7 Things you need to know about social connection and your health
Your relationships influence your physical health, not just your mood. Psychologist Dr. Julianne Holt-Lunstad explains what social connection really is, and how much of it you actually need.
When Julianne Holt-Lunstad pooled 148 studies covering more than 300,000 people, one finding stood out: the strength of a person's social ties predicted survival about as powerfully as whether they smoked, and more so than obesity or physical inactivity. Relationships, it turned out, are not just good for the spirit. They are a physical-health variable in their own right.
That 2010 meta-analysis was Holt-Lunstad's, and it helped establish social isolation and loneliness as risk factors for early death. In a recent episode of A Whole New Level, Levels editorial director Mike Haney spoke with Holt-Lunstad, a professor of psychology and neuroscience at Brigham Young University and director of the Social Connection and Health Lab, who has spent her career building the scientific case that connection is not just good for our mood but is a fundamental driver of physical health. Her research since, spanning millions of participants, has only sharpened the picture, and she contributed to the U.S. Surgeon General's 2023 advisory on loneliness and serves as a technical adviser to the World Health Organization's work on social connection.
Her core message is that connection is a basic biological need, and that the way we usually talk about it, mostly in terms of loneliness, misses a large part of the risk. Here are seven things she wants people to understand about social connection and their health.
1. Social connection is a basic health need, not a nice-to-have
For most of her career, Holt-Lunstad has worked against a common assumption: that relationships are good for our mood but do not carry real, physical consequences. The data have steadily overturned that view. Her meta-analyses find that being socially connected predicts a substantially greater likelihood of survival, and that this holds even after accounting for age, initial health status, and lifestyle factors like exercise and smoking. This is not simply a case of connected people happening to live healthier lives.
"But what we see is that being socially connected has an effect over and above any kind of healthy lifestyle," she says.
"Social connection is more likely to increase that health span, not just lifespan."
Just as important as how long we live is how well. Holt-Lunstad draws a distinction between lifespan and healthspan, the number of healthy, functioning years we get. Plenty of medical interventions can extend lifespan while doing little for healthspan, or even eroding it. Connection appears to be one of the few factors that lengthens the healthy, high-functioning years rather than just the total on the odometer, which moves it from a feel-good extra to something closer to a pillar of the basics, alongside food, movement, and sleep.
2. Isolation and loneliness are two different things
The two terms get used interchangeably, but they describe different situations. Social isolation is objective, defined as having few relationships and infrequent social contact. Loneliness is subjective, the uncomfortable and sometimes distressing feeling that comes from a gap between the connection you want and the connection you actually have.
These often go hand in hand, since being objectively alone can make you likelier to feel alone. But not always. You can be isolated and perfectly content in your solitude, and you can be surrounded by people and still feel profoundly alone.
"People talk about feeling lonely in a crowd, feeling lonely at a party, and even feeling lonely in a marriage," Holt-Lunstad says.
The distinction matters because the remedy differs. Feeling lonely and being objectively cut off call for different responses, and, as the next points make clear, they also carry different levels of risk.
3. Every relationship breaks down into structure, function, and quality
When Holt-Lunstad asks her students how they would measure whether someone is socially connected, the answers pile up: How many friends do they have? Can they count on people? Are the relationships meaningful? Can they trust them? Science has measured connection just as many ways. Synthesizing all of it, she and other researchers landed on three core elements: structure, function, and quality. (The Surgeon General's advisory adopts the same three-part definition.)
Structure is the objective side, whether you have people in your life at all: the size of your network, the variety of relationship types, how often you engage. It is critical but, in her phrase, necessary but not sufficient. Function is whether those people actually meet your needs and goals, captured in measures like felt support and a sense of belonging. Quality is the tenor of the relationships, and it includes the negative as much as the positive.
"So you can have people in your life that you can count on but might be really judgy in the process, or really controlling, or you can think of relationships that might be associated with all sorts of strain and conflict, right?" she says.
When all three are robust, a whole host of good outcomes tends to follow. When any one is low, risk rises. Isolation is mainly a failure of structure. Loneliness, in how it is actually measured, taps mostly into function. Neither really captures quality, which is why Holt-Lunstad argues we need to think comprehensively rather than fixate on a single deficit.
4. You can be isolated without feeling lonely, and that may be the biggest risk
Early in her career, Holt-Lunstad expected the subjective side to win out. Her lab studies showed that perception drove physiology: people had worse stress responses when interacting with someone they saw as conflictual, and simply believing support was available, even if they never used it, was enough to calm the body. So she assumed the objective measures were crude indicators and that what really mattered was how we perceive our relationships. Her big meta-analysis upended that expectation.
"And so it was really surprising to me when I did that large meta-analysis that the strongest predictors were the objective ones, not the subjective ones," she says.
"Loneliness increases risk for earlier death by 14%. But isolation predicts earlier death by 32%."
Both matter, but in the most recent meta-analysis she points to, objective isolation is more than twice as strong a predictor of early death as loneliness. Her explanation is that perceptions shift week to week, while the objective features of a life, such as whether you are married or who you live with, tend to be stable and exert a chronic, day-in, day-out influence on your behavior, emotions, and physiology.
The unsettling implication is that the danger does not depend on the warning light being on. Loneliness is at least a signal that something is off, but isolation can be silent.
"But we might have a false sense of security, or a false sense of maybe a lack of risk or invulnerability, if we don't feel that loneliness when we're isolated," Holt-Lunstad says.
She pairs that with a note of self-awareness: we live in a society built for convenience and comfort, and our preferences are not always what is good for us. The routine that feels easiest, another night in with Netflix instead of seeing friends, can quietly be the unhealthy choice.
5. Being disconnected registers in your body as a threat
Why would relationships reach into our physiology at all? Holt-Lunstad starts with the biology, because without a plausible mechanism the links look like mere correlation. Across social species, being part of a group aids survival, from evading predators to regulating body temperature to obtaining food. Humans are among the most vulnerable species at birth; an infant left alone will not survive.
"And so what we see is that the human brain really expects social proximity," she says. Some researchers call this the social baseline: the brain runs most efficiently in the presence of others and treats being alone as a vulnerable state.
On your own, the brain has to work harder, staying vigilant to threats and shouldering every decision solo. That vigilance shows up in the body as a stress response and, over time, as inflammation. A short burst of inflammation can be protective. Sustained inflammation is not, and it has been tied to depression, heart disease, and dementia, the same outcomes that lack of connection has been linked to.
There is a behavioral layer on top of the biological one. Isolation makes people less trusting and more guarded, which invites less warmth from others, a self-reinforcing loop. And the people around us shape our health in ordinary, practical ways, reminding us to sleep, to take a medication, to finally get a lingering symptom checked.
"Interestingly, living alone, if there was some kind of acute event, whether that is a fall or a cardiac event, having someone in the home can literally be a matter of life and death in whether or not you get quick help in the moment," Holt-Lunstad says.
Her point is that it is not biology versus behavior but both at once, and that her mortality findings hold even after adjusting for lifestyle behaviors like physical activity and smoking.
6. You need a range of relationships, probably more than you think
A single great relationship is better than none, but Holt-Lunstad pushes back firmly on the popular idea that one or two is all anyone needs. The evidence, she says, simply does not support it. Relying on one person for everything is a fragile setup, and she has watched that fragility play out.
"I think about the data that we have on international students on college campuses where they may only have one or two friends, and then summer comes and they go away and suddenly they have no one, right?" she says.
Different relationships meet different needs, so variety is part of the prescription: partners, family, friends, neighbors, coworkers, and even acquaintances and the people you see in passing. On numbers, she frames this as a rough range on a continuum rather than a magic figure, but the pattern across national and population health studies points to roughly four to six relationships as adequate, with added benefit up to around six to nine. There also seems to be a floor for contact: studies of extreme isolation, from solitary confinement to lighthouse keepers and astronauts, suggest we need some daily contact with others. And much like a diet heavy in junk food, a social life made up mostly of "processed" interactions, remote and through screens, is a poor substitute for the real thing.
The casual, low-stakes connections matter too. "There are in fact studies that look specifically at weak ties and have demonstrated their importance," she says. Yet modern life is quietly eroding them. Self-checkout, delivery, and automation strip out the small daily exchanges with a cashier or a neighbor. In her Social Connection in America survey, Holt-Lunstad found that 67% of American adults report never participating in clubs or groups, not even once a year. That disengagement, she notes, is no longer a worrying minority. It has become the norm.
7. AI companions can quiet the feeling without meeting the need
As connection moves online and now toward AI companions, the obvious question is whether these count. The early research suggests not really. In a study by one of her graduate students, people either talked with a human, talked with an AI chatbot, or read a dull camera manual as a control.
"So what was found was that engaging with the AI chatbot was somewhat better than reading a manual, but it wasn't the equivalent of engaging with a human," Holt-Lunstad says. Other work finds that while a chatbot can reduce loneliness in the moment, the relief does not seem to last, and loneliness may even climb over time.
To explain why, she reframes loneliness as a signal, much like hunger or thirst, a biological prompt that motivates us to meet a real need.
"Imagine a world where we no longer feel thirsty, but we actually haven't been given any water. We're now chronically dehydrated."
That is her worry about AI companionship: it may be very good at dampening the signal without supplying the water. You stop feeling lonely, so you stop seeking people out, but the underlying need for human contact goes unmet. She also flags an ethical dimension. Scalable AI companions are cheap, while putting an actual person in a lonely older adult's home is expensive.
"And so we want to be really careful that we don't now create a class system where only the upper class have access to human contact," she says.
The bottom line: the fix is partly personal and partly structural
If connection is a basic need, how do we act on it? Holt-Lunstad is candid that for many people the barrier is simply that meeting new people feels awkward and uncomfortable. Her first suggestion is a reframe, the same move that sits at the heart of cognitive behavioral therapy, one of the better-studied ways to reduce loneliness. Rather than dwelling on her own unease in a room full of strangers, she looks for the other person who also has no one to talk to and focuses on putting them at ease.
"And sometimes if we flip how we think about it, and often just reframe it in some way, it can feel less vulnerable, less scary, less intimidating, and get us to put ourselves out there in ways that feel more comfortable," she says.
But she is adamant that the burden should not fall on individuals alone. She describes a physician whose isolated older patient was told to join a club, until the real obstacle became clear.
"But then realize the reason this person was severely isolated is that they lived in a community that was a very high crime rate, and they did not feel safe going out in their community," Holt-Lunstad says. "And so I use this example to illustrate that in some ways we often put the burden on the individual to fix something that might actually need to be more of a systems change."
We have robust evidence of the consequences of disconnection and much thinner evidence on what actually fixes it, partly because many programs never rigorously test whether people end up more connected, only whether they enjoyed themselves. And our environments often work against us: neighborhoods without common spaces, cities designed around cars rather than people, buildings built for efficiency rather than connection, and social norms that make us feel we are imposing if we ask for a ride instead of paying for one.
So the goal is not just to help someone feel less lonely, but to change what left them isolated in the first place, and to make sure they end up genuinely connected rather than merely not lonely. On a personal level, that can be as ordinary as doing the uncomfortable thing: joining the club, making the awkward call, and building the kind of network that means you are not calling a rideshare home from the hospital. It may not feel urgent, especially if you do not feel lonely. That, as Holt-Lunstad's research keeps showing, is precisely the point.
This article is based on an episode of the A Whole New Level podcast featuring Dr. Julianne Holt-Lunstad, a professor of psychology and neuroscience at Brigham Young University and director of the Social Connection and Health Lab.
Why Social Connection Is a Basic Health Need | Dr. Julianne Holt-Lunstad & Mike Haney
In a recent episode of A Whole New Level, Levels editorial director Mike Haney sits down with Dr. Julianne Holt-Lunstad, professor of psychology and neuroscience at Brigham Young University and director of the Social Connection & Health Lab. Holt-Lunstad has spent her career building the scientific case that social connection isn't just good for our emotional well-being but is a fundamental driver of physical health — her landmark 2010 meta-analysis, and later work spanning millions of participants, established social isolation and loneliness as risk factors for early mortality on par with well-known threats like smoking.
The conversation covers how she distinguishes isolation, loneliness, and living alone; the biological and behavioral pathways that connect our relationships to our bodies; what the research actually says about AI companions; how much connection we really need; and why the hardest part of the problem may be less about individuals and more about how we've built our communities.
"The human brain really expects social proximity. When we are alone outside the group, that's a very vulnerable place to be, and has been throughout human history." — Dr. Julianne Holt-Lunstad
How she came to study social connection
Mike Haney: Well, Dr. Julianne Holt-Lunstad, thanks so much for joining us today.
Julianne Holt-Lunstad: I'm happy to be here.
Mike Haney: So, maybe by way of bio, you've been at this quite a long time. Your sort of first big kind of landmark meta-analysis I think was 2010, correct? We just mentioned that you've been at BYU since 2001. So it seems like you've been at this since before it was as popular a topic as it is today, or at least as sort of widely accepted. So what got you into studying this? How did you end up looking at social connection in health?
Julianne Holt-Lunstad: Right. So when I was in grad school, my training is in health psychology, and I was working in a psychophysiology lab. I was doing studies where we were looking at the biological mechanisms by which stress influence health, and always part of that was folks' social relationships. So we looked at social relationships as a buffer to stress and how social support can dampen that stress signal, but we also looked at how relationships can be sources of stress and how that could exacerbate these physiological responses. And so really I started thinking a lot about my own social connections and how important they are not just in times of stress but more broadly. And so this really led to a lot of my work going forward in my career around the broader importance of social connections and really putting some robust evidence behind it. Because at the time also there was this perception that our relationships are just something that are associated with maybe emotional well-being but not something that has like hard evidence of the physical health outcomes.
Mike Haney: Well, I'm glad you phrase it that way because that's actually sort of the genesis of this episode. So we talk a lot about longevity and my sort of obsession is what I call sort of nailing the basics, right? Like I'm less interested in all the kind of optimization stuff. That's fine and interesting, but I feel like people aren't nailing the basics enough. And I always describe that as eat real food, sleep enough, move every day, and calm down. So, basically, diet, exercise, sleep, and stress. And I was doing an episode with a longevity expert who said, "Right, and also have friends." And I was like, "Right, of course."
