Randomized trial evidence in older adults with hypertension has suggested that meditation may reduce mortality, but the association has not been tested in a nationally representative US sample with linked mortality data. This study pooled multiple National Health Interview Survey cohorts to estimate associations between meditation practice and long-term mortality, to examine variation by family income, and to synthesize survey estimates with trial evidence for a subgroup comparable to trial participants.
The analysis combined four NHIS cohorts spanning 2002 to 2017, linked to mortality follow-up through 2019. The pooled analytic sample included 90,041 adults, of whom 8,402 (9.3%) reported practicing meditation in the past 12 months. The pooled follow-up encompassed 818,248 person-years.
The primary analytic strategy used doubly robust survey-weighted Cox proportional hazards models. These models combined stabilized inverse-probability weights with covariate adjustment to reduