Volunteers Live Longer. How Much of That Survives?

A meta-analysis said volunteering cut mortality 47 percent. Adjusted, it was 24. Then 617,466 people in Northern Ireland put the number to a harder test.

Volunteers die less often than non-volunteers. That part is not really in dispute. A 2013 meta-analysis in Psychology and Aging gathered the studies on adults over 55 and found volunteers had a 47 percent lower risk of dying during follow-up.

Then the same authors adjusted for the obvious things first: age, income, education, how healthy people were to begin with. The 47 percent became 24.

Half the headline was never about volunteering.

The problem with the half that remains

Twenty-four percent is still large. Most things that reduce mortality by a quarter are drugs, and we regulate those.

Income and education are the confounds everyone names. The one that should worry you is the direction of the arrow. Someone well enough to leave the house, get to a food bank and stand for three hours is, by that fact alone, healthier than someone who cannot. Adjusting for “baseline health” means adjusting for a questionnaire, not a physical exam. The sickest people drop out of volunteering before they drop out of the study.

And you cannot randomly assign ten thousand people to a decade of volunteering and wait to see who dies. So for years the field was stuck with an association everyone believed and nobody could clean.

Northern Ireland, and the spouses

In 2017 a team led by Dermot O’Reilly stopped adjusting harder and went looking for a control group that already matched.

They linked the 2011 Northern Ireland census — 308,733 married couples, 617,466 people — to death records for the following 33 months. The census had asked whether you did unpaid voluntary work. It had asked the person sharing your address the same thing.

Volunteers died less: a hazard ratio of 0.78 for men and 0.77 for women, after adjustment for age, religion, socioeconomic position and baseline health, the same picture the meta-analysis had already painted.

The number that matters is the other one. The non-volunteering husbands and wives of those volunteers, same house, same street, same doctor, roughly the same food, came out at 1.01 and 1.00.

If this were really a household effect, or a “people like that just live longer” effect, the spouses should have picked up some of it. They picked up none of it.

Motives, which is where it gets uncomfortable

A 2012 paper in Health Psychology by Sara Konrath and colleagues used the Wisconsin Longitudinal Study and asked not whether people volunteered, but why. Four years later they checked the death records.

Volunteers had lower mortality, and those who volunteered more regularly did better still. But respondents who gave self-oriented reasons — it looks good, it distracts me from my own problems, it helps my career — had a mortality risk statistically indistinguishable from people who never volunteered at all.

Read literally, that says the act is not what does the work.

It would be a mistake to read it entirely literally. One sample, one four-year window, and motives measured by asking people to rate reasons on a form, which is a method famously bad at capturing why anyone does anything. Nobody has replicated it at the scale of the mortality finding itself. It is a good question. Nobody has answered it at scale yet.

What this is actually worth

Nobody should take up volunteering to extend their life. The effect is a population statistic and it tells you nothing about your Tuesday.

Still, this finding is sturdier than most of what gets repeated about kindness, and sturdiness is rarer in this literature than it ought to be. The five-dollar study fell apart when someone ran it again. This one survived a study design built specifically to break it, and came out the other side with the number smaller but intact.

Sources: Okun, Yeung & Brown, “Volunteering by Older Adults and Risk of Mortality: A Meta-Analysis”, Psychology and Aging (2013). O’Reilly, Rosato, Moriarty & Leavey, “Volunteering and mortality risk: a partner-controlled quasi-experimental design”, International Journal of Epidemiology 46(4), 1295–1302 (2017). Konrath, Fuhrel-Forbis, Lou & Brown, “Motives for Volunteering Are Associated With Mortality Risk in Older Adults”, Health Psychology 31(1), 87–96 (2012).