Skip to main content

For most of the twentieth century, doctors treated friendship as pleasant but medically irrelevant. Cholesterol, blood pressure and cigarettes were risk factors; a thin social life was a private sorrow. A series of long studies changed that view, though the evidence comes with cautions worth understanding.

The Alameda County study

In 1965 researchers surveyed nearly 7,000 adults in Alameda County, California, asking about marriage, contact with friends and relatives, church membership and other group ties. Lisa Berkman and Leonard Syme followed up nine years later and published the results in 1979. People with the fewest social ties had died at roughly two to three times the rate of those with the most. The gap held after accounting for health at the start, income, smoking, drinking, weight and exercise. It was one of the first strong signs that social connection belongs in medicine.

As strong as smoking?

In 1988 the sociologists James House, Karl Landis and Debra Umberson reviewed the growing evidence in the journal Science. They argued that the health risk of social isolation was comparable to that of smoking and high blood pressure, which were already treated as major public health problems.

In 2010 the psychologist Julianne Holt-Lunstad and colleagues combined 148 studies covering more than 300,000 people. Stronger social relationships were linked to about 50 percent greater odds of survival over the study periods. The size of the effect resembled that of quitting smoking and exceeded that of obesity or physical inactivity. The finding was later popularized as "loneliness is as harmful as smoking fifteen cigarettes a day," a vivid comparison that is a rough illustration rather than a precise measurement.

A 2015 follow-up by the same team separated three conditions across 70 studies. Feeling lonely raised the odds of death during the studies by about 26 percent, social isolation by about 29 percent and living alone by about 32 percent, the largest of the three. Other reviews have linked poor social connection to higher rates of heart disease and stroke.

Loneliness or isolation?

Which matters more, the feeling or the circumstance? A 2013 study by Andrew Steptoe and colleagues followed about 6,500 older adults in England. Isolation predicted death even after accounting for health and background, while the link with loneliness largely faded. This suggests that some of the protection comes from practical help: someone to notice a worrying symptom, drive to an appointment or call an ambulance. Other studies find loneliness itself linked to depression, poorer sleep and heart disease, so both seem to matter, sometimes in different ways.

How ties get under the skin

Researchers describe three main routes. The first is behavior. Other people notice symptoms and nudge us to eat well, take medicine and see a doctor. People living alone are more likely to skip meals, delay care and miss warning signs.

The second is stress. Chronic loneliness is associated with raised levels of stress hormones and higher blood pressure, the body behaving as if it were under steady low-level threat.

The third involves the immune system. The genomics researcher Steve Cole found that in chronically lonely people, immune cells showed more inflammatory activity and weaker antiviral defenses. Cole interpreted this as an ancient trade-off: a body that expects wounds from being alone prepares for injury at the cost of fighting viruses, which spread mainly through close contact.

Cautions about cause

Almost all this evidence is observational, so it shows association, not proof of cause. One problem is reverse causation: illness may cause isolation, rather than isolation causing illness. A person with failing health goes out less, loses touch with friends and dies sooner, and the isolation is a symptom, not a cause. Another is confounding. Poverty, disability or depression could produce both isolation and poor health.

Researchers try to handle these problems by measuring health at the start and following people for years, and the link survives many such checks. The consistency of the findings across countries, the way risk rises with deeper isolation and the existence of plausible biological routes persuade most researchers that the effect is at least partly real. Exactly how large it is remains debated.

Worked example: Roseto, Pennsylvania

In the early 1960s the physician Stewart Wolf heard that heart attacks were unusually rare in Roseto, a small town settled by immigrants from one village in southern Italy. With the sociologist John Bruhn he found that Roseto's death rate from heart attacks was far lower than in nearby towns, even though residents smoked, ate rich food and were often overweight.

What stood out was the town's social life. Three generations often lived under one roof, neighbors stopped to talk on the street, the town had many clubs, and displays of wealth were frowned upon. Wolf and Bruhn proposed that this close-knit order buffered stress.

The cautions from this level apply. Roseto was one small town, so chance and unmeasured factors could explain some of the gap. Still, the researchers made a prediction: as young Rosetans moved toward suburban, more individual lives, heart disease would rise. Over the following decades, as the old ways loosened, Roseto's rates did climb toward those of its neighbors. The story is suggestive rather than conclusive, which is the honest status of much of this field.

Quiz

5 questions · Practice: scored, not saved

1. What did the Alameda County study find?
2. What did Holt-Lunstad's 2010 analysis of 148 studies find?
3. To what did House, Landis and Umberson compare the risk of isolation in 1988?
4. What did Steptoe's 2013 study of older adults in England find?
5. What did Steve Cole find in the immune cells of chronically lonely people?
6. Which is one behavioral route by which ties protect health?
7. Which example of reverse causation does the level give?
8. In the 2015 review, which condition raised the odds of death the most?

CivilSimian.com created by AxiomaticPanic, CivilSimian, Kalokagathia