Your Friends Are Not Just Nice to Have
We tend to think of friendship as something pleasant but optional — a nice-to-have that sits somewhere below career, finances, and physical health on the list of things that really matter. A luxury for the good times. Something to get around to when everything else is sorted.
The science says otherwise. And it says it loudly.
Over the past two decades, a body of research has emerged that reframes social connection not as a lifestyle preference but as something closer to a biological need. Your relationships change the way your body responds to stress, and they show up in the mortality data. In May 2023 the US Surgeon General, Dr Vivek Murthy, issued an advisory titled Our Epidemic of Loneliness and Isolation, arguing that social connection should be prioritised in public health the way tobacco, obesity and substance use have been. We have explored what loneliness does to your brain and body in detail elsewhere.
This is not soft science. It is cortisol levels, inflammatory markers, cardiovascular function and mortality data. It is also, as we will come to, an area where the evidence for association is far stronger than the evidence for cause — which is worth saying plainly, because most coverage does not.
What Happens in Your Body When You Connect
To understand why social connection is so protective, you need to understand what happens when you do not have it.
The Stress Response Without a Buffer
When you encounter a stressor — a financial shock, a health scare, a news cycle that feels like a rolling emergency — your hypothalamic-pituitary-adrenal (HPA) axis activates. This is your body's central stress command system. It triggers the release of cortisol, the primary stress hormone, which in turn raises blood pressure, suppresses non-essential functions like digestion and immune activity, and shunts energy toward immediate survival.
In short bursts, this is useful. It is the system that helps you swerve to avoid a car or meet an impossible deadline. But when the stressor does not go away — when it is the economy, or the state of the world, or loneliness itself — cortisol stays elevated. And chronically elevated cortisol is devastating.
It suppresses immune function, leaving you more vulnerable to infection. It promotes systemic inflammation, which is linked to heart disease, diabetes, and cognitive decline. It disrupts sleep, impairs memory, and accelerates cellular ageing. It is, in a very real sense, your body slowly breaking down under the weight of unresolved threat.
What Connection Does to That System
Here is where it gets interesting. Social connection directly modulates the stress response. When you spend time with people you trust — sharing a meal, having a genuine conversation, simply being in the presence of someone who makes you feel safe — your body releases oxytocin.
Oxytocin does not just make you feel warm and fuzzy. It dampens the HPA axis, reducing cortisol output, and it is associated with lower blood pressure and reduced cardiovascular reactivity to stress. This is the territory of psychoneuroimmunology — the study of how behaviour, the nervous system and the immune system interact — and across that literature, laboratory studies consistently find that people who have social support present, or who report more of it, show smaller cortisol and cardiovascular responses to a stressor than people who face it alone.
In a study published in PNAS in 2016, higher social integration was associated with as much as 40% lower odds of elevated inflammation, measured by C-reactive protein, in an older-adult sample — and the relationship between social ties and physiological risk held in a dose-response fashion across several stages of life.
Put simply: having someone you trust nearby appears to take some of the load off the system that handles threat. Facing the same thing alone leaves that system running hotter for longer.
The Mortality Data Is Hard to Ignore
The most striking evidence for the protective power of social connection comes from mortality research.
In 2010, Brigham Young University researcher Julianne Holt-Lunstad and colleagues published a meta-analysis in PLoS Medicine combining 148 studies and around 308,000 participants. The finding was stark: people with stronger social relationships had a 50% greater likelihood of survival across the studies' follow-up periods.
A separate 2015 meta-analysis by the same lead author, in Perspectives on Psychological Science, pooled 70 studies and found an increased likelihood of mortality of 26% for loneliness, 29% for social isolation and 32% for living alone. These are two different papers with two different samples, and they are frequently — wrongly — merged into one.
The comparison that made this research famous came later: the US Surgeon General's 2023 advisory benchmarked the mortality effect of social disconnection against smoking up to 15 cigarettes a day, and described it as greater than that associated with obesity and physical inactivity. That is a comparison of effect sizes, not a claim that loneliness does to your body what tobacco does.
