Why Can’t Americans Sleep?

I didn’t tolerate my Paxil brain for long. I weaned myself off, returned to normal for a few months, and assumed that my sleeplessness had been a freak event, like one of those earthquakes in a city that never has them. But then my sleep started to slip away again, and by age 31, I couldn’t recapture it without chemical assistance. Prozac worked for years on its own, but it blew out whatever circuit in my brain generates metaphors. When I turned to the antidepressants that kept the electricity flowing, I needed sleep medication too—proving, to my mind, that melancholy couldn’t have been the mother of my sleep troubles, but the lasting result of them. I’ve used the lowest dose of Klonopin to complement my SSRIs for years. In times of acute stress, I need a gabapentin or a Unisom too.

Unisom is fine. Gabapentin also turns my mind into an empty prairie.

Edibles, which I’ve also tried, turn my brain to porridge the next day. Some evidence suggests that cannabis works as a sleep aid, but more research, evidently, is required. (Sorry.)

Which brings me to the subject of drugs. I come neither to praise nor to bury them. But I do come to reframe the discussion around them, inspired by what a number of researcher-clinicians said about hypnotics and addiction during the SLEEP 2024 panel on the subject. They started with a simple question: How do you define addiction?

It’s true that many of the people who have taken sleep medications for months or years rely on them. Without them, the majority wouldn’t sleep, at least in the beginning, and a good many would experience rebound insomnia if they didn’t wean properly, which can be even worse. One could argue that this dependence is tantamount to addiction.

But: We don’t say people are addicted to their hypertension medication or statins, though we know that in certain instances lifestyle changes could obviate the need for either one. We don’t say people are addicted to their miracle GLP-1 agonists just because they could theoretically diet and exercise to lose weight. We agree that they need them. They’re on Lasix. On Lipitor. On Ozempic. Not addicted to.

Yet we still think of sleep medications as “drugs,” a word that in this case carries a whiff of stigma—partly because mental illness still carries a stigma, but also because sleep medications legitimately do have the potential for recreational use and abuse.

But is that what most people who suffer from sleep troubles are doing? Using their Sonata or Ativan for fun?

“If you see a patient who’s been taking medication for a long time,” Tom Roth, the founder of the Sleep Disorders and Research Center at Henry Ford Hospital, said during the panel, “you have to think, ‘Are they drug-seeking or therapy-seeking ?’ ” The overwhelming majority, he and other panelists noted, are taking their prescription drugs for relief, not kicks. They may depend on them, but they’re not abusing them—by taking them during the day, say, or for purposes other than sleep.

Still, let’s posit that many long-term users of sleep medication do become dependent. Now let’s consider another phenomenon commonly associated with reliance on sleep meds: You enter Garland and Hendrix territory in a hurry. First you need one pill, then you need two; eventually you need a fistful with a fifth of gin.

Yet a 2024 cohort study, which involved nearly 1 million Danes who used benzodiazepines long-term, found that of those who used them for three years or more—67,398 people, to be exact—only 7 percent exceeded their recommended dose.

Not a trivial number, certainly, if you’re staring across an entire population. But if you’re evaluating the risk of taking a hypnotic as an individual, you’d be correct to assume that your odds of dose escalation are pretty low.

That there’s a difference between abuse and dependence, that dependence doesn’t mean a mad chase for more milligrams, that people depend on drugs for a variety of other naturally reversible conditions and don’t suffer any stigma—these nuances matter.

“Using something where the benefits outweigh the side effects certainly is not addiction,” Winkelman, the Harvard psychiatrist and chair of the panel, told me when we spoke a few months later. “I call that treatment.”

The problem, he told me, is when the benefits stop outweighing the downsides. “Let’s say the medication loses efficacy over time.” Right. That 7 percent. And over-the-counter sleep meds, whose active component is usually diphenhydramine (more commonly known as Benadryl), are potentially even more likely to lose their efficacy—the American Academy of Sleep Medicine advises against them. “And let’s say you did stop your medication,” Winkelman continued. “Your sleep could be worse than it was before you started it,” at least for a while. “People should know about that risk.”

A small but even more hazardous risk: a seizure, for those who abruptly stop taking high doses of benzodiazepines after they’ve been on them for a long period of time. The likelihood is low—the exact percentage is almost impossible to ascertain—but any risk of a seizure is worth knowing about. “And are you comfortable with the idea that the drug could irrevocably be changing your brain?” Winkelman asked. “The brain is a machine, and you’re exposing it to the repetitive stimulus of the drug.” Then again, he pointed out, you know what else is a repetitive stimulus? Insomnia.

“So should these things even be considered a part of an addiction?” he asked. “At what point does a treatment become an addiction? I don’t know.”

