The seasons of a life

What kills Americans isn't one list. It's a stack of them, and the deck reshuffles about every fifteen years. Read the federal cause-of-death tables by age and you watch the leading killers trade places in a slow, predictable procession.

For a child between one and four, the leading injury killer is water. Drowning takes more of them than any other accident, often in a few unwatched minutes around a pool or a bathtub. Through grade school the road takes over, and the car stays the top accidental threat into the early teens.

From the late teens through the early thirties, the fast deaths rule. For Americans fifteen to thirty-four, the top three causes are unintentional injury, suicide, and homicide. Not one of them is a disease. Within injury, the largest killer of young adults is now drug overdose, then the car and the gun. For a person in their twenties, the dramatic deaths aren't a distortion of fear. They are the leading risks. It's the one stretch of life where what feels dangerous and what is dangerous agree.

Then they part ways. In the late thirties the slow killers move in, and by the forties heart disease and cancer have climbed into the top tier alongside injury. From forty-five to fifty-four they take the top two spots outright and never give them back. Somewhere in midlife, without anyone noticing, the leading threat to your life stopped being the thing that could happen today and became the thing you've built quietly for decades.

Reassess risks as age, health, activities, and location change.

That's the trap of the season of life. The fears we carry hardest are the ones we installed young, when the car and the stranger and the accident were the most likely ways to die. They were accurate then. They don't update on their own. A fifty-five-year-old who still organizes caution around the crash and the intruder is defending the perimeter of a twenty-year-old while the real breach forms silently in an artery. The danger didn't disappear. It changed shape, and the fear didn't follow.

The dangers that arrive late

Some risks rise with age. Falls warrant assessment of mobility, balance, medications, vision, and the home environment.

For most of adult life, a fall is an embarrassment, not a threat. That changes in the sixties. Falls are the leading cause of injury death for Americans sixty-five and older, killing more than 40,000 of them in a recent year, one in every fifty-six deaths in that age group. The rate doesn't rise gently. It explodes. From sixty-five to seventy-four, the fall death rate is around 19 per 100,000. By eighty-five and older it's around 340 per 100,000, about eighteen times higher. A fall a sixty-year-old walks away from is, for a body two decades older, often the beginning of the end.

In a bathroom, assess wet surfaces, lighting, support for transfers, and the person's mobility. Grab bars, a suitable seat, and lighting address specific hazards but do not replace an individualized fall assessment.

Ladder work requires assessing balance, strength, reach, footing, and the task. If those demands exceed the person's ability, arrange qualified help. Do not assume a younger helper is competent or that the work will cost a fixed nominal amount.

Even here, hold the proportion. Those 40,000 deaths are worth preventing, but they're a fraction of what heart disease and cancer take from the same age group. Falls matter because they're preventable and rising, not because they're the main event. The ladder and the grab bar are worth a Saturday afternoon. The plate, the blood pressure, and the daily walk are worth the rest of the year.

The seasons of the year

The calendar moves risk too, though less than folklore suggests. Some of the movement is real. Drowning is a summer event: July is reliably the deadliest month, more than 750 drownings in a recent July against a small fraction of that in winter, because the whole country gets in the water at once. Heat illness follows the same curve, concentrated in July and August. Impaired-driving deaths, around 12,000 a year, cluster in summer and around a few holidays. The week of the Fourth of July is the deadliest stretch on the road, and New Year's Day the deadliest day, when fatal drunk-driving crashes run roughly ninety percent above normal. Autumn brings hunting season, and injuries that come more often from falling out of a tree stand than from a gun. Winter brings house fires, space-heater and candle accidents, carbon monoxide poisoning, and a bump in falls as older adults climb ladders to hang and take down holiday decorations in December and January, the two months when fatal falls peak.

National totals provide scale, not permission to disregard a household's seasonal exposures. Use layered protection around water, wear seatbelts, and maintain appropriate alarms. No child is drown-proof, and those measures do not cover every seasonal hazard.

The season that matters, and the one we get backwards

Two seasonal patterns are large enough to change behavior, and the culture flags neither, because neither looks like a seasonal danger.

The first is the heart. Cardiac deaths aren't spread evenly across the year. They climb every winter and peak in late December and early January, sharply enough to have earned a grim nickname in the medical literature, the "Merry Christmas Coronary." More Americans die of cardiac causes on December 25 than on any other day, with December 26 and January 1 close behind. Cold, holiday stress, heavy food and drink, disrupted routines, fewer staff in the hospital, and above all the habit of ignoring chest pain so as not to ruin the holiday converge into one deadly window. This is the seasonal danger that earns your respect, and it's invisible because the biggest killer in the country doesn't register as a thing that has a season. It does.

It keeps a stranger calendar still. It notices when we change the clocks. Every spring, when daylight saving time lifts an hour of sleep out of the night, heart attacks tick up; one widely cited study found the Monday after we spring forward brought roughly a quarter more of them than an ordinary Monday. Every fall, when the hour is handed back, they drop. One hour of sleep, added or removed by an act of Congress, shows up days later in the cardiac unit. The biggest killer we have can feel the clock on your wall move.

The seasonal suicide pattern described here does not support a winter-holiday peak. Population seasonality does not identify when an individual needs help, and it does not establish how outreach resources should be allocated. Support must remain available throughout the year.

Use age and season to refine an assessment, not replace it. Plan for relevant water, road, fall, heat, fire, and health exposures. Respond to cardiac symptoms or signs of psychological distress whenever they occur.

Age and season add context to risk estimates. Combine them with health, activity, exposure, and available protective measures.

How to Evaluate a Real Risk

If the odds can change, you need a reliable way to read them. Risk assessment is a skill, and skills can be built. Here's the framework I use, in the courtroom and in my own life, when someone tells me something is dangerous.

Probability

How likely is this to happen to you, under your actual conditions of exposure? Base rates matter, and so does your situation. A risk that's common in one population is rare in another. Rest the estimate on evidence, historical data and empirical studies, and update it when the evidence changes. When something is called dangerous, ask: dangerous for whom, under what conditions, how often? The answers are usually nothing like the impression a news story leaves behind.

Severity

If it happens, how bad is it? A low-probability, low-severity risk deserves few resources. A high-probability, high-severity risk demands attention now. Combine the two and you get expected harm, and expected harm is what should drive where you spend your protective energy. Not drama. Not visibility. Expected harm.

Counterfactual

Ask what additional benefit a proposed action is likely to produce. Existing controls do not necessarily eliminate the benefit of another control. Compare the remaining risk, evidence of effectiveness, cost, and possible adverse effects.

Concentrated harm versus distributed harm

Compare concentrated and distributed harm over the same period and population. Emotional response and coverage are separate questions from the number and severity of outcomes.

Control

What can you do about this? Worry about a risk you can influence does work. Worry about one you can't mostly doesn't. The fear economy sells anxiety about large, systemic, uncontrollable threats and underplays the harms your own behavior, training, and attention could prevent. Eat differently. Ask whether they washed their hands. Wear your seatbelt every time, no exceptions. Get the screenings your risk profile calls for. Learn to restart a heart. None of it is glamorous. None of it delivers the satisfaction of preparing for a dramatic threat. All of it tracks where the risk actually is, and that's the only test that matters.

What Good Safety Culture Looks Like

A person can run these numbers alone. Keeping a whole group safe takes more. I've worked where safety culture was genuine and where it was theater. The difference isn't subtle, and it isn't mainly about equipment or rules.

Genuine safety culture

Where safety culture is real, safety isn't a department. It's a value. Everyone, senior to junior, knows the standards aren't negotiable, and that anyone who sees a violation has the right and the duty to say so. Near-misses and errors get reported and investigated, not buried, because an error shows where a system is fragile before the fragility produces a catastrophe.

Staff need permission and a workable process to report hazards and stop unsafe work. Reports need a response from someone with authority and resources. Do not assume that someone recognized the danger before every accident or that speaking alone would have prevented it.

The people closest to the work hold real authority. A stunt coordinator who judges conditions unsafe calls a hold, and that authority is real, not theoretical. It gets used routinely, respected without resentment, and never turned against the person who used it. When something does go wrong, the first question is systemic: what in this organization allowed this? Individual accountability follows, but the system comes first. That isn't excusing individual failure. Individual failures almost always happen inside systems that invited them.

Safety theater

Theater has its own signature. Rules everywhere, enforcement selective. Visible risks managed conspicuously, quiet ones piling up. Documentation elaborate, practice inconsistent. Violations that don't hurt anyone today become normal. When an accident finally comes, the response reaches for blame and visible new countermeasures instead of asking what the system did. People who raise concerns get handled as problems, not thanked as assets.