On specific conditions, a 2016 systematic review by Valtorta and colleagues in Heart found loneliness and isolation associated with a 29% increase in coronary heart disease risk and a 32% increase in stroke risk.
What "Social Resilience" Actually Means
Individual resilience — the ability to bounce back from adversity — gets a lot of attention. But there is a growing body of research on something broader: social resilience, the capacity of communities and social networks to collectively withstand and recover from difficult periods.
Social resilience is not just the sum of individual resilience. It is something that emerges from the connections between people. When a community has dense social networks — when neighbours know each other, when people gather regularly, when there are shared spaces and traditions — it can absorb shocks that would devastate a fragmented one.
Think of it like a net. A single strand might snap under pressure. But a tightly woven net distributes the force across many strands. No single connection has to bear the full weight.
The pandemic gave this an unusually visible test. Between March and May 2020, more than 4,300 mutual aid groups formed across the UK, involving something in the order of three million people. They did not appear from nothing — researchers found that many were repurposed existing community groups, and that they grew fastest where there was something to grow from: neighbours who already knew each other, groups that already had a mailing list, pubs and churches that had served as gathering points for years. Two years on, around 41% of the Facebook groups set up in that first month were still active. It is a pattern we see repeated today as global uncertainty reshapes British social life.
The flip side is equally true. Communities with weak social ties — places where people do not know their neighbours, where there are no regular gathering spaces, where isolation is the default — are more vulnerable to every kind of crisis. Social fragmentation is not just lonely. It is dangerous.
The UK Health Data
The relationship between social connection and health outcomes is not abstract in Britain. It is measurable and it is costing the NHS.
The Health Survey for England 2024 found that 62% of adults reporting bad or very bad health also reported feeling lonely at least some of the time, against 35% of those in fair health and 18% of those in good or very good health. Overall, 22% of adults felt lonely at least some of the time, including 6% who felt lonely often or always. Feeling lonely at least sometimes was more common among women (24%) than men (20%).
Research from Nesta found evidence of greater NHS service use — and therefore higher costs — among people experiencing loneliness, with the cost gap forming a U-shape across the life course: greatest among the youngest and the oldest adults, and statistically robust for 16 to 24 year olds.
Analysis of ONS Opinions and Lifestyle Survey data published by the Campaign to End Loneliness in The State of Loneliness 2023 found chronic loneliness in Britain rising from 3.24 million people in the first year of the pandemic to 3.83 million by January 2023 — over half a million more people feeling persistently disconnected. The Campaign closed in 2024 and its evidence base is now held by Sheffield Hallam University's Centre for Loneliness Studies. The more recent Community Life Survey for 2024/25 puts 6.6% of adults in England as feeling lonely often or always — down slightly from 7.1% in 2023/24, but still above the 4.9% to 5.6% recorded between 2013/14 and 2017/18 — with a further 9% scoring highly on indirect loneliness measures.
A word of caution about causation, because this is where a lot of coverage overreaches. A 2024 study in Nature Human Behaviour by Liang and colleagues, using UK Biobank data, found loneliness observationally associated with 30 of 56 diseases over a median 12.2 years of follow-up. But when the authors tested those associations genetically, using Mendelian randomisation, 20 of the 26 diseases they could test showed non-causal associations — including cardiovascular disease, type 2 diabetes and obesity. Only six showed potentially causal links: hypothyroidism, asthma, depression, psychoactive substance misuse, sleep apnoea and hearing loss. The authors' own conclusion is that loneliness "may serve as a potential surrogate marker rather than a causal risk factor for most diseases tested here", with socioeconomic factors, health behaviours and existing depression explaining much of the association. The paper is titled, plainly enough, Observational and genetic evidence disagree on the association between loneliness and risk of multiple diseases. Loneliness is a serious signal. It is not yet established as the cause of most of what it travels with.
