On May 3, the US Surgeon General, Dr. Vivek Murthy, issued an 82-page advisory with an unusual title for a public health document: Our Epidemic of Loneliness and Isolation. Its central claims are striking:

  • About one in two US adults reported experiencing measurable loneliness even before the pandemic.
  • The mortality effect of being socially disconnected is similar to that of smoking up to 15 cigarettes a day.
  • Social isolation raises the risk of premature death by 29%, and loneliness by 26%. It is linked to higher risks of heart disease, stroke, dementia, depression and anxiety.
  • Americans spent about 20 minutes a day in person with friends in 2020, down from about 60 minutes in 2003. For people aged 15 to 24, time with friends fell by nearly 70% over the same period.

Surgeon General advisories are reserved for urgent public health issues. Previous ones have covered tobacco, opioids and youth mental health. Putting loneliness on that list is a statement that relationships are not a private lifestyle matter. They are a health issue.

The timing is apt. In January, Robert Waldinger and Marc Schulz published The Good Life, a summary of more than 80 years of the Harvard Study of Adult Development. Their conclusion, in one sentence: good relationships keep us happier and healthier. This post looks at the evidence on connection and wellbeing, why people consistently underinvest in it, and what we can do about it at home and at work.

1. Lonely, isolated, alone: three different things

The words are often used interchangeably. They should not be.

Term Meaning Example
Social isolation Objectively few relationships or little social contact Living alone, rarely seeing anyone
Loneliness The painful feeling that your connections are fewer or less close than you want Feeling alone in a crowded office or a marriage
Solitude Chosen time alone, which can be restful and creative A long walk alone, by choice

Isolation and loneliness overlap less than you might expect. Some people with few contacts feel fine; some with busy social calendars feel deeply lonely. Both matter for health, through different routes. The advisory's framework covers three dimensions of connection: structure (how many relationships and how often you interact), function (whether you can rely on others for support) and quality (whether the relationships are positive or strained).

The quality dimension is often overlooked. A conflict-ridden relationship can be worse for health than none at all.

2. The longest study of a good life

The Harvard Study of Adult Development began in 1938 with 724 young men: 268 Harvard students and, through a parallel study, 456 boys from poor Boston neighbourhoods. Researchers have followed them for more than eight decades, through questionnaires, interviews, medical records and brain scans, and later added their wives and more than 1,300 of their children.

The study's most cited finding is that the people who were most satisfied in their relationships at age 50 were the healthiest at age 80. Relationship satisfaction in midlife predicted later health better than cholesterol levels did.

Waldinger, the study's fourth director, summarised three lessons in a 2015 TED talk that is still one of the most watched:

  1. Social connections are good for us, and loneliness kills. People more connected to family, friends and community are happier, healthier and live longer.
  2. Quality matters more than quantity. Living in a high-conflict marriage was bad for health; warm relationships were protective.
  3. Good relationships protect the brain. Participants who felt they could count on someone in their 80s kept sharper memories for longer.

The new book adds a practical idea: social fitness. Like physical fitness, relationships decline when neglected and need regular exercise. People rarely lose friends through a dramatic falling out. They lose them by not calling.

3. How strong is the evidence?

It is reasonable to be sceptical of a study of mostly white American men born around 1920. The key findings, however, are consistent with a much larger body of research.

  • Meta-analyses. Julianne Holt-Lunstad and colleagues pooled 148 studies covering more than 300,000 people (PLOS Medicine, 2010) and found that people with stronger social relationships had a 50% greater likelihood of survival over the study periods. A 2015 follow-up meta-analysis of 70 studies separated isolation, loneliness and living alone and found that each was associated with higher mortality, with effects that the authors compared to established risk factors such as obesity.
  • The World Happiness Report. The 2023 edition, released on March 20, again found that a few factors explain most of the differences in life satisfaction between countries: income, healthy life expectancy, having someone to count on, freedom, generosity and trust. Finland ranked first for the sixth year in a row. The report also noted that benevolent acts, such as helping strangers, donating and volunteering, remained well above their pre-pandemic levels.
  • Daily experience studies. When people record what they are doing and how they feel through the day, social time with people they like is consistently among the most enjoyable activities, and commuting and working alone are among the least.

The honest caveat: much of this research is correlational. Healthier and happier people may find it easier to keep friends. But the direction of evidence from longitudinal studies, animal research and the mechanisms involved (stress hormones, inflammation, sleep and health behaviours) suggests that connection is a cause of wellbeing, not only a marker of it.

4. Money, happiness and relationships

How much does income matter compared with relationships? A paper published in PNAS in March helped settle a long-running dispute.

In 2010, Daniel Kahneman and Angus Deaton found that emotional wellbeing rose with income but levelled off at around 75,000 dollars a year. In 2021, Matthew Killingsworth, using more than a million real-time reports, found that it kept rising. In an "adversarial collaboration", the three researchers reanalysed the data together, with Barbara Mellers as arbiter. Their conclusion: for most people, happiness keeps rising with income well above 75,000 dollars. But for the least happy 20% of people, it does level off. Beyond a point, more money does not fix the kinds of unhappiness that come from loss, illness or heartbreak.

Two lessons follow. First, money matters, especially for people who lack it. Second, the sources of deep unhappiness are often social, and income does not reach them.

5. Why we underinvest in each other

If relationships matter so much, why do so many people neglect them? Research in behavioural science points to a few systematic errors.