Most organizations running on theater don't know it. They believe, sincerely, that their binders and their rules are a safety system. The test is pressure: when the schedule collides with the protocol, when a shortcut looks low-risk, when a concern comes at an inconvenient moment. That's when the difference shows, and by then it's usually too late to learn it.

Building the real thing

Leadership must support training, supervision, reporting, and corrective action when they cost time or money. Technical expertise and suitable equipment remain necessary. Evaluate what the organization actually does under pressure rather than treating leadership and technology as alternatives.

The Expert in the Room

Expert analysis requires a method for testing explanations against the evidence and stating the limits of the result. A litigation record is not a substitute for that work.

In a courtroom my job is to help a decision-maker understand technical facts that sit outside ordinary knowledge. My obligation is to the evidence. Not to the party paying me, not to a conclusion someone reached before I arrived, not to a story that's already been decided. I tell the truth about what the evidence shows, whichever side it helps.

How I approach a case

I start with the evidence. What actually happened? What were the conditions? What did the people involved know, and when? What was the accepted standard of care, and how does what happened measure against it? Those questions produce uncomfortable answers. Sometimes the party that retained me has a weaker case than they believed. Sometimes the evidence points at the other side. I follow it anyway.

An investigation should establish the applicable standard, reconstruct what happened, and test how a departure contributed to injury. A difference between expected and actual conduct does not itself establish causation or legal responsibility. Compare alternative explanations and the evidence supporting each.

The foreseeability questions this book keeps returning to are a different tool, and they come after mine: my investigation asks how an accident happened, and foreseeability asks whether anyone should have stopped it. Two halves of one refusal to accept that a death simply happened.

Credibility requires accurate claims and acknowledgment of uncertainty. Evaluate the evidence whether it supports or weakens the retaining party's position. A favorable result does not by itself validate an expert's method.

What makes an expert credible

Trust the expert who has been retained by both plaintiffs and defendants in similar cases, whose opinions track the evidence rather than the paying party, and who can describe the limits of their knowledge as plainly as its reach. An expert who's always certain, never admits uncertainty, and reliably produces exactly the opinion the client needs isn't an expert. They're an advocate with credentials.

Hold the experts and sources in your own life to the same standard. Who's telling you what's dangerous? What are their incentives? Do their claims track the evidence, or what their audience wants to hear? That isn't cynicism. It's the minimum due diligence owed to any claim about what might kill you.

The non-expert's toolkit

When evaluating a risk claim, identify the evidence, assumptions, competing explanations, and the claimant's incentives. Ask what evidence would contradict the claim. Those checks help assess reliability without guaranteeing protection from error.

It Has Been Done Before

By now you might be feeling something close to despair. If our fear is miscalibrated by evolution, amplified by industry, and sorted by a moral instinct that lets the worst killers off the hook, what hope is there of pointing it correctly? More than you'd think. Within living memory this country has taken a danger that was normalized, profitable, and even celebrated, and taught itself to be afraid of it. The results are measured in hundreds of thousands of lives. Not once. Several times. Misallocated fear isn't a law of nature. It's a habit, and habits can be broken on purpose.

The examples below involve changes in knowledge, behavior, equipment, policy, treatment, and access. Assess their contributions separately rather than assigning each improvement to a single moment when public fear changed.

The road that learned to scare us

In 1972 the American road killed 54,589 people, the worst year on record. Driving was dangerous the way weather is dangerous, a fixed cost of modern life nobody expected to change. Then it changed, because people changed it. A consumer-safety movement put vehicle design on trial. A new federal agency set crash standards. Seat belts, then crumple zones, then air bags were engineered into the machine. Mothers Against Drunk Driving and the families behind it turned drunk driving from a punch line into a disgrace, and the "designated driver" entered the language. Enforcement followed.

The result is one of the great quiet victories of the age. The death rate per mile driven fell by roughly eighty percent, even as Americans tripled the miles they drive. Seat belt use climbed from about one in seven drivers in the early 1980s to about nine in ten today. By the government's own estimate, those decades of safety improvements saved more than 600,000 lives. None of it was inevitable. Every piece was a fear installed on purpose, around a danger that used to feel like the price of getting around.

