← Longevity & Supplement Guides
Does a Sense of Purpose Really Make You Live Longer?
Purpose used to be something you discussed with a priest, a therapist, or nobody. This summer it acquired a hazard ratio.
In July, a team led by Angelina Sutin published the largest analysis yet of purpose in life and death: 25 samples, 488,765 people, 48,928 deaths, some followed for as long as 32 years. Those who scored higher on purpose were roughly 30% less likely to die during follow-up — a pooled hazard ratio of 0.76, with a confidence interval narrow enough (0.70 to 0.83) that the finding is not going to quietly evaporate.
Weeks earlier, a different paper working in one of the same datasets argued that most of that effect is an artefact of sloppy measurement. Both papers are careful. Both are probably right about something. The gap between them is the most useful thing published on longevity this month, and it is not the gap the headlines picked.
What "purpose" actually is on a questionnaire
The thing being measured is narrower and stranger than the word suggests. Most of these studies use a seven-item version of the Ryff and Keyes psychological wellbeing scale, where people rate agreement on a six-point scale with statements about whether their life has direction, whether they make plans and follow through on them, whether the things they do seem to add up.
Notice what is absent. Nobody is asked about a calling, a legacy, a mission, or whether the work is meaningful. The scale measures something closer to directedness — the sense that your days point somewhere and that you are the one aiming them. Purpose, in the epidemiological literature, is much less romantic than purpose in the retreat brochure, which sells it in three-day intensives with a view of the sea.
The narrowness is a feature: it makes the thing answerable in seven questions and comparable across half a million people. It also means your score can fall for reasons that have nothing to do with your character.
The number that made everyone pay attention
The modern interest traces to a 2019 paper in JAMA Network Open. Alimujiang and colleagues followed 6,985 adults over 50 in the Health and Retirement Study and recorded 776 deaths in four years. Those scoring in the lowest purpose band had a hazard ratio of 2.43 against the highest — a 2.4-fold difference in the risk of dying — and that was after adjusting for smoking, physical activity, chronic illness, depression and anxiety.
An effect that survives that adjustment list is either important or a warning that the adjustment list is missing something. For six years the field mostly assumed the former.
The strongest objection is that illness takes purpose first
In May, Sias and Turtle published the objection properly, in PLOS ONE. Same cohort family, 5,953 adults over 50, followed 2006 to 2018. They reproduced the familiar result — low purpose, higher mortality, hazard ratio 2.59 in the crude model — and then did something the earlier papers had not: they measured baseline health properly.
Not "do you have a diagnosis," which is what self-report captures, but lung function and grip strength. Objective, unglamorous, sensitive to the illness a person has not been told about yet. With those in the model, the hazard ratio fell to 1.77 and stopped being statistically significant. Excluding people who died early in follow-up pushed it lower still.
Their conclusion is blunt: the strong associations in the literature may largely reflect inadequate measurement of baseline health rather than any effect of purpose on how long you live. Which is a polite way of saying that a body quietly failing tends to stop making plans, and the questionnaire notices before the medical record does.
The mechanism is unglamorous, which is a point in its favour. Grip strength predicts mortality well enough to embarrass more sophisticated biomarkers. If it can absorb most of the purpose effect, purpose was partly a slow readout of the same thing.
One construct, three honest numbers
The purpose–mortality association, before and after the health you cannot see
Hazard ratio on a logarithmic axis. The dashed line is no difference (1.0). Rows two and three report the opposite direction, so they sit to its right.
How to read a hazard ratio without over-reading it
A hazard ratio of 0.76 means that at any given moment during follow-up, the higher-purpose group was dying at about three-quarters the rate of the lower-purpose group. It is a ratio of speeds, not a promise of years, and it says nothing about any individual.
Statistical significance is not a measure of importance. The 1.77 above is a larger number than the 0.76 and a weaker finding, because the uncertainty around it grew wide enough to include "no difference at all."
Why the objection does not erase the finding
The PLOS ONE analysis is one cohort of six thousand people. The meta-analysis pools 25 samples and 48,928 deaths, and reports that the association held across sociodemographic groups, with only small differences by age, race and education. It also states that behavioural, clinical and psychological risk factors explain the association in part but not completely. That residue is not nothing.
The follow-up length matters too. Sick-people-stop-planning is an excellent explanation for a four-year study. It is a strained explanation for a signal still visible after three decades, because the illness that was silently present in 1994 has long since declared itself and been adjusted for or killed its host.