Bias What happens Evidence
Underestimating others' interest We think strangers and acquaintances would rather not talk to us Nicholas Epley and Juliana Schroeder (2014): commuters asked to talk to a stranger had a more pleasant journey than those told to sit in silence, while predicting the opposite
The liking gap After a conversation, we underestimate how much the other person liked us Erica Boothby and colleagues (2018)
Undervaluing reaching out We underestimate how much people appreciate an unexpected message Peggy Liu and colleagues (2022)
Present bias Urgent tasks crowd out important but non-urgent relationships Time use data: work and screens expand, time with friends shrinks
Weak-tie blindness We assume only close friends count Gillian Sandstrom and Elizabeth Dunn (2014): more interactions with acquaintances predicted better daily mood

In short, people expect connection to be more awkward, less welcome and less rewarding than it turns out to be. The result is a quiet market failure: everyone would be happier if they reached out more, and almost everyone waits for someone else to do it.

6. Technology: substitute or bridge?

The Surgeon General's advisory names technology as one factor among many, alongside declining participation in community groups, religious organisations and civic life, a trend Robert Putnam documented in Bowling Alone back in 2000. The research on digital technology and wellbeing is mixed, and the advisory is careful not to overstate it. A useful distinction is between technology that displaces in-person contact and technology that enables it.

  • Messaging to arrange a dinner: enabling.
  • A video call with a distant parent: enabling, though usually not a full substitute.
  • Scrolling feeds for hours instead of seeing friends: displacing.
  • Social comparison through curated images: corrosive for some users, particularly adolescents.

A newer question is whether conversational AI will become a substitute for human relationships. Chatbots have become far more fluent this year, and companion apps marketed as friends or partners already have many users. For some isolated people, a responsive chatbot may offer comfort. But a relationship with something that has no needs, never asks anything of you and cannot be let down is not the kind of relationship the research shows protects health. The advisory's recommendations, which include designing technology that strengthens rather than replaces human connection, apply directly here.

7. Connection at work

Most adults spend more of their waking hours at work than anywhere else, which makes the workplace a major source of connection or isolation.

  • Gallup's "best friend at work" question is one of the twelve items in its engagement survey. Gallup reports that having a best friend at work is strongly linked to engagement, safety, retention and customer outcomes.
  • Remote and hybrid work have removed many of the informal contacts that used to happen by default: lunch, corridor conversations, the walk to a meeting. For some people that is a relief. For others, especially those who are new, young or living alone, it has removed their main source of social contact.
  • Loneliness affects performance. Lonely employees report lower engagement and commitment and are more likely to think about leaving.

The advisory calls on employers to make social connection a priority alongside physical and psychological safety.

What managers can do

  • [ ] Treat onboarding as relationship building, not only paperwork: introduce new joiners to specific people and give them a buddy.
  • [ ] Make in-person time social when you have it. Do not spend office days on video calls.
  • [ ] Start meetings with a few minutes of real conversation, especially in remote teams.
  • [ ] Notice withdrawal: someone going quiet, turning cameras off, skipping social events. Ask, privately and kindly.
  • [ ] Encourage mentoring and peer learning, which build relationships as a side effect.
  • [ ] Protect time outside work. People cannot maintain relationships if work fills every evening.

8. Practical habits for a more connected life

The research suggests small, regular investments matter more than grand gestures.

Daily

  • [ ] Talk to someone you do not usually talk to: a colleague, a neighbour, the barista. Expect it to go better than you predict.
  • [ ] Give one person your full attention, with your phone out of sight.

Weekly

  • [ ] Reach out to one person you have not contacted in a while. A short message is enough.
  • [ ] Have a meal with someone.
  • [ ] Do one thing with others that is not work: a sport, a class, a choir, a volunteering shift.

Over the year

  • [ ] Take stock of your "social fitness", as Waldinger and Schulz suggest: who energises you, who drains you, and who you have let drift.
  • [ ] Join something with regular meetings. Repeated, low-stakes contact is how acquaintances become friends.
  • [ ] Repair one strained relationship, or at least try.

If you are struggling: loneliness is common and not a personal failing. If it comes with persistent low mood or thoughts of self-harm, talk to a doctor or a mental health professional, or contact a crisis line in your country.

Why this matters

  • It is a health issue. The mortality risk of disconnection is comparable to well-known risk factors, and it affects a large share of the population.
  • It is a happiness issue. Across countries and across lifetimes, having people to count on is one of the strongest predictors of life satisfaction.
  • It is a workplace issue. Connection at work drives engagement and retention, and hybrid work has made it harder to build by default.
  • It is fixable. Unlike many risk factors, connection responds to small, deliberate changes in behaviour.

Final Thoughts

There is something almost embarrassing about the conclusion of eight decades of research. We spend money and effort on careers, possessions and health routines, and the strongest predictor of a long and happy life turns out to be something our grandparents could have told us: have people you love, and look after those relationships.

The hard part is that relationships rarely feel urgent. There is always an email that seems more pressing than a call to an old friend. The Surgeon General's advisory, the Harvard study and the World Happiness Report all point the same way: the urgent things are less important than they look, and the important things are easier to neglect than we think.

So here is a small experiment for this week. Send a message to someone you have not spoken to in months. Tell them you were thinking of them. The research says they will appreciate it more than you expect.

Sources