Something strange has been happening in the story of smoking in America. For decades, public health officials have delivered a simple and urgent message: cigarettes kill, and they kill in many ways—through cancer, heart disease, strokes, lung disease, and a host of other ailments. That message has gotten through. Smoking rates have fallen dramatically, from a time when cigarettes were nearly everywhere, sold in vending machines, smoked in offices and airplanes and even hospital lobbies, to an era in which lighting up is banned in most public places and a teenager who smokes is increasingly the exception rather than the rule. According to the American Lung Association, adult smoking fell from 42.6 percent in 1965 to just 11.6 percent in 2022. Among young people, the decline has been even more extraordinary: from 36.4 percent in 1997 to 3.8 percent in 2021. And yet, despite all this progress, something seemed wrong. Study after study appeared to show that the risk of dying from a smoking-related disease was climbing even as fewer people smoked. Between 1987 and 2013, the overall risk of death associated with smoking appeared to jump by nearly 50 percent. It was a troubling paradox, and it led to a unsettling possibility: maybe cigarettes themselves had become more deadly, or maybe the people who continued to smoke were somehow more vulnerable than previous generations of smokers. A team of researchers decided to investigate. Using data from the U.S. National Health Interview Survey, linked to death records through 2018, they asked a question that sounds simple but turns out to be subtle: Has smoking actually become more harmful, has the way people use cigarettes changed, or have the people who smoke simply come to look different from everyone else? Their answer, published in PNAS Nexus, is a fascinating lesson in how statistics can mislead when they ignore the human dimension of age. The increase in smoking’s mortality risk was mostly an illusion—not a sign that cigarettes became more toxic, but a reflection of the fact that the average smoker today is much older than the average smoker of the past.
To understand why, the researchers, led by Ralph Lawton and his colleagues, began by doing what earlier researchers had done. They took the entire population of adults, divided them into smokers and nonsmokers, and compared the death rates of the two groups. When they did this, they reproduced the earlier finding: the overall risk of death associated with smoking grew by nearly 50 percent between 1987 and 2013. This is the headline that has generated alarm, and it is mathematically real. But it does not mean what it appears to mean. The researchers then took a closer look, separating the data by age. Instead of comparing the average smoker with the average nonsmoker, they compared smokers and nonsmokers at the same stage of life: people in their forties, fifties, sixties, seventies, and so on. When they did this, the apparent increase largely disappeared. At any given age, the mortality gap between smokers and nonsmokers has not changed in any meaningful way over the past several decades. In other words, a 55-year-old smoker in 2013 faced roughly the same increased risk of dying compared with a 55-year-old nonsmoker as a 55-year-old smoker in 1987 did. The danger of a cigarette, as measured by its effect on an individual’s life expectancy at a particular age, has not suddenly become greater. What had changed was not the cigarette, and not the response of the human body to smoking, but the age composition of the smoking population. When all age groups are pooled together, older people dominate the average, and because older people are much more likely to die, this makes the average risk for smokers look higher than before. The researchers controlled for age, not by treating age as a nuisance to be eliminated, but by genuinely looking at what happens within each age group. And within each age group, the story has remained remarkably stable.
This demographic shift is the hidden engine driving the alarming numbers, and it has its roots in the success of decades of tobacco control. The people who once made up the great tide of American smokers, who began lighting up in the 1950s, 1960s, and 1970s, when smoking was not only accepted but often expected, are now older adults. As they age, they carry with them the cumulative consequences of years of exposure. And because younger generations have increasingly rejected smoking, there are fewer and fewer young people to replace them in the smoking population. This means the average smoker is no longer a young adult experimenting with cigarettes or a middle-aged person who can still imagine they have plenty of time. Instead, the average smoker is older, often much older, and age itself is one of the strongest risk factors for death. According to the Centers for Disease Control and Prevention, adults aged 45 to 64 now have the highest prevalence of cigarette smoking. The result is a self-reinforcing pattern: older people are more likely to have accumulated decades of smoking-related damage, more likely to have other health conditions such as high blood pressure, diabetes, or heart disease, and less able to repair the harm that smoking causes. Smoking-related diseases like lung cancer and chronic obstructive pulmonary disease take years to develop, so older smokers are exactly the people most likely to die from them. Lawton explained that the impacts of smoking on mortality are larger for older people, and the population—especially the population of smokers—has been getting older over time. When researchers produce estimates for the full population, they pool across older and younger individuals. Given recent demographic trends, older people have been getting more weight in the estimate, making it look like the risk of smoking is increasing for everybody, even though the age-specific risks have not changed.
