A nationwide study of people with chronic illness found that those who prayed daily were 48% more likely to survive over the following 6 years, after controlling for depression, social support, and health behaviors

A 2023 study published in the Journal of Religion and Health followed a nationwide sample of adults living with chronic illness and found that those who prayed daily were 48 per cent more likely to be alive six years later, after researchers controlled for depression, social support, and a range of health behaviours. The result was reported after multivariate adjustment. The researchers accounted for whether participants exercised, whether they smoked, whether they had people around them, and whether they were experiencing the kind of sustained low mood that itself tracks with poorer outcomes. After all of that, the prayer association remained.

This is one study, not settled consensus, and the finding should not be read as the final word on prayer, religiosity, or survival. What it is worth doing is sitting with the result carefully, because the reflex to explain it away quickly is just as much a bias as the reflex to declare it confirmed.

What the study did and did not control for

The value of a finding like this depends almost entirely on what the researchers actually adjusted for, and the controls here were more thorough than much of the earlier literature on religion and health. Depression was included because it is both independently associated with mortality in chronic illness and correlated with lower rates of religious practice. Social support was included because community and belonging are independently associated with health outcomes, and religious practice often bundles social connection in with it. Health behaviours were included because people with stronger religious commitments sometimes report lower rates of smoking and alcohol use.

After those adjustments, something was still there. That is the finding. The 48 per cent figure is not raw; it is what remained when the researchers tried to subtract the obvious candidates.

What they could not fully account for is harder to specify. Self-reported prayer, like most self-reported data, carries its own noise. People who describe themselves as daily prayer practitioners may differ from others in ways the study’s instruments did not capture: in their attitude toward the illness itself, in how they think about time and the future, in what they do with distress that does not resolve. The study cannot separate prayer as a practice from prayer as a marker of something broader. That limitation is not a reason to dismiss the result. It is a reason to read it precisely.

A long and contested body of literature

The relationship between religious practice and health has been studied for several decades, with results that range from strong positive associations to negligible ones depending on the population, the health outcome, the measure of religiosity, and the analytical approach. Harold Koenig at Duke University has produced some of the most thorough reviews of this literature, and his work consistently finds that religious attendance, more reliably than private prayer, is associated with longer survival and lower rates of a range of illnesses. A meta-analysis by Michael McCullough and colleagues, published in Health Psychology in 2000, reviewed more than 40 studies and found a survival advantage associated with religious involvement, with an odds ratio of 1.29 — somewhat lower than the finding reported here, which used private prayer rather than general religious involvement as the measure.

Private daily prayer is a different measure from attendance, which complicates comparison. Attendance bundles social contact, physical activity in getting to a service, and the structure of a weekly ritual into the same behaviour. Prayer, studied separately, removes most of that. If the association persists after controlling for social support, something other than the social dimension of religious life seems to be doing some of the work.

What that something is remains genuinely unclear. Proposed mechanisms range from psychological: a sense of meaning and coherence under stress, a practised orientation toward acceptance, lower physiological arousal over time; to social: the structure of daily practice itself, the role of belief in regulating behaviour across years rather than weeks; to what the literature sometimes calls the transcendent, which no study design has yet found a way to measure directly. Naming these is not endorsing them. They are the candidates the research community has discussed. None has been confirmed. The field does not have a causal account.

The limits of adjustment

In studies like this one, what the researchers can control for is always a version of what they thought to measure. Depression was accounted for, but how depression was measured matters. A short questionnaire captures certain things. It does not capture the texture of how someone lives inside their illness across years: whether they feel that time is open or foreclosed, whether their days have a shape they find legible, whether they have a framework for why they are still trying. These are not clinical categories. They are ordinary features of the experience of being unwell for a long time.

It is possible that prayer is functioning in this dataset as a proxy for a cluster of orientations the study could not directly measure. People who pray daily in the context of serious illness are, by definition, persisting in a practice that requires returning to something they believe matters. That quality of sustained return, whatever its object, may be part of what is being measured.

That reading does not diminish the finding. It complicates the interpretation of what prayer is doing, which is not the same as explaining the result away.

The direction of influence

One methodological concern worth naming is the direction of any relationship between prayer and survival. Studies of this kind are observational, and while a six-year follow-up design is more robust than a cross-sectional snapshot, it still cannot confirm causation. People who prayed daily at the study’s outset may have been, in ways not fully captured by the measures available, already on a different trajectory. They may have had a relationship with their illness that was already more integrated, a set of expectations about the future that were already more stable. The prayer may be downstream of something rather than upstream of survival.

Conversely, there is a straightforward argument that the practice itself produces effects that accumulate. A daily practice of any kind, sustained over years, affects the structure of time. It affects what gets attended to. In the context of illness, where one of the quiet losses is often the sense that each day has a shape, that may be more material than it sounds.

What to do with a result like this

The temptation with research in this area runs in two directions at once. The sceptical reader dismisses the finding as residual confounding that better methods would eliminate. The credulous reader extracts a prescription. Neither response is quite right.

What the study supports is a careful attention to the category of practice in the lives of people with chronic illness, not the specific content of prayer but the dimension of daily life it represents: regularity, orientation toward something beyond the immediate, a felt sense of meaning under conditions that otherwise resist it. The finding belongs to this dataset and this population. The hierarchy of what helped is scoped to this model. None of it translates cleanly to advice.

If there is an implication worth drawing out, it is not that people with chronic illness should begin praying. It is that the variables researchers routinely include in health outcome models may not yet be capturing everything that matters, and that what gets left out tends to be whatever is hardest to operationalise. Meaning is hard to operationalise. So is the felt texture of daily life over years. The survival advantage associated with daily prayer, in a study that controlled for what it could, is a signal that something in that territory is doing real work. What exactly, the study cannot say.

The finding will likely be contested, replicated imperfectly, and partially explained by better-designed future studies. That is how the field moves. For now, it is a result that resists the quick summary, which is probably the right place to leave it.

If you or someone you know is managing a serious chronic illness and finding it difficult, speaking with a doctor, specialist, or counsellor is more useful than an article. This piece is a reading of a research finding, not a treatment path.