Julianne Holt-Lunstad: I was about to jump in and say the same thing.
Mike Haney: Yeah. And it struck me that I've been underplaying that, because I think with so much of health it's so easy to get kind of self-focused, right? Like all the things I can do within the confines of my body and how that affects me, and think less about the outside world and all of that. But what I realized when I thought about that is, oh, I don't have a really pithy three-word phrase, because when we get into the specifics — and this is all what we're going to talk about today around social connection — there's a lot of different aspects of that. And so maybe just by way of starting, let's define some terms. The ones I wrote down here that seem to come up in the research a lot is social isolation, living alone, and loneliness, which are in fact three different things. And maybe there's other terms, but maybe just start by defining the terms that we're actually talking about and how they're different, right?
Defining the terms: isolation, loneliness, and living alone
Julianne Holt-Lunstad: Okay. So let's start with those three because those are all forms of lacking connection, and then we'll talk about, well, what is social connection? So, and I'm glad you started off — particularly isolation and loneliness are terms that get used interchangeably a lot, right? And so social isolation is really more objective. And so it is often defined as having few relationships and infrequent social contact. And now loneliness is more that subjective feeling of alone. So isolation is more objectively being alone. So that subjective feeling alone — and I should say it's an uncomfortable or even distressing form of a subjective feeling that is based on the discrepancy between one's desired level of connection and one's actual level of connection. And so of course these often go hand in hand because objectively being alone can increase your likelihood of feeling alone. And yet you can be isolated and not feel lonely. You might actually really enjoy your time alone. And you can be surrounded by other people, not isolated, and yet still feel profoundly lonely. People talk about feeling lonely in a crowd, feeling lonely at a party, and even feeling lonely in a marriage. And so I think those are helpful distinctions that can help us distinguish those, and we can talk later, but I want to make clear both are important for health. But the way in which we might go about them is a bit different.
Living alone is exactly what it sounds like. It's a single occupancy household. So there's only one person living there, and so you don't have anyone else living with you. That can sometimes be a marker of isolation, but as you can imagine, you can live alone but still have a wide social network, and so you may not be entirely isolated.
Now, those are all forms of lacking connection, but let me say a little bit more about what social connection is. And that really came about as an umbrella term to encompass all the ways in which scientists have studied and measured this. So let me just kind of give you a more informal way of thinking of this, because I'll sometimes with my students, I'll ask them, okay, if you were to measure how socially connected someone is, what would you do? And I start writing on the board, right? And people would list like, well, okay, how many friends you have? Others it's, well, whether you can count on people. Well, it's about how meaningful those relationships. Oh, it's about whether you can trust them. And as you can imagine, so many different ways come up. And similarly in science, so many ways have been measured, right? And so when we look across these and synthesize this evidence, it seems to fall under three core elements. And so social connection really is an umbrella term that encompasses these three core elements, and that is the structure, function, and quality.
So structure really gets at a lot of those objective kinds of measures. And really when you get down to it, it's to what extent do you have people in your life? So it gets at the size of your social network, but also the variety of types of relationships you have and the roles that you have and how frequently or how engaged you are. And so we know that that is critically important, to have people in your life, but we also know it's not sufficient. So it's one of those like necessary but not sufficient, because we know that we need people in our lives that we can actually count on. And so the function really gets at the extent to which people in our lives fulfill various needs and goals. And so it's often measured by things like social support and a sense of belonging and other things like that. And then the last one is quality. So you can have people in your life that you can count on but might be really judgy in the process, or really controlling, or you can think of relationships that might be associated with all sorts of strain and conflict, right? And so the quality gets at those positive and negative aspects. And what we see is when structure, function, and quality, when all of those are robust and high, it's associated with a whole host of positive kinds of outcomes. But when any one of those is low, it can lead to risk. And so when we think of isolation, that's more of lacking the structural element, and so is living alone. And loneliness, although it's often colloquially used as lacking any one of those, really when it comes to how it's measured, it really taps more into that functional element. And so neither of those really get at the quality actually. And so we really need to think comprehensively, not just about measuring the deficits, but we also need to think about how the positive and protective effects of social connection and how robust those are as well.
Mike Haney: That's really helpful, and I think we'll come back to those as we talk about some of the specific mechanisms here and ultimately some of the specific kind of remedies, or how we address this in our own life. But I want to look at one other set of definitions, which is around the kind of health side of this, right? So we talk about mortality, we talk about risk, we talk about sort of health outcomes. How do you think about those terms? We get into this a lot in longevity where you realize, like, wait, are we talking about being a healthy 80-year-old or are we talking about living to 130? Because those are really two different things, right? And my sense of the research around this is that it's sort of all of the above. We've got studies that look at some of those components you just talked about and tie them to some of these sort of outcomes. And I think the great trick of some of your work, your big meta-analysis, is kind of synthesizing all of that work into some kind of takeaways. But how do you think about the outcome side of this and the kind of definitions we should be having in our head around that?
Julianne Holt-Lunstad: Right. So mortality of course is death, right? And so first off, it's one of the most objective outcomes that we can look at because there's no ambiguity there. And in fact, that's why I studied in my meta-analyses mortality, because it's such a clear marker, right? But in these studies, how this is looked at is how they follow people over time and determine what factors predict who is alive or dead at the end of the follow-up period. And so oftentimes those follow-up periods can vary. But they also look at and try to isolate the particular factors. So for instance, when it comes to social connection, we need to identify it as an independent protective factor, or if we're looking at isolation and loneliness, an independent risk factor. And what I mean by that is independent of other factors that we already know increase our risk for mortality. So we know that as we get older, we are at increased risk for dying, right? All our various lifestyle factors, some of the ones you just mentioned, right? Those can influence it. And so we need to make sure that we're talking about how do these social factors independently predict our likelihood of earlier death compared to whether it's sleep or exercise or smoking or age or even initial health status. And so the one thing that I think is really quite clear and robust is that these are independent of those other factors that also predict risk.
I will also note, we talk about mortality, but in public health there's a lot of conversation around lifespan versus health span. And so a lot of medical interventions can increase our lifespan but can deteriorate health span. And what I mean by that is health span is those number of healthy functioning years. And so we can all imagine all sorts of medical interventions that can keep us alive, but our quality of life is terrible, right? And so when we think about, okay, if you wanted to live to 100 — in fact, I've heard a lot of people say, why on earth would I ever want to live to 100? That sounds miserable — because what they imagine is a life where maybe their functionality has declined, their physical functionality, cognitive functioning, and imagine a life in which it's not a fulfilling and full life that they currently enjoy. And so what I think is also critically important to recognize is that social connection is more likely to increase that health span, not just lifespan.