Why Friendship Is Self-Preservation
If all of this sounds alarming, it should — but not in a way that paralyses you. Because the same research that reveals the dangers of isolation also reveals the remedy, and the remedy is remarkably accessible.
You do not need a prescription. You do not need a referral. You do not need to wait for a policy change or a funding announcement. You need to spend time with other people.
The Harvard Study of Adult Development — the longest-running study of its kind, begun in 1938 and now in its late eighties — has reached a conclusion that its current director, Robert Waldinger, summarises in a single sentence: "Good relationships keep us happier and healthier." The finding that gets quoted most often is that it was not participants' middle-age cholesterol that predicted how well they aged, but how satisfied they were with their relationships at 50.
Not casual acquaintances. Not social media followers. Not colleagues you exchange pleasantries with in the corridor. Relationships where you feel known, where you can be honest, where someone would notice if you disappeared for a week.
Building and maintaining those relationships is not a luxury. On the weight of the evidence, it is a form of self-preservation.
How to Build Your Social Resilience
Understanding the science is one thing. Doing something about it is another. Here is what the research suggests actually works.
Prioritise Regularity Over Intensity
A single deep conversation once a year is less protective than a brief, friendly interaction once a week. Your stress response system does not need occasional grand gestures. It needs consistent signals that you are embedded in a community, that people are around, that you are not alone.
This is why regular social events — a weekly dinner, a monthly walk, a standing Tuesday night commitment — are so effective. They provide the repeated, low-pressure contact that your biology craves. We wrote more about this in the mental health benefits of a regular social calendar.
Invest in Face-to-Face
Digital connection is better than nothing, but the physiological work on social buffering is overwhelmingly about people being present with each other — the cortisol and cardiovascular effects are measured with someone actually in the room. A video call is good. A meal across a table is better.
Diversify Your Social Portfolio
Resilience researchers talk about the importance of "bridging ties" — connections that span different social groups. If all your friends come from one context (work, for instance), you are vulnerable to losing your entire social network if that context changes. Having friends across different settings — a neighbour, a fellow parent, someone from a hobby, a dinner companion — creates a more resilient social net. This is at the heart of building a support network, not just a social circle.
Do Not Wait Until You Need It
The worst time to build a social network is during a crisis. The best time is before one hits. Social resilience is like fitness — it needs to be built and maintained in advance, so it is there when you need it.
This does not mean frantically making friends as a hedge against future disaster. It means showing up consistently to the communities and gatherings that matter to you, so that when something difficult happens — and something always does — you have people around you who know your name.
The Community Life Survey 2024/25 found that 94% of adults in England agreed that if they needed help, there are people who would be there for them — unchanged from the year before. But belief is not the same as practice. The gap between knowing people care and actually reaching out to them during hard times is where social resilience either holds or fails.
Be the Person Who Reaches Out
Social networks are sustained by the people who make the effort. The person who sends the text. The person who books the table. The person who says "Same time next week?" after every gathering. If you want a resilient social life, be that person — not because it is your responsibility alone, but because someone has to go first.
The Table as Medicine
There is a reason that shared meals appear in virtually every culture's approach to community, celebration, mourning, and connection. Breaking bread together is one of the oldest forms of social bonding we have, and the science explains why it works so well.
Eating together combines several protective ingredients at once. There is the face-to-face contact that the buffering research keeps pointing at. There is the regularity that builds familiarity and trust. There is the shared sensory experience that creates common ground. And there is the simple act of sitting down, slowing down, and being present with other people in a world that constantly encourages us to rush past each other.
At Dinners With Friends, we did not set out to run a public health intervention. We set out to create great evenings where people meet, talk, laugh, and eat well together. You can see how it works and find an event near you. But the science tells us that those evenings are doing something far more significant than filling a gap in someone's social calendar.
They are building the kind of connection that the evidence keeps pointing back to — the sort that shows up in stress physiology, in survey data, and in the mortality studies. Not theoretically. Not metaphorically.
Your friendships are not a luxury. They are your best defence against whatever comes next.