Changes in smoking and tobacco control

In 1964 the Surgeon General told Americans that cigarettes cause cancer. Smoking was then glamorous, advertised on every screen, and entirely normal: more than two in five adults smoked. What followed was sixty years of warning labels, advertising restrictions, taxes, smoke-free laws, and a patient reframing of what a cigarette meant. Today the adult smoking rate is around one in nine, the lowest in sixty years, with roughly twice as many former smokers as current ones. Tens of millions of people walked away from one of the most addictive products ever sold.

The decline in smoking accompanied warning labels, advertising restrictions, taxes, smoke-free laws, and changes in public attitudes. Fear alone does not account for the combined effect of those measures.

The emergency we met head-on

By 1995, AIDS was the leading cause of death for Americans between twenty-five and forty-four, killing more than 50,000 people a year. The response was neither fast enough nor fair, and it came at a terrible and unequal cost. But it shows what reallocated fear can do at speed. The communities most at risk changed their behavior at a pace public health rarely sees. Activists forced faster drug approval and broader access. In 1996 combination antiretroviral therapy arrived and cut the annual death toll by nearly half in a single year, and by more than sixty percent within five. A death sentence became, for those who could get the medicine, a manageable condition. Fear, aimed correctly and backed by science and money, turned one of the deadliest plagues of the century into a controllable illness.

What the wins have in common, and what they expose

Road safety, tobacco control, and HIV prevention involve different mechanisms and outcome measures. Each illustrates the need to connect evidence with specific measures and sustained implementation.

Diet-related prevention requires evidence about specific foods, diseases, and interventions, together with practical access to healthier choices and appropriate care. A diagnosis of insufficient public fear does not establish which measures would work or what they would cost.

We have taught ourselves to fear the right thing before. The only question left is whether we will do it again, for the danger that's quietly the largest of all.

Personal decisions can support prevention, but the preceding examples also involve equipment, policy, research, treatment, and access. Examine the contributions and limits of each.

The Courage to Look

A sound method helps test a risk estimate; it does not guarantee a correct answer or make action easy. Identify the evidence, uncertainty, and practical barriers before deciding what to do.

The idea is older than any of the science in this book. Aristotle taught that courage isn't fearlessness, which he counted as a kind of recklessness, but a trained judgment that sits between cowardice and recklessness: the courageous person feels fear, and feels it in proportion, fearing the right things, to the right degree, at the right time. He thought a person could be faulted for fearing what they shouldn't, and equally for failing to fear what they should. Strip away the philosophy and it's the argument of this whole book in a single ancient sentence. To fear well isn't to feel less. It's to feel accurately, and then to act on what's true rather than on what's loud.

The wrong kind of brave

We celebrate the wrong courage. The culture rewards the person who runs toward the fire, draws the weapon, takes the dare. On a film set, the performer who agrees to the dangerous stunt gets the applause. I've spent decades on those sets, and I can tell you where the real courage lives. It's not in the person who takes the risk. It's in the person who stops the shot. The bravest act on any production is the one voice that says we're not rolling until that rig is rebuilt, knowing everyone is tired, the light is going, the budget is bleeding, and you'll be the reason for the delay. That person fights every social instinct in the room. That's courage, and it's the courage this book asks of you.

Real safety is almost never dramatic. It's a seat belt clicked for the ten-thousandth time, a screening scheduled, a vegetable eaten, a hard conversation with a doctor, a question asked out loud when staying quiet would be easier. None of it photographs well. Safety is doing the boring, correct thing long after the feeling of danger has faded. That's a quieter, more demanding bravery than the kind we put on screens.

Method is courage's tool

Courage without method is just nerve, and nerve gets people killed. Professionals stay calm around real hazard because they've replaced feeling with procedure. They don't trust the adrenaline; they trust the checklist. In my world we say it as more science and less cowboyism.

Run a pre-mortem on the risks relevant to the household or workplace. Consider what could go wrong, the evidence for its likelihood, and which measure could reduce the harm. Do not decide in advance that the answer must be a neglected personal habit rather than an external threat.

The discipline of the unspoken concern

Raise a specific concern before work or care proceeds. Identify the observation, request a check, and confirm who will respond. Speaking can make a hazard visible to others; preventing injury also requires an effective response.

