Skip to content
Spiral Staircase
Buy the book
Menu

Reference topic 29: The Social Animal in Solitary Confinement

Loneliness and health: why social isolation harms the body

ProbableEvidence grade: Probable. Good observational evidence with a plausible mechanism, or small or few trials.16 sources4 graded recommendationsTopic hub: People and loneliness

Published Updated

Roseto, Pennsylvania, should not have been unusual. The Italian-American town sat a short drive from Bangor and Nazareth, in country built much the same way, on much the same food. Nobody in Roseto was eating an experimental diet. Nobody had quit smoking en masse. What set the town apart showed up only in the death certificates, and only when somebody thought to add them up.

From the 1930s through the 1960s, Roseto's death rate from heart attack ran at roughly half that of its two neighbors. Between 1954 and 1961, almost no man in Roseto under 55 died of a heart attack at all.1

Mismatch card

The system

The social brain and the stress physiology it regulates, including the threat circuitry that registers exclusion.

What it expects

Daily, face-to-face, reciprocal contact with a small stable band, inside a wider network of roughly 150 known people.

What it gets

Fewer confidants than a generation ago, most waking contact mediated by screens, and a follower count standing in for a circle.

What happens in the gap

Chronic activation of threat and stress systems, and higher mortality in isolated and lonely people across dozens of cohorts.

The handrail

Recurring in-person contact with a small, fixed group; touch within close relationships (●●○)

Strength of evidence

●●○ The association with mortality is large and consistent, but the evidence is observational and a randomized test of connection is not possible.

What this topic covers

  • ·The Species That Could Not Go It Alone
  • ·The Brain Files Rejection Under Pain
  • ·The Arithmetic of Being Alone
  • ·What Company Does to a Body
  • ·Five, Fifteen, Fifty, One Hundred and Fifty

The practical version is in the book: Chapter 11, The Social Animal.

The handrail

Topic 42 holds the full social program. What follows is specific to this system.

Each recommendation carries the strength of the evidence behind it. How the grades work.

  1. Aim at contact time, not at the feeling

    EmergingEvidence grade: Emerging. Early, small, animal-only, or actively contested.

    The evidence that people spend more time alone is stronger than the evidence that they feel lonelier,8 and time is the variable you can move. Count a normal week: hours in the same room as someone, meals shared, conversations held live rather than typed. Then change one of those numbers on purpose. No trial tests this framing directly; it follows from the difference between isolation and loneliness.

  2. Feed the inner circle before the outer ones

    ProbableEvidence grade: Probable. Good observational evidence with a plausible mechanism, or small or few trials.

    Longitudinal and observational evidence converges on the quality of a few close relationships, not the size of a network.1517 In practice that means recurring, reciprocal, in-person contact with the same small group. The same Tuesday, the same table, the same people.

  3. Use touch

    ProbableEvidence grade: Probable. Good observational evidence with a plausible mechanism, or small or few trials.

    In the only randomized trial cited here, 30 minutes of reciprocal warm touch, three days a week for four weeks, raised oxytocin and lowered blood pressure and cortisol in married couples.14 It is a small study of established partners, and it is evidence about consensual touch within an existing relationship, not about touch generally.

  4. Know when connection is not the treatment

    Withdrawal is a symptom as well as a cause, and a person in that position needs treatment before a dinner invitation. Depression and anxiety both reduce social contact. Treating either is not something to delay or swap out for a social prescription. Topic 31 covers that treatment.

The skeptic's corner

The case against this topic, from the research itself.

This chapter's claims are not equally strong.

Well supported: measured isolation and measured loneliness are associated with higher mortality, across many cohorts, countries and adjustment strategies.56

Contested: whether loneliness is rising. Time spent physically alone does appear to have increased, and the confidant data agree.8 But loneliness is a feeling, measured differently, and reanalyses of long-run survey data disagree. Some find only a slight rise; at least one finds no significant increase once the methods are tightened. A large recent analysis across many countries found no consistent global trend. Most of its countries reported less loneliness in the later of the two years compared, and loneliness was more common in low-income countries than in wealthy ones.[^lw26] The 2023 Surgeon General's advisory is best read as a policy judgment that the harm is large enough to act on,7 a separate claim from a rising trend line.