"Social connection is more likely to increase that health span, not just lifespan." — Dr. Julianne Holt-Lunstad
The biology of connection and the social baseline
Mike Haney: I think it's a good lead into the kind of relationship side of this, right? The causality versus the correlation part. Like, you call out in one of the papers this idea that if we're not looking at these adjusted data sets, then we might miss some of this kind of — I think you called it sort of reverse causality — meaning we could say, okay, social isolation for reasons we'll talk about might affect your physiological health and then you go into heart failure, but it could also be that you have heart failure and then that leads to social isolation because you can't go out and play pickleball anymore, and then that actually worsens it as well. So maybe talk first about just the physiological side and what do we know about the actual effects on our body of these different forms of sort of social isolation.
Julianne Holt-Lunstad: Yeah, I think that's a really important point, and we can talk about how we've eliminated some of the reverse causality kinds of elements, but I think it's first, as you say, important to really just identify like what are the plausible biological pathways that would even account for this. Because a skeptic could easily say these are just interesting correlations. But if you don't have those plausible pathways to explain it, it's less robust. So I think really stepping way back is talking about this from a conceptual standpoint and then the real data behind it.
And so what we see from across scientific disciplines, we see also across species, that being part of a group is essential for survival. So if you look across social species, being part of a group has a number of really critical survival elements, from evading predators to protection from the elements, regulating body temperature, efficiency of effort, obtaining food, all of these kinds of things. Being part of a group increases likelihood of survival. And if we think about this, humans are one of the most vulnerable species at birth. If a human infant is left alone, it will die. It cannot survive without the care of others. And so what we see is that the human brain really expects social proximity. And so it has adapted to expect that. And we see this in many studies where — and in some cases researchers have even dubbed the neural default as social baseline — and this idea that when we are around others that our brain and our physiology functions optimally, but when we are alone outside the group, that's a very vulnerable place to be and has been throughout human history. And so our brains need to be far more active, whether it is being vigilant to threats in our environment. So that may show up in terms of a stress response, but also defensive kinds of behaviors, to really also just having to be more active — brain being more active — to manage the demands of everyday life on your own. All the decision-making, all of the effort.
And so when our brains are more active, this also signals our periphery. And so this shows up in our physiology. And so this activation of these various systems on the one hand can be adaptive in the short term. So for instance it might signal an inflammatory response, which if there are threats in your environment an inflammatory response might be helpful if you were to get injured. However, if the inflammatory response is sustained long term, we know and we have good evidence around how that systemic inflammation creates dysregulation and can put us at increased risk for a variety of chronic illnesses. And I use the inflammation example because we know that it's been linked to depression. It's been linked to heart disease. It's been linked to dementia. And interestingly, lacking connection has also been linked to mental health, cognitive health, physical health. And so I use that example to illustrate a common biological mechanism that could potentially explain these diverse kinds of outcomes. But we also have seen through the evidence that there's robust evidence of dysregulation across multiple systems, and that this dysregulation over time can put wear and tear on our bodies, increasing our risk of developing chronic illnesses and earlier death. And so those biological processes — some of them are well understood, some we're still learning a lot about. But it's a clear pathway that explains these findings. And so it's really helpful to understand that, and understand also even how these can influence behavior as well.
Mike Haney: So I'm glad you brought up the social baseline theory because I find this idea interesting. It seems to me — and tell me if this is correct — we talk a lot about the sort of evolutionary response to stress, right? That our body is wired so that if it thinks there's a lion in the grass, all these hormonal things start happening in our body, and that's a good thing and there's a reason that evolutionarily evolved. But the problem in modern-day context is that there's never a lion in the grass, it's just traffic or it's an email I don't like or it's somebody upset me on social media, and so now I'm having these kind of physiological responses that are there for a reason but they're misdirected. It seems like the connection here could sort of go both ways, in the sense that if there is an actual or perceived threat and I have some form of community that can help me, that can help maybe calm that physiological response, so I'm not going to have the cortisol surge, I'm not going to have the inflammation. But also just the lack of social connection, absent any particular stressor, is itself a sort of causal mechanism. Is itself something that could raise inflammation, or that the body might simply perceive the lack of social connection as a stressor itself. Does that distinction make sense? Like absent things specific versus just not having connections.
Julianne Holt-Lunstad: Right. And I think there's also another element that we need to acknowledge too, and that is the behavioral element. Because while we have good evidence of these physiological responses and all of the biomarkers that are happening — and yes, indeed, we do see these effects — is that when, for instance, when we are again more isolated, we also are less likely to be trustful of others. And so when we do that, we start behaving in more defensive ways. And when we behave more defensive, it's less friendly, less open, which reinforces and can elicit less friendly responses in return. So that can be a reinforcing cycle that then can also trigger a threat response because we're seeing less friendly responses from others and further alienation.
But also we see how behaviorally this can influence other aspects of health. So for instance, we know that having others around us can have practical influences on our health. So from the time we're little, we have people who are encouraging us to brush our teeth, look both ways before we cross the street. And as we get older, it might be someone to remind us to get more sleep, or go see a doctor when we've got something lingering that we've been putting off, or to remind us to take our medication. Interestingly, living alone, if there was some kind of acute event, whether that is a fall or a cardiac event, having someone in the home can literally be a matter of life and death in whether or not you get quick help in the moment. And so there are these behavioral kinds of elements that interact with our biology as well. And so it's a both, right? When we see this stress response, we behave in certain ways, but also our behaviors can elicit stress responses. But then there's also the practical things.
I do want to acknowledge, however, that all of our big meta-analyses on mortality, we did adjust for and control for a lot of lifestyle behaviors. So we adjusted for physical activity, smoking, and other kinds of things. And so even though that might be one pathway by which our relationships can influence our health, it's not entirely accounting for it. So what I'm saying is someone could argue, well, maybe people who are more socially connected just eat better and have a healthier lifestyle, but really it's just the healthy lifestyle that's accounting for it. But what we see is that being socially connected has an effect over and above any kind of healthy lifestyle.
How we know what we know
Mike Haney: So, it sounds like I want to talk a little bit about the sort of research side of this and where we're at and how we sort of conduct this research, but it sounds like, over the past several decades of folks diving into this, what we've landed on is less of a kind of, oh, it's really the physiological mechanisms, or it turns out it's just somebody to remind you to take your medication, but it is all of the above, and they are intertwined and there really is no kind of separating them from each other. And I'm curious just on the research side of this. I mean, this is one of those topics where we can't run a massive randomized control trial and isolate 10,000 people and give 10,000 other people friends and follow them for 50 years. So how do we sort of conduct this research? And what is the kind of range? Are there mouse studies? Are there cell culture studies? Are there sort of intervention-type studies or just natural experiment studies? How do we learn what we know here?
Julianne Holt-Lunstad: All of the above. So this ranges from, as you say, animal models where they do indeed randomly assign animals to be housed in isolation or housed socially. And we see very robust kinds of effects on things like stroke and even mortality, cancerous tumor development. So there is experimental evidence that we see, and causal evidence in animals. And interestingly, I will say we have this accumulation of evidence across these different methodologies — and I'll share a few more — but the bottom line is that across these we see a convergence of evidence that gives us greater confidence in this. And in fact, I published a paper that took the same criteria that established smoking, the causality of smoking causing cancer, right? And if you think about it, there's somewhat of an analogy there in the sense that you can't randomly assign people to be smokers and some to be non-smokers, right? These things kind of naturally happen. And yet we very readily accept that smoking causes cancer. And so there was this criteria, the Bradford Hill criteria, that has nine criteria that helped to initially establish that. Now, of course, there are other ways that you can — and now we have more robust ways of establishing causality. But I just want to note that when I put the evidence on social connection head-to-head on all of these nine criteria, it met all but one of the nine criteria. And the only one that it didn't meet was specificity, meaning that there was only one pathway or only one kind of outcome. And interestingly, the same is true for smoking. It also does not meet that criteria, because it doesn't just cause cancer, it's linked to heart disease, it's linked to other kinds of things.