Calvinist about sleep meds, blasé about sleep meds—whatever you are, the fact remains: We’re a nation that likes them. According to a 2020 report from the National Center for Health Statistics, 8.4 percent of Americans take sleep medications most nights or every night, and an additional 10 percent take them on some. Part of the reason medication remains so popular is that it’s easy for doctors to prescribe a pill and give a patient immediate relief, which is often what patients are looking for, especially if they’re in extremis or need some assistance through a rough stretch. CBT‑I, as Ronald Kessler noted, takes time to work. Pills don’t.

But another reason, as Suzanne Bertisch pointed out during the addiction-and-insomnia-meds panel, is that “primary-care physicians don’t even know what CBT-I is. This is a failure of our field.”

Even if general practitioners did know about CBT-I, too few therapists are trained in it, and those who are tend to have fully saturated schedules. The military, unsurprisingly, has tried to work around this problem (sleep being crucial to soldiers, sedatives being contraindicated in warfare) with CBT-I via video as well as an online program, both shown to be efficacious. But most of us are not in the Army. And while some hospitals, private companies, and the military have developed apps for CBT-I too, most people don’t know about them.

For years, medication has worked for me. I’ve stopped beating myself up about it. If the only side effect I’m experiencing from taking 0.5 milligrams of Klonopin is being dependent on 0.5 milligrams of Klonopin, is that really such a problem?

There’s been a lot of confusing noise about sleep medication over the years. “Weak science, alarming FDA black-box warnings, and media reporting have fueled an anti-benzodiazepine movement,” says an editorial in the March 2024 issue of The American Journal of Psychiatry. “This has created an atmosphere of fear and stigma among patients, many of whom can benefit from such medications.”

A case in point: For a long time, the public believed that benzodiazepines dramatically increased the risk of Alzheimer’s disease, thanks to a 2014 study in the British Medical Journal that got the usual five-alarm-fire treatment by the media. Then, two years later, another study came along, also in the British Medical Journal, saying, Never mind, nothing to see here, folks; there appears to be no causal relationship we can discern.

That study may one day prove problematic, too. But the point is: More work needs to be done.

A different paper, however—again by Daniel Kripke, the fellow who argued that seven hours of sleep a night predicted the best health outcomes—may provide more reason for concern. In a study published in 2012, he looked at more than 10,000 people on a variety of sleep medications and found that they were several times more likely to die within 2.5 years than a matched cohort, even when controlling for a range of culprits: age, sex, alcohol use, smoking status, body-mass index, prior cancer. Those who took as few as 18 pills a year had a 3.6-fold increase. (Those who took more than 132 had a 5.3-fold one.)

John Winkelman doesn’t buy it. “Really,” he told me, “what makes a lot more sense is to ask, ‘Why did people take these medications in the first place?’ ” And for what it’s worth, a 2023 study funded by the National Institute on Drug Abuse and published in the Journal of the American Medical Association found that people on stable, long-term doses of a benzodiazepine who go off their medication have worse mortality rates in the following 12 months than those who stay on it. So maybe you’re damned if you do, damned if you don’t.

Still, I take Kripke’s study seriously. Because … well, Christ, I don’t know. Emotional reasons? Because other esteemed thinkers still think there’s something to it?

In my own case, the most compelling reasons to get off medication are the more mundane ones: the scratchy little cognitive impairments it can cause during the day, the risk of falls as you get older. (I should correct myself here: Falling when you’re older has the potential to be not mundane, but very bad.) Medications can also cause problems with memory as one ages, even if they don’t cause Alzheimer’s, and the garden-variety brain termites of middle and old age are bummer enough.

And maybe most generally: Why have a drug in your system if you can learn to live without it?

My suspicion is that most people who rely on sleep drugs would prefer natural sleep.

So yes: I’d love to one day make a third run at CBT-I, with the hope of weaning off my medication, even if it means going through a hell spell of double exhaustion. CBT-I is a skill, something I could hopefully deploy for the rest of my life. Something I can’t accidentally leave on my bedside table.

Some part of me, the one that’s made of pessimism, is convinced that it won’t work no matter how long I stick with it. But Michael Irwin, at UCLA, told me something reassuring: His research suggests that if you have trouble with insomnia or difficulty maintaining your sleep, mindfulness meditation while lying in bed can be just as effective as climbing out of bed, sitting in a chair, and waiting until you’re tired enough to crawl back in—a pillar of CBT‑I, and one that I absolutely despise. I do it sometimes, because I know I should, but it’s lonely and freezing, a form of banishment.

And if CBT-I doesn’t work, Michael Grandner, the director of the sleep-and-health-research program at the University of Arizona, laid out an alternative at SLEEP 2024: acceptance and commitment therapy, or ACT. The basic idea is exactly what the name suggests. You accept your lot. You change exactly nothing. If you can’t sleep, or you can’t sleep enough, or you can sleep only in a broken line, you say, This is one of those things I can’t control. (One could see how such a mantra might help a person sleep, paradoxically.) You then isolate what matters to you. Being functional the next day? Being a good parent? A good friend? If sleep is the metaphorical wall you keep ramming your head against, “is your problem the wall?” Grandner asked. “Or is your problem that you can’t get beyond the wall, and is there another way?”