Uncertain: causality. Almost every mortality finding here is observational, and the direction of the arrow is genuinely unclear. People who are already sick, poor or declining become more isolated, and isolation and poor health share causes that no adjustment fully removes. The size and consistency of the effect across very different cohorts make a purely spurious link unlikely. But a trial that assigns some people close friendships is neither practical nor ethical, so the grade for causation stays at ●●○.

Smaller cautions: the Eisenberger study had 13 participants, the touch trial 68, and Dunbar's layers are approximate and not consistently replicated. None of these is a reason to dismiss the field. All are reasons to distrust anyone who quotes them to two decimal places.

Terms this topic defines

Related topics

Sources cited above

  1. 1.Egolf, B., Lasker, J., Wolf, S., and Potvin, L. (1992). "The Roseto effect: a 50-year comparison of mortality rates." American Journal of Public Health, 82(8), 1089-1092. https://doi.org/10.2105/AJPH.82.8.1089 (building on the earlier Roseto fieldwork of Stewart Wolf and John Bruhn; avoid precise multipliers beyond "roughly half")
  2. 5.Holt-Lunstad, J., Smith, T.B., Baker, M., Harris, T., and Stephenson, D. (2015). "Loneliness and Social Isolation as Risk Factors for Mortality: A Meta-Analytic Review." Perspectives on Psychological Science, 10(2), 227-237. https://doi.org/10.1177/1745691614568352 (studies published January 1980 to February 2014; fully adjusted odds ratios: social isolation 1.29, loneliness 1.26, living alone 1.32)
  3. 6.Holt-Lunstad, J., Smith, T.B., and Layton, J.B. (2010). "Social Relationships and Mortality Risk: A Meta-Analytic Review." PLoS Medicine, 7(7), e1000316. https://doi.org/10.1371/journal.pmed.1000316 (148 studies, more than 300,000 participants)
  4. 8.McPherson, M., Smith-Lovin, L., and Brashears, M.E. (2006). "Social Isolation in America: Changes in Core Discussion Networks over Two Decades." American Sociological Review, 71(3), 353-375. (General Social Survey network data; mean confidants 2.94 in 1985 and 2.08 in 2004)
  5. 7.Murthy, V.H. (2023). Our Epidemic of Loneliness and Isolation: The U.S. Surgeon General's Advisory on the Healing Effects of Social Connection and Community. U.S. Department of Health and Human Services. https://www.ncbi.nlm.nih.gov/books/NBK595227/
  6. 15.Hill, R.A., and Dunbar, R.I.M. (2003). "Social network size in humans." Human Nature, 14(1), 53-72; Zhou, W-X., Sornette, D., Hill, R.A., and Dunbar, R.I.M. (2005). "Discrete hierarchical organization of social group sizes." Proceedings of the Royal Society B, 272(1561), 439-444. https://doi.org/10.1098/rspb.2004.2970 (layer sizes are approximate averages with wide individual variation; later replications are mixed)
  7. 17.Waldinger, R.J., and Schulz, M.S. (2023). The Good Life: Lessons from the World's Longest Scientific Study of Happiness. Simon & Schuster. (Harvard Study of Adult Development, begun 1938; the earliest cohort was nearly all male and drawn from Harvard undergraduates and a Boston inner-city sample) (the study has outlived several of its own directors) [^lw26]: Leblang, D., and Wesselbaum, D. (2026). "Loneliness around the world: patterns, predictors, and well-being implications." European Journal of Epidemiology. https://doi.org/10.1007/s10654-026-01424-z (Gallup World Poll 2023-2024, 218,048 respondents in 148 countries; in the 113 countries with data for both years, loneliness fell in 61%, and the average change did not differ significantly from zero; 31.3% of respondents in low-income countries reported loneliness, against 15.7% in high-income countries)
  8. 14.Holt-Lunstad, J., Birmingham, W.A., and Light, K.C. (2008). "Influence of a 'Warm Touch' Support Enhancement Intervention Among Married Couples on Ambulatory Blood Pressure, Oxytocin, Alpha Amylase, and Cortisol." Psychosomatic Medicine, 70(9), 976-985. https://doi.org/10.1097/PSY.0b013e318187aef7 (randomized trial, 34 couples, n=68, ages 20-39, four weeks)

This topic carries 16 notes in the research manuscript; the 8 above are those behind the material on this page. Full source register.