But so we have the animal studies, we have experiments. So even though we can't randomly assign people to have lots of friends or no friends, we have done situations where we randomly assign them to social situations and measure their physiology. And we do this all the time. In some cases, these are interventions where they are given some kind of social enhancement and another group is not, and we are able to see the kinds of effects that that has. Some of the most robust evidence we have are the kinds of large epidemiological studies where they're population-based studies, and these include in some cases literally millions of participants and follow them over time. This is what I did in my meta-analysis, one of which included over 3.4 million participants. And so we're able to look at how these things unfold over time, not just a short period, but really long-term kinds of outcomes. We also have studies that follow people not only in the lab but in their natural environments. And so there are so many different ways in which we can study this. And the beauty of it is that it's all converging to generally the same finding, and that is those who are more socially connected have better health outcomes, and those that lack that are associated with worse kinds of outcomes.
Different kinds of relationships — and AI companions
Mike Haney: Well, let's dive into some of the specifics then of what we mean when we talk about social connections, because I would imagine over the course of your career studying this, our kind of social environment has changed. I mean, the thing that most comes to mind is there wasn't really social media in the early 2000s. I don't know if MySpace was around yet, or Friendster, or whatever our early ones were, and now that's a big component of the broad umbrella of sociality, right? Text chains, I guess, maybe were around, but probably not really around. And now — and I want to get into more specifically AI and AI companions, right?
Julianne Holt-Lunstad: Right.
Mike Haney: So let's just talk about some of the kind of specific interventions here. What do we know about the differences between friends, family, weak ties, co-workers, the kind of — just talk about real world now, and people maybe in your actual physical vicinity as opposed to people I talk to on the phone or co-workers I see on Zoom? What do we know about the differences among those kinds of relationships and the effect that they have?
Julianne Holt-Lunstad: Well, we do have evidence that it's not just the size of our social network, that we need a variety of relationships, and that different kinds of relationships — as you mentioned, whether it's a close intimate relationship like a romantic partner, or family, friends, to our neighbors, our co-workers, and even acquaintances and strangers. All of these relationships are important and can fulfill different kinds of needs and goals that we have in our lives. And having a variety of them is really important. And so I always kind of balk at when I hear people say, "Oh, you only really need one or two good relationships." And the evidence just simply doesn't support that. And in fact, while of course having one or two great relationships is better than none, it does potentially put you at risk. Particularly, I think about the folks where their spouse or partner is their only relationship and they rely on that person for everything to fulfill all their needs. And if that relationship were to dissolve or end, whether it's a breakup or death, now that person is incredibly vulnerable, right? I think about the data that we have on international students on college campuses where they may only have one or two friends, and then summer comes and they go away and suddenly they have no one, right? And so having just a couple, if anything, it's vulnerable, because if anything were to happen, they don't have a backup of anyone else. But we know that these different relationships are really important.
And so now we're at a point where the way in which we engage with our social relationships and our social contacts is changing a lot. And so it's not to say that we haven't had other forms of remote means of connecting. So first there was letter writing. That's a remote form of engagement. There's talking on the phone. That again is also remote. But now we're getting into texting or responding through social media, to now, as you say, AI companions. And so we historically have looked at the extent to which people talk to others and how likely or how often they get together with others in person. And so we're starting to learn a little bit more about these remote kinds of interactions. And interestingly, now we're having interactions with non-humans, right? And so in some ways we've had interactions with other forms of non-humans. So pets, even a higher power, right? But there's something different about AI compared to some of these. And so we're starting to really learn a little bit more about that. And what the early research is suggesting is that it's not the same. And this is not surprising. Of course, we all lived through the pandemic and felt the insufficiency of only being able to engage with people at a distance, right? We saw the kinds of suffering that occurred. So we know that it's not entirely the same, but it helped to some extent, right?
Well, we're seeing — so I'll just give you one example of a study that one of my grad students did, where he had individuals have a conversation with a human versus an AI chatbot, and looked at not only physiological responses but also their reports of loneliness. And what they found — oh, and then a control condition was they were not interacting with anyone, but they were still talking, because physiologically we need to control for that. So they had to read a really boring — I think it was like a camera manual. And so what was found was that engaging with the AI chatbot was somewhat better than reading a manual, but it wasn't the equivalent of engaging with a human. And there are other studies that have shown that while engaging with an AI chatbot can reduce loneliness in the moment, it doesn't seem to reduce it — or that loneliness, that reduction doesn't seem to be sustained, and that there may even be increases in loneliness over time. So we really need to be thoughtful about what elements these things can tap into, where they might help facilitate social connection, and where they might be displacing, and that may even potentially cause harm.
Mike Haney: Yeah. Let's follow up on the AI side. I want to get to this eventually anyway, but it feels like there are so many variables here, and I suspect we're just at the beginning of understanding what it is, right? Like you just mentioned displacing. Okay, that makes sense. Like we did an episode about AI psychosis, right? So people who lose touch with their human connections and only talk to the AI, or come to trust or love the AI more than humans, one could imagine where that might be detrimental. But for socially isolated elderly people, widowers, et cetera, that don't have people in their lives, you could imagine the AI being better than nothing, right? That I can actually have something to kind of converse with. And then I wonder about the kind of skinning of it. There's one thing to open up the GPT window and just start to have a conversation. There's another thing if we start to embody that in some kind of a humanoid robot, which could be better or worse. Do we know anything about — when you mentioned that the AI companions may even increase loneliness over time — is it because there's some, do we know, is it because those same sort of evolutionary mechanisms that have trained us that we need to be around others of our species, that something in us knows this is not another of our species, and that while it's sort of getting that surface level, something's better than nothing, that it's not durable for that reason, that something in our brain is going, right, but it's not another human here?
Julianne Holt-Lunstad: Yeah. I think it's helpful to really think about the fact that we know that social connection is this biological need. And when we think about loneliness, it is really the biological signal that we are not meeting that need. So it's often been described as akin to hunger or thirst, where hunger is this biological signal that helps us recognize that we need to eat, and it motivates us. It's really uncomfortable and it motivates us, similar to thirst, right? And so from this standpoint, loneliness is actually somewhat adaptive in the short term, because it motivates us to meet that biological need. Now, what I worry about is that these AIs can do a really good job at dampening that signal so that we feel like we no longer feel that loneliness, right? But imagine a world where we no longer feel thirsty, but we actually haven't been given any water, right? And we're now chronically dehydrated. This could be very problematic because you no longer have that motive or desire to meet that biological need, but that dampening of the signal isn't actually meeting the need, right? And so we're not actually with other humans.