Behind these abstract statistics are real people with real stories, and this is where the human meaning of the study becomes clear. Imagine a man in his late sixties who started smoking when he was seventeen, back when the dangers of tobacco were buried in internal memos and public health warnings were rare. He has tried to quit more times than he can remember, but nicotine has had a grip on him for fifty years. Or imagine a woman in her seventies who smoked through pregnancies and menopause and grief, whose social life still revolves around the coffee shop where she and her friends smoke outdoors. These are the faces of the American smoker today. They are not the carefree teenagers of old movies, nor are they the small number of young people who still take up the habit. They are older, and their bodies are more fragile. The same exposure to the hundreds of toxic chemicals in cigarette smoke that might have taken decades to catch up with a thirty-year-old has now caught up with them. This helps explain why the overall risk appears to have risen so sharply: not because a single cigarette has become more dangerous, but because a forty-year-old smoker has become a sixty-year-old smoker, and a sixty-year-old smoker is always more likely to die than a forty-year-old smoker. This does not mean smoking is safe, or that the threat is over. Smoking remains a leading modifiable cause of death, and the study makes clear that at any given age, the harm caused by smoking is essentially unchanged. But it changes the emotional tone of the conversation. Instead of panicking over a mysterious new danger, we can recognize that the danger was always there, and that the apparent increase is largely a statistical reflection of an aging generation. It also reminds us that older smokers deserve compassion, not judgment. Many of them began smoking before the full truth was known, and many have spent years trying to quit. The health system has rightly focused on preventing young people from starting, but it must also do more to help older adults break free from a habit that is now costing them the most.
The implications for research and policy are significant, and the study offers a clear lesson for anyone who tries to measure the harms of smoking. If researchers fail to account for changes in the age distribution of smokers, they will continue to produce misleading estimates. They may attribute rising mortality to a change in the tobacco itself or to some mysterious vulnerability in the current generation of smokers, when in fact the explanation is simpler and more structural: the people who smoke are older, and older people die at higher rates. Lawton acknowledged that models for the harms of smoking need to account very carefully for underlying changes in age, but he also struck a hopeful note, saying that “we show this is quite doable with thoughtful application of commonly-used tools. For smoking specifically, accounting thoroughly for aging gets you almost all the way towards explaining changes in risk over time.” This is a reassuring message, because it means researchers do not need to invent a new science to make sense of these trends. They just need to be more disciplined about separating the effects of age from the effects of smoking, and more attentive to the ways demographic shifts can masquerade as changes in risk. The same lesson applies beyond smoking. Lawton and his colleagues are already exploring smoking as a key marker of broader inequalities in midlife mortality in the United States, and they have been thinking about how similar methods can be used to examine changes in mortality risks associated with factors like obesity and education. In all of these areas, the population affected by a risk factor changes over time. It becomes older or younger, richer or poorer, more educated or less educated, more diverse or less diverse. If researchers forget to account for those changes, they can easily mistake shifts in the population for shifts in the danger itself. Careful attention to age, and to the broader social and demographic context, is not an obscure technical detail. It is essential to understanding who is suffering, why, and what can be done to help.
In the end, this study is good news, though it is not a reason to relax. It tells us that the apparent rise in smoking-related mortality is not evidence that cigarettes have become more deadly or that the remaining smokers are somehow doomed to suffer more than their predecessors. Rather, it is a sign of progress: fewer young people started smoking, the generation that smoked has aged, and the harms of smoking remain concentrated in the later years of life, as they always have been. The phrase “at any given age” changes everything. A 60-year-old smoker today faces a similar extra risk compared with a 60-year-old nonsmoker as a 60-year-old smoker did three decades ago. That is still a serious risk—the leading modifiable cause of death is not a small thing—but it is not a new or escalating danger. The real story is demographic, and it is also human. It is about a generation shaped by cigarette advertising and wartime rations and Hollywood glamour, a generation that is now reaching the age when smoking-related diseases come due. It is about the extraordinary success of a public health movement that convinced most young people never to start, and about the ongoing need to help the people who did start before that movement reached them. Quitting at any age brings benefits, sometimes almost immediate ones, and it is never too late. The study also invites all of us to be more thoughtful when we see scary headlines about risk. It asks us to wonder who is being counted, how old they are, and what else has changed in their lives. The danger of smoking is real, but it is not a mystery. It is a predictable consequence of age, exposure, and time. And while statistics can hide that truth, they can also reveal it, if we are willing to look closely enough. The rise in smoking’s mortality penalty is largely a story of aging—and understanding that story is the first step toward a more compassionate and accurate approach to public health.