How to practice it

Like any discipline, this one is built by repetition, not insight. When you notice fear rising, name its source before you act: smoke detector, headline, or data? The naming buys a second of reason. Once a year, write down the handful of things most likely to end your life early, in order, and check what you're doing about each. The gap between that list and your daily worries will be instructive. And treat the boring maintenance of your safety, the screenings, the seat belt, the sleep, the hard conversations, as the high-status act it is. You're not neurotic. You're doing the most rational thing available to someone who'd like to keep living.

The reward

Use the assessment to focus preparation on relevant risks without assuming that uncommon hazards no longer matter. Revisit the plan when circumstances or evidence change. Reduced worry is not proof of safety.

Courage isn't the absence of fear. It's the redirection of it.

That redirection takes no talent. It takes the quiet, repeated, unglamorous work of looking. Dead is dead, and the whole point of seeing that clearly is to keep yourself, for as long as you can, on the living side of it.

The Thirty: How Not to Die

Use this checklist to identify prevention needs. The order is not a measured ranking of lives saved by each action, and the appropriate priorities depend on the person and setting.

Include dietary and cardiovascular risks in the assessment, alongside tobacco, alcohol, medical conditions, and other relevant exposures.

Choose relevant measures, implement them, and review the result. The combined benefit cannot be calculated by adding overlapping disease and risk-factor totals.

Train Your Fear

Fear isn't the enemy. It's a survival mechanism, and pointed at real threats it's one of the most useful tools we have. The problem is that modern life hijacked the mechanism and aimed it at targets chosen by media cycles, commercial incentives, and political calculation, not by evidence.

Training your fear means building a habit. When anxiety about a hazard rises, ask: is this calibrated to real risk? What does the evidence say? Am I afraid because this is dangerous to me, or because someone whose interests differ from mine made it vivid?

Treat diet as a health exposure, ask about hand hygiene during care, and review the leading causes of death relevant to age and circumstances. Use those checks to identify practical actions, not to rank one precaution above everything else in every setting.

Ask what was foreseeable, what was known, and what action could change the risk. Then examine feasibility, implementation, and evidence of benefit. Knowledge and attention alone do not establish preventability.

Recognize the hazard, assess the evidence, and implement a suitable control. Recognition starts that work; it does not complete it.

The armed robber is real. The plane crash is real. The wildfire is real. So is the cheeseburger. So is the unwashed hand. So is the accumulated weight of a thousand small daily choices. The question is never which of these dangers exists; they all do. The question is which ones deserve your attention, in proportion to the harm they're likely to do.

Deaths from preventable causes warrant attention regardless of how dramatic the event was. Prevention requires evidence about causes, feasible controls, and the people or institutions able to implement them.

Use the evidence to choose actions, then check their implementation and limits.

Where These Figures Come From

Every claim in this book rests on the numbers, so the numbers deserve an honest accounting of their sources and their limits.

The leading-cause-of-death figures come from final United States mortality data published by the National Center for Health Statistics, drawn from the National Vital Statistics System, for the most recent full year available when this edition went to press. Firearm death totals and their breakdown by intent come from the same federal mortality system. The violent crime figures come from the Federal Bureau of Investigation's national crime reporting. Drug overdose figures come from the National Center for Health Statistics provisional and final counts. Traffic deaths come from the National Highway Traffic Safety Administration. Older-adult fall deaths draw on federal mortality data and the National Safety Council. Excessive-alcohol deaths come from the Centers for Disease Control and Prevention's alcohol-related disease estimates. Tobacco figures come from the Centers for Disease Control and Prevention. The dietary evidence draws on a 2024 umbrella review of ultra-processed foods, a 2025 Lancet series, large prospective cohort studies, and the clinical work of Dr. Caldwell Esselstyn at the Cleveland Clinic, alongside the population research associated with Dr. T. Colin Campbell. The hospital-safety figures and patient guidance draw on the World Health Organization, the Centers for Disease Control and Prevention, the Agency for Healthcare Research and Quality, and my own book Deadly Hospital Mistakes.

Two honest caveats. First, mortality figures shift from year to year, and some of the most recent are provisional and will be revised; I've used round numbers in the text so that small revisions don't make the argument wrong. Second, the estimates for preventable medical harm vary widely by method, and I've deliberately avoided a single headline number in favor of an honest range. None of this softens the central point. The proportions aren't close, and the proportions are the argument.