So the honest position is neither headline. Purpose is not a 30% discount on death. Nor is it a mirage. It is a real association of uncertain and probably modest size, wrapped around a much larger core of "your body already knows something and the questionnaire is picking it up."
The loneliness result reframes the whole question
The paper that actually changed how I think about this appeared in Social Science & Medicine in January and barely made the rounds. Working with 8,351 adults from the Health and Retirement Study, mean age 68, followed 11 years with 1,191 deaths, the authors asked how loneliness gets from a feeling to a death certificate. Their answer: through purpose. Purpose in life statistically explained an estimated 88% of the association between loneliness and mortality risk — and crucially, most of that ran through declines in purpose over time rather than differences in baseline levels. The pattern held after adjusting for depression, social isolation and neuroticism.
Read that as a mechanism rather than a statistic and it lands differently. Loneliness does not appear to act directly on the body so much as dismantle the reasons to keep aiming at things. First the calendar empties, then the plans stop being made, then the physiology follows. That is a considerably more actionable story than "have more purpose," and it sits neatly alongside what we know about social connection and longevity and about how beliefs about your own ageing track with how it goes.
Purpose is a calendar more than a feeling
If purpose can be eroded by loneliness, then it is not a fixed trait you either possess or lack. It is downstream of arrangements. And arrangements are the part of this you can inspect on a Tuesday.
What the seven questions quietly detect, I think, is whether anything expects you. Not whether you feel inspired — whether something on Thursday would notice your absence, and whether next month requires your participation. A choir, a five-a-side game, a grandchild collected from school, an allotment, a class you teach badly.
The useful thing to notice this week is not "do I have a purpose," which is unanswerable at eight in the morning. It is: what is on my calendar that involves another person and would fail if I did not show up? If the answer is nothing, that is a finding — the same finding these papers are picking up in aggregate, arriving in time to be acted on.
The honest caveat, stated plainly: no trial has taken people with empty calendars, assigned half of them a standing commitment, and shown they lived longer. That study does not exist and may never. What exists is a large, consistent association, a credible mechanism running through loneliness, and a serious argument that much of the raw effect is your health showing through. Acting on the first two costs you almost nothing and is pleasant on its own terms, which is a rare combination in this field.
A number you cannot wear
Agen's premise is that the invisible becomes manageable once measured, and I want to be careful about where that premise stops. Purpose will not arrive as a metric on your wrist, and it should not. There is no sensor for whether your life points somewhere; anything claiming one is a mood ring with a subscription.
What a measurement system can do is catch the things purpose is entangled with, early enough to matter. The PLOS ONE paper is a gift here rather than a debunking: it tells you that low purpose and declining physical capacity travel together, and that the physical side is measurable now. Grip strength, aerobic fitness, how fast you walk, how quickly your heart rate settles after effort — these are the objective baseline health the purpose literature kept failing to capture, and they are exactly what the biomarkers worth tracking are for. If your grip and your walking pace are drifting down while your plans thin out, you are not having a philosophical crisis. You are having a physiological one with philosophical symptoms.
The opposite case is commoner and more cheerful: a measurable body, every trend pointing the right way, and an empty Thursday. No wearable will name that for you. It belongs with the healthspan gap — the difference between a long life and a life you would want extended — and it is the part of a longevity system the instruments cannot reach.
Use the numbers to correct the fantasy, not to replace the experience. The Agen Band can tell you your recovery is fine. Whether anything is expecting you on Thursday is a question you have to answer yourself, and it appears to be one of the load-bearing ones.
The bottom line
The largest analysis to date — 488,765 people, 48,928 deaths, up to 32 years — puts higher purpose alongside roughly 30% lower mortality risk, a pooled hazard ratio of 0.76. A serious 2026 objection shows that in one well-measured cohort most of that dissolves once lung function and grip strength enter the model: an unknown share of the effect is illness casting a shadow backwards.
Between those two sits the result I would keep: loneliness appears to reach mortality mainly by eroding purpose, and mainly through purpose falling over time rather than starting low. Purpose behaves less like a trait and more like a structure that can be dismantled — which is also why it can be rebuilt.
So take the 30% with proper scepticism and the mechanism seriously. Then look at the week ahead and find the appointment that would fail without you. If there isn't one, that is the most informative health number you will read today, and it did not require a device.