And so I think about also just the ethical side, the example that you gave of the older adults. Imagine there is a subpopulation of older adults that are homebound and that are quite isolated, and many of which are also very lonely. And so you could see where the natural response would be, let's give them these AI companions and this can fulfill a companionship need. And so again, it might reduce loneliness, but they haven't actually interacted with any humans, right? And so you're not actually fulfilling that biological need. And from a more kind of ethical standpoint, it also gets at cost. And so we see the desire for these kinds of solutions that are easily scalable. But the kinds of solutions that would actually bring a human into their home are quite costly. And so now you have in essence human connection as something that only those who are the most affluent can actually afford. And so we're also seeing where this might be human contact might be a luxury good, right? Where the masses can get access to these various tech AI companions, but to actually get someone to physically be in your presence and to care — and then we also think about even the kinds of engagement of actually being able to be in person. Sometimes that is also costly, whether it's time or the expense of being able to geographically be in the presence of others. And so we want to be really careful that we don't now create a class system where only the upper class have access to human contact.
"Imagine a world where we no longer feel thirsty, but we actually haven't been given any water. We're now chronically dehydrated." — Dr. Julianne Holt-Lunstad
Mike Haney: Yeah, I feel like we could probably spend another hour going down the AI dystopian rabbit hole. But I want to come back to sort of the — let's pull back from AI because there are so many things there we don't yet know and there's other stuff that I suspect we've got more studies on.
How much connection do we actually need?
Mike Haney: Do we know — is there a sort of — when we again just talk about kind of the in-person stuff for now. Is there a kind of minimum effective dose of social connection? Meaning, if I'm somebody who's fairly socially isolated, is asking the barista how her day is better than nothing, or is there a kind of spectrum here? Or is there kind of a minimum threshold of, like, if I do that, but that's still my only social connection and I don't have a real friend or a trusted person, I'm going to have the same kind of outcomes. What do we know about how much we actually need, or what can sort of be enough?
Julianne Holt-Lunstad: Yeah, it's a great question, because it's one I get all the time. And in fact I remember during the pandemic it was probably the number one question that journalists would ask, is just how much do we need? And so I wrote a policy brief on establishing national health guidelines for social connection, just like we have dietary guidelines and guidelines for physical activity and sleep. And in essence, what do we know from the evidence about what these guidelines might be? And I will note that there is some effort towards establishing these, but they are not imminent yet. So I went through this, and it's not as simple as we might think, but we were able to provide a few kinds of ranges, right?
And so just like we know that even the dietary guidelines are really complex because they can vary depending on if you're a child or someone who's pregnant, but nonetheless there are some general guidelines, right? And so when we give these guidelines, they're often for the average. But we also know that these guidelines are more behaviorally oriented, right? So I first started off with the most common question I would get, and that is, well, just how many people do I need, or how many friends do I need, or how many people do I need in my life? And so I will say that most of the evidence is measured on a continuum, and we do have evidence across that structure, function, and quality, those elements, of a dose-response effect. Meaning that for every increase in the indices of social connection, there was a dose-response of reduction in risk across some biomarkers and physiological biomarkers and even risk for mortality. So I'm caveating anything I say here, that there is this kind of potential dose-response effect and that this is more on a continuum. So we first start off with the evidence that we know certainly that having no relationships is clearly detrimental, and so we know clearly that you're at risk for that. So then we tried to think about, well, what would be adequate and what would be optimal? And so when it comes to numbers of relationships, adequate seems to be somewhere around four to six. And this has been shown across a variety of national health studies, population health studies. And then from beyond six, there are several studies that show that there are added benefits to that. Now I will say with the caveat that we don't have good evidence of what the upper limit is, and perhaps the reason is because kind of like exercise — because the majority of the population isn't getting enough exercise, we know less about kind of where people might be overdoing it at the upper limits. Similarly, when it comes to social connection, not enough people are getting enough and have enough that we don't fully understand the upper limits. But generally that four to six is adequate, six to nine seems to have added benefits, but we don't know whether there's some kind of threshold by which those benefits may not go beyond that.
So that's one recommendation. Another is around how much contact we need. And so this is also — we can look to evidence around those that have been severely isolated, solitary confinement, those that have no access to others. So professions where there's very limited contact. So those who work in lighthouses, many of the astronauts — there have been studies of very limited contact. And what we see is that when there's limited contact, that is associated with all sorts of detrimental effects. And so really what we see is that we do need some kind of daily contact with others. And so that's another one. We also see, similar to the dietary guidelines around how you should limit processed foods or sweets and salty, these kinds of what we often colloquially refer to as junk food, right, that we might think of making — even though we ought to limit the kinds of foods, we similarly ought to limit the kinds of processed social connections. So the kinds of interactions that we might have that are more remote or through social media, that these more processed kinds of interactions should not be the bulk of our social diet, but rather should be more limited. And so these are just a few examples of what we know from some of the evidence. But when it comes to actually creating national health guidelines, this does need to be a consensus recommendation. And just like these other recommendations, they periodically get updated based on newer data. And so similarly, any kind of social recommendations will also need to be periodically updated as well.
Mike Haney: So we talk about the studies that lead to the conclusion of say the four to six or six to nine. How much are they able to isolate the types of relationship? I guess getting back to that sort of function and maybe quality component to it. If the four people include the person at the convenience store I see every day, and I say hi Bob and he says hi Mike, does that count as my one? Or when we talk about four to six, do we need two family and two friends who we're actually hanging out with and sort of talking to?
Julianne Holt-Lunstad: Yeah, great question. And of course, the way in which we look at these is looking across the body of evidence. And what makes this somewhat complicated is that not all studies look at it in the same way. So it's important to think about how it's actually asked and assessed of those that are in the study. And so some look at it in terms of the number of people that they feel close to. So however that is defined for the individual. So for some people that would include their neighbor and for some people that would not include their neighbor, right? And so these different kinds of relationships, it ends up being what the respondent perceives as that. In some studies they actually specifically ask about and label it as friends. And again it would be how the respondent defines a friend. And in some that looks at family. So in some studies they do break down by different types of relationships, and then in others it looks at just kind of people in your life more broadly. And interestingly there's a high correspondence between these numbers, and so that's where summing across that we get to these people in our lives. But I would say that the bulk of the studies look at either people you feel close to, friends and family.
Mike Haney: So sort of less research on — I guess there probably have been specific studies, but in terms of synthesizing that into some kind of guideline, less research on the kind of what we might call weak ties, the casual co-workers, et cetera.