Because there often is another way. To be a good friend, to be a good parent, to be who and whatever it is you most value—you can live out a lot of those values without adequate sleep. “When you look at some of these things,” Grandner said, “what you find is that the pain”—of not sleeping—“is actually only a small part of what is getting in the way of your life. It’s really less about the pain itself and more about the suffering around the pain, and that’s what we can fix.”

Even as I type, I’m skeptical of this method too. My insomnia was so extreme at 29, and still can be to this day, that I’m not sure I am tough enough—or can summon enough of my inner Buddha (barely locatable on the best of days)—to transcend its pain, at once towering and a bore. But if ACT doesn’t work, and if CBT-I doesn’t work, and if mindfully meditating and acupuncture and neurofeedback and the zillions of other things I’ve tried in the past don’t work on their own … well … I’ll go back on medication.

Some people will judge me, I’m sure. What can I say? It’s my life, not theirs.

I’ll wrap up by talking about an extraordinary man named Thomas Wehr, once the chief of clinical psychobiology at the National Institute of Mental Health, now 83, still doing research. He was by far the most philosophical expert I spoke with, quick to find (and mull) the underlayer of whatever he was exploring. I really liked what he had to say about sleep.

You’ve probably read the theory somewhere—it’s a media chestnut—that human beings aren’t necessarily meant to sleep in one long stretch but rather in two shorter ones, with a dreamy, middle-of-the-night entr’acte. In a famous 2001 paper, the historian A. Roger Ekirch showed that people in the pre-electrified British Isles used that interregnum to read, chat, poke the fire, pray, have sex. But it was Wehr who, nearly 10 years earlier, found a biological basis for these rhythms of social life, discovering segmented sleep patterns in an experiment that exposed its participants to 14 hours of darkness each night. Their sleep split in two.

Wehr now knows firsthand what it is to sleep a divided sleep. “I think what happens as you get older,” he told me last summer, “is that this natural pattern of human sleep starts intruding back into the world in which it’s not welcome—the world we’ve created with artificial light.”

There’s a melancholy quality to this observation, I know. But also a beauty: Consciously or not, Wehr is reframing old age as a time of reintegration, not disintegration, a time when our natural bias for segmented sleep reasserts itself as our lives are winding down.

His findings should actually be reassuring to everyone. People of all ages pop awake in the middle of the night and have trouble going back to sleep. One associates this phenomenon with anxiety if it happens in younger people, and no doubt that’s frequently the cause. But it also rhymes with what may be a natural pattern. Perhaps we’re meant to wake up. Perhaps broken sleep doesn’t mean our sleep is broken, because another sleep awaits.

And if we think of those middle-of-the-night awakenings as meant to be, Wehr told me, perhaps we should use them differently, as some of our forebears did when they’d wake up in the night bathed in prolactin, a hormone that kept them relaxed and serene. “They were kind of in an altered state, maybe a third state of consciousness you usually don’t experience in modern life, unless you’re a meditator. And they would contemplate their dreams.”

Night awakenings, he went on to explain, tend to happen as we’re exiting a REM cycle, when our dreams are most intense. “We’re not having an experience that a lot of our ancestors had of waking up and maybe processing, or musing, or let’s even say ‘being informed’ by dreams.”

We should reclaim those moments at 3 or 4 a.m., was his view. Why not luxuriate in our dreams? “If you know you’re going to fall back asleep,” he said, “and if you just relax and maybe think about your dreams, that helps a lot.”

This assumes one has pleasant or emotionally neutral dreams, of course. But I take his point. He was possibly explaining, unwittingly, something about his own associative habits of mind—that maybe his daytime thinking is informed by the meandering stories he tells himself while he sleeps.

The problem, unfortunately, is that the world isn’t structured to accommodate a second sleep or a day informed by dreams. We live unnatural, anxious lives. Every morning, we turn on our lights, switch on our computers, grab our phones; the whir begins. For now, this strange way of being is exclusively on us to adapt to. Sleep doesn’t much curve to it, nor it to sleep. For those who struggle each night (or day), praying for what should be their biologically given reprieve from the chaos, the world has proved an even harsher place.

But there are ways to improve it. Through policy, by refraining from judgment—of others, but also of ourselves. Meanwhile, I take comfort in the two hunter-gatherer tribes Wehr told me about, ones he modestly noted did not confirm his hypothesis of biphasic sleep. He couldn’t remember their names, but I later looked them up: the San in Namibia and the Tsimané in Bolivia. They average less than 6.5 hours of sleep a night. And neither has a word for insomnia.