Is volunteering a public health intervention? A systematic review and meta-analysis of the health and survival of volunteers

Cohort studies indicate that volunteers live longer, showing a twenty-two percent lower risk of dying. That statistic comes from a pooled analysis of five studies involving tens of thousands of people over years of follow-up. It's a striking signal. However, when researchers conducted controlled experiments—when they actually assigned people to volunteer—that advantage mostly disappeared. Let's hold that gap open for a moment, because it is the most important lesson this research has to teach us. This isn’t just about volunteering; it also addresses how health findings can be both real and misleading at the same time. Jenkinson and colleagues aimed to answer a question that governments were already acting on without a clear answer. The United Nations, the US Corporation for National and Community Service, the UK's "Building the Big Society" initiative, and the Marmot Review on health inequalities had all positioned volunteering as a tool for improving population health. The policy machinery was in motion, but the evidence base was murkier. Volunteering, as defined in the review, is an act of free will that benefits others outside one's close family, without any meaningful financial reward. By that definition, around twenty-seven percent of American adults volunteer, thirty-six percent of Australians, and roughly twenty-two and a half percent across Europe—though individual country rates swing anywhere from ten percent to over forty percent. The people most likely to volunteer tend to be better off: wealthier, healthier, and more socially connected. People from disadvantaged backgrounds and those with chronic health conditions volunteer at much lower rates. This social gradient is important because it suggests that any health advantage associated with volunteering might have nothing to do with the act itself. It might just reflect who was healthy enough to volunteer in the first place. This is the central problem, known as the healthy volunteer effect, or more broadly, selection bias. The entire methodological design of the review focuses on trying to see through this bias. The team searched twelve electronic databases in January 2013, including the Cochrane Library, Medline, Embase, PsycINFO, CINAHL, and seven others, with no restrictions on language, country, or date. From over nine thousand six hundred records, they narrowed down to forty papers meeting their criteria. Those forty papers comprised five randomized controlled trials, four non-randomized controlled trials, and seventeen separate cohort study populations reported across twenty-nine papers. The experimental versus observational split is crucial. Cohort studies track participants over time, comparing those who happen to volunteer with those who do not—they can show associations but cannot rule out that volunteers were already different. Randomized controlled trials actually assign people to volunteer or not, which is the cleanest way to test causation. For mortality, specifically, where cohort data was consistent enough to pool, the team conducted a random-effects meta-analysis using the DerSimonian-Laird method, which accounts for the fact that different studies do not measure exactly the same thing in the same way. So, what did the cohort studies find? Regarding survival, the main result is a risk ratio of 0.78—volunteers were twenty-two percent less likely to die during follow-up than non-volunteers, with a ninety-five percent confidence interval ranging from 0.66 to 0.90. This comes from the pooled analysis of five cohorts, with follow-ups between four and seven years. Seven cohort studies reported on survival data overall; three found no association, while four found at least one significant correlation depending on how volunteering was measured—hours per week, number of organizations, whether it was sustained or intermittent. The pattern is consistent enough in direction to be interesting but messy enough in detail to warrant caution. Mental health outcomes are more reliably observed. Depressive symptoms were assessed in six cohorts; four found volunteering associated with reduced depression, while two found no benefit. Life satisfaction, tracked in five cohorts over periods ranging from three to twenty-five years, showed improvement in four of the five. Wellbeing associations appeared in three cohorts with follow-ups of ten to twenty-nine years. One notable finding is that quality-of-life benefits were observed only when volunteering was reciprocal—when volunteers felt appreciated. Remove that sense of meaning, and the effect disappears. Physical health, however, tells a different story. Functional ability, self-rated health, frailty, and the number of chronic conditions—none of these showed consistent improvement across the cohort studies. Whatever volunteering does, it appears to influence mood and mortality before it impacts the body in any measurable way. Or, possibly, it doesn’t reach the body at all, and the mortality association operates entirely through psychological and social pathways. Now, let's consider the experimental evidence. This is where the findings become humbling. The randomized controlled trials and non-randomized trials, largely studying intergenerational volunteering programs where older adults were placed in schools or community settings, did not confirm what the cohort studies found. For depression, self-rated health, and self-esteem, the experimental results were diverse and generally null. The trials were small, likely underpowered, and used populations and program designs that do not cleanly map onto the wide variety of volunteering contexts captured in the cohort studies. Yet, the gap is real. One cannot view a twenty-two percent mortality reduction in the observational data alongside a null result in the experiments and conclude that the causal story is settled. The proposed mechanisms are plausible. Volunteering might increase physical activity simply by encouraging people to get out more—and if that's the operative pathway, then frequency would matter, meaning that more trips out would lead to greater effects. Social integration is another possibility: studies have found that social connectedness partly mediates the volunteering-survival relationship, and individuals with stronger social networks have independently lower mortality rates. Psychological mechanisms could also include a sense of purpose, altruism, and reciprocity. The review notes that there may be an optimal engagement level for mental wellbeing—perhaps around ten hours a month—beyond which volunteering could become burdensome, leading to burnout and eroding benefits. Improved mental wellbeing could then potentially feed back into physical health and survival, although that chain remains speculative. What complicates understanding these mechanisms is that the literature rarely describes volunteering precisely. Jenkinson and colleagues found conflicting results on the type and intensity of volunteering—formal versus informal, weekly versus monthly, and one organization versus several—and could not pinpoint which features matter most. This is not a minor gap. If one wants to design a volunteering program as a health intervention, it is crucial to know what to recommend. Selection bias remains the dominant interpretive concern throughout. Volunteers in the cohort studies tend to be more affluent and in better baseline health, and even after adjusting for sociodemographic, lifestyle, and health differences, a statistically significant survival advantage persists. That adjustment reduces but does not eliminate the effect. One cannot fully rule out that some unmeasured quality—let's call it vitality, social appetite, or baseline optimism—drives both the decision to volunteer and the longer life. Jenkinson and colleagues stop short of endorsing volunteering as a clinical intervention. What they explicitly call for is pragmatic randomized controlled trials that are large enough to be adequately powered and diverse enough in population and program design to be generalizable. Until those trials are conducted, the evidence base cannot support volunteering as a public health prescription in the way governments have been framing it. The review identifies specific gaps, including optimal dose and frequency, the type of activity that produces health-relevant effects, the populations most likely to benefit, and the mediating pathways researchers should be measuring—such as physical activity, social integration, stress, and functional ability. The tension at the heart of this review—a compelling observational mortality signal that does not withstand experimental scrutiny—is a lesson that extends well beyond volunteering. Observational data can indicate what is associated with better health in the real world, while experiments determine whether those associations are causal. When those two findings disagree, the honest answer isn't to choose the number one prefers. It's to conduct better experiments. That work, regarding volunteering and health, is still unfinished. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.