Julianne Holt-Lunstad: There are in fact studies that look specifically at weak ties and have demonstrated their importance. And even in my big study, I lead the Social Connection in America project. It's a national survey that assesses social connection both at the individual and community level. One of the questions — not only do we ask about the number of people that they feel close to, but we also ask about in essence acquaintances. And so that is also important in understanding how many people are in that broader network beyond what people might perceive as just close relationships. And because we do have evidence that the broader social network and these weak ties are important. And what is interesting is our modern life is in some ways starting to shrink that outer circle, those weaker ties, because we're having less engagement with people in our community, because much of our lives are automated. And the kinds of things where we would get things, whether it's shopping or other kinds of activities, we can often get those delivered now, right? Or we do the self-checkout, so we don't engage with a checker anymore. And so we're starting to see the erosion of some of these weaker ties. And so because we know of the importance of them, we're going to have to track over time what the kinds of consequences this might have. And already there is concern about how this might be related to more community-based and societal kinds of outcomes rather than individual-based outcomes. And of course all of these are really important because the health and functioning of our communities influences the health and functioning of our individuals, right?
Mike Haney: It sort of provides the infrastructure. The book that kept coming to mind when I was doing this research was Bowling Alone, which is obviously at this point, I don't know, 25 years old. But feels relevant to that exact point, right? As the community structures that would provide some of those connections — clubs and civic institutions, et cetera — as those go away, then we're sort of losing infrastructure for that kind of connection. The other thing I wanted to touch on here is, you know, as we're talking about this —
Julianne Holt-Lunstad: Can I interrupt for one second? I'll just note that, speaking of Bowling Alone, part of what he talks about is the importance of groups, and his focus is on democracy.
Mike Haney: Right.
Julianne Holt-Lunstad: Right. And so in our Social Connection in America project, what we found is that — in fact, I'm going to make sure I get it right, because I don't want to guess here, but I want to say it's 67%. Here we go. I have it right here. Yep. 67% of American adults report never participating in clubs or groups. The alternative was at least once a year. Okay, that's a low bar, right? So more than two-thirds of Americans are not participating in clubs and groups. This is back to the Bowling Alone, right? That this disengagement is now the norm. It's not something that is a concerning percentage of Americans. It's the majority of Americans are not doing this.
Perception versus objective isolation
Mike Haney: Well, that hits at the kind of functional component to this, or the sort of structural component, but I'm also curious — which seems real and a part of this — but so much of the language we use when we talk about this is, do people feel that they have a friend? Do they feel close to people? Do they have the perception that they are alone? How much of this comes down to the perception, right? And maybe this is another way of asking how much individuality there is within this. Because you can imagine — I'll just give my own example. So we moved to San Diego here about seven years ago solely for the weather. We knew nobody here and we had no family here, and I always say the natural experiment we ran was like, is weather more important than friends? And after seven years of being here I'm like, yeah, it is. I'm actually happier here. Now that said, I have a really good group of friends, but they're remote. They're guys I see three, four times a year on trips that we do. There is a sort of text chain where I sort of still feel connected to them. And I feel perfectly fine with it. But I do worry about this because I think about my health and I think about nailing the basics. And I think, even though I don't feel lonely, I don't feel socially isolated — I have a wife and a kid, so there's people around my house — but I do live far away from family, I do live far away from friends. This came home recently when I had to have a medical procedure. I needed a ride home from the hospital and my wife was scheduled to be out of town. And I realized, oh, I don't have anybody in San Diego I could easily that I would feel comfortable asking. There are people kind of on the periphery that I could, and they'd say yes, but I'd feel weird about it. And it was the first time it hit me that even though I don't feel socially isolated, I don't feel lonely, I'm perfectly happy with sort of the structure of my life — like maybe that's an indication that I am actually doing damage to my health and I just don't realize it. So I guess that's a sort of way of asking, how much of this comes down to just our feeling and how much of that is individual? Some people can do really well with some set of connections and other people need a whole different set.
Julianne Holt-Lunstad: Yeah, it's a great question and something I've grappled with throughout my career in my research. In fact, remember that when I first started in this research, I was doing these laboratory-based studies. And some of the things that we would do is we would look at how people's perceptions of their relationships influenced how they responded biologically. And we saw that when they were interacting with someone that they perceived to be more conflictual with, that was associated with worse physiological responses than if they responded to someone that they perceived to be more supportive. But what was really interesting is that even we saw that when — even if they just knew that support was available, even if they didn't get any support, just the perception that it was available to them if they needed it was enough to dampen that stress response. And so from these studies, I thought the kinds of more objective measures of social relationships are just these crude indicators. It really has to do more with how we perceive these relationships.
And so it was really surprising to me when I did that large meta-analysis that the strongest predictors were the objective ones, not the subjective ones. And that was a big surprise. And so I started really thinking about, well, why is that the case? Why would these objective indicators — first off they were both important, I just want to clarify that they were both important, but there was a stronger effect for some of these more objective ones. And I started thinking about the fact that what I was looking at was the long-term effects, right? And that our perceptions are more malleable and change more frequently over time, whereas some of these more objective ones are much harder to move. They're more stable, right? And so some of the most historically measured indicators are things like marriage, right? Whether you're married or not, whether you live with someone or not, those things don't change frequently. And the person that you live with has a chronic influence on you for good or for bad, right? That is a very persistent influence on a host of your behaviors, emotions, and which of course influence your physiology.
And so it was really interesting to me also that we see in meta-analyses that when it comes to objective isolation versus the perception of loneliness, that the latest meta-analysis shows that loneliness increases risk for earlier death by 14%. But isolation predicts earlier death by 32%. And so they're both important, but objective isolation is a significantly stronger predictor. And so I've thought a lot about this, because of course we talk a lot about loneliness and the dangers of loneliness, and these are of course legitimate concerns. We have good evidence of the concerns around loneliness. But we might have a false sense of security, or a false sense of maybe a lack of risk or invulnerability, if we don't feel that loneliness when we're isolated. And so there may be less reason to actually make any changes.
And so I'm cautious, of course, but I think about also our individual preferences, right? And we have a society that's designed for convenience and comfort, and sometimes that reinforces isolation. And so I think about the other kinds of health behaviors where we might take great pleasure in things that are not good for us and not healthy, and the kinds of things where we might not enjoy the things that nonetheless are good for us. And how to really come to terms with the fact that maybe our preferences aren't always what is good for us, and that sometimes we can be biased and desire things that may not be healthy. And so we need to be really careful of getting settling into these really comfortable routines of just watching Netflix instead of going out and engaging with friends. But we also need to acknowledge some of the societal barriers that people face, because I don't ever want people to take this as some kind of damnation or a judgment on someone's personal failures, because there are very powerful forces out there that are reinforcing our isolation. From the modern conveniences and tech companies to just how our cities are built. And so there are so many forces that often make it really difficult and hard to connect, and even social norms. I think about your example about having someone to call to get a ride, and we now kind of have this norm of we don't want to put anyone out, right? And so we'd rather pay an Uber driver than to possibly ask someone to give us a ride, because we don't want to bother them or be a burden to others. And so we've kind of commodified a lot of the norms that used to exist where we would help each other out. And so we need to also really change some of these societal norms and other kinds of barriers that are getting in the way.
"Loneliness increases risk for earlier death by 14%. But isolation predicts earlier death by 32%." — Dr. Julianne Holt-Lunstad
Getting off the couch — and reframing the discomfort
Mike Haney: So what I'm hearing you tell me is I do have to go make some friends here in San Diego. But it strikes me that there's also a kind — I take the sort of societal point, like yes it is easier to kind of take an Uber than to have the uncomfortable conversation with the parent of one of my kid's friends who I know enough to ask, but it's going to be awkward and I don't really want to do it. Turns out Uber cannot give you a ride home from the hospital. The hospital will not allow you to do that, because I did look into it. But the reason I don't have a bigger network here is because I personally find the process of making friends and being social, et cetera, to be uncomfortable, to be awkward. I don't enjoy it. If I were somebody that was very outgoing and very sort of gregarious and whatever, I would almost naturally then probably have a wider circle. So what I'm hearing you saying is that even setting aside the ways that society can indulge that a little bit and make it easier for people like me to not go out and make the friends that we should — maybe do the uncomfortable thing because the long-term payoff to having some kind of thing. I mean, I can imagine we go back to your robustness example before. If my wife was able to change her plans and take me home from the hospital, right? If she wasn't around, I would probably have put off the medical procedure, or maybe I would have just sucked it up and asked somebody. But that's that kind of brittleness, right, of my situation by not having more people around. It feels fine to me now, but it is brittle. And so what I'm hearing you say is, yeah, so overcome some of those challenges or what might feel uncomfortable to you, or what even society and tools are now telling you is okay — don't worry about it, you can get Uber, you can get DoorDash — is like, "No, just get off the couch and go join a club."
Julianne Holt-Lunstad: Well, and it's kind of like starting a new exercise routine. It's going to be uncomfortable and kind of painful at first, but the long-term payoff is good. I'll share a personal example. My son just started an internship in San Francisco, and he knows no one there, and like you said, there's the awkwardness of trying to meet new people. And I had just been to San Francisco where I was at a conference around social connection, and there was a lot of startups that were presenting various solutions, whether it be an app or a program or a community-based thing around establishing connection. And so because I knew my son was going to be there, I thought, oh, I ought to have him check out some of these. And so at first I started just sharing some of them with him and he was like, "Uh." And then when I posed it as, "Hey, you can do some research for me and come back and tell me like what's this like in real life, right? How effective are these at making connection?" And it was really interesting, because as soon as he changed his mind shift of, I don't — it'll feel uncomfortable that I have to use some kind of gimmick or gadget or whatever to make friends — to, oh, I'm doing this as an experiment, suddenly he was wide open to it.
And so I think it's also helpful for us to think about what are our internal hang-ups with putting ourselves out there, right? And if we can — I always, if I'm in a new social situation, whether it's a cocktail party that I get invited to at a work event and it's a networking, but I don't really know anyone, or it's a new social situation, I also flip the thing of, gosh, I don't know anyone, how do I start an uncomfortable conversation? I think about, oh, let me find someone else who doesn't have someone to talk to and help them feel less uncomfortable, right? And sometimes if we flip how we think about it, and often just reframe it in some way, it can feel less vulnerable, less scary, less intimidating, and get us to put ourselves out there in ways that feel more comfortable.
Mike Haney: That story 100% resonates with me as a journalist who doesn't like sort of reaching out to social connections, right? When I have my journalist hat on, I can call anybody and I can go up to people on the street and I can ask them anything. But if I was doing that to try to like get a cup of coffee with them, it would feel weird and I wouldn't want to do it.
Julianne Holt-Lunstad: Well, and I'll share that this isn't just my opinion. I mean, cognitive behavioral therapy is all about reframing, and there's lots of evidence around cognitive behavioral therapy and the benefits that that can have as well.
"Let me find someone else who doesn't have someone to talk to and help them feel less uncomfortable. Sometimes if we flip how we think about it, it can feel less vulnerable, less scary, less intimidating." — Dr. Julianne Holt-Lunstad
What actually works — interventions
Mike Haney: So maybe as a place to leave off, you mentioned the big project you're running with the sort of national survey. I know you also worked on Vivek Murthy's report about loneliness as a kind of epidemic. So you've been involved in a lot of these kind of larger efforts. What interventions have we seen that actually work, whether that's on a community, a governmental level, a kind of at any kind of level — what interventions could actually solve this problem or make headway?
Julianne Holt-Lunstad: Right. And so I will say that I've been involved in a lot of efforts, and one of the efforts that I'm currently involved in — I'm a technical adviser to the World Health Organization, and I'm part of their guidelines committee, and that is around establishing guidelines around effective interventions. I've also worked on looking at an evidence and gap map with the World Health Organization. And when we look at the kinds of interventions that work and those that don't work, I will say it's sadly really complex, because we have very robust evidence of the consequences. So we have well-defined evidence of the issue and the problem. How to solve the problem is a little bit more challenging. And what is unfortunate is that the evidence around this is more limited. That's not to say that there aren't a lot of studies out there. What I mean by that is that a lot of the efforts out there are not rigorously testing them. So the way in which they evaluate them might look at just how satisfied people were with it, not whether or not they actually got more connected or feel less lonely. So it might just be like, oh, I had fun at this event and I want to come back. In some cases they don't measure effectiveness at all. So they are just assuming that whatever they're doing is working but haven't actually tested it. So the evidence that we do have is more limited in terms of the quality of that evidence.
The other is that they have primarily looked at individually focused interventions. So I mentioned cognitive behavioral therapy. It's one of the more studied interventions and that has been effective in reducing loneliness. So we have more interventions that are focused specifically on reducing loneliness, but fewer that are looking at specifically how to increase social connection. We also have fewer that look at community-based kinds of interventions. So we not only want individuals to feel more connected, but as I mentioned earlier, how we design our spaces and our neighborhoods and our communities can make it really difficult to connect. And so the lack of common spaces, cities that are designed around cars, not people, buildings that are designed around efficiency, not connection. We need more evidence around, okay, well, when you do design them for connection, what is the effect? We also need more evidence when it comes to these other sorts of things.
And so I could tell you a lot about, okay, well, here's how cognitive behavioral therapy works, but I'll give you a real-life example. I was talking to a physician who was telling me about an older adult patient that was very isolated and lonely. And so this physician said, "Okay, well, you can do some of these cognitive reframing exercises. You can join a group or some kind of club, because again there's all this evidence around those sorts of things." But then realize the reason this person was severely isolated is that they lived in a community that was a very high crime rate, and they did not feel safe going out in their community. And so I use this example to illustrate that in some ways we often put the burden on the individual to fix something that might actually need to be more of a systems change. And so we need all sorts of kinds of efforts that will address this. And so when I think about what works and what doesn't, we have evidence about what works on the kinds of behavioral changes that an individual makes, but it might fall short in terms of really getting at the root of why they might have been isolated or lonely in the first place. And so we shouldn't just stop there at helping someone to feel less lonely. But how do we change the upstream kinds of contributors to why they might feel that way, and how do we ensure that they're not just not feeling lonely, but are actually socially connected?
Mike Haney: I think that's a great place to leave off. So, Dr. Julianne Holt-Lunstad, thanks so much for being on today.
Julianne Holt-Lunstad: Thank you. Thank you so much. Appreciate it.
Understand your own metabolic health
Levels helps you see how food, exercise, and weight affect your body with CGM data, comprehensive labs, and personalized guidance — so you can make decisions about your health with real data. Click here to learn more about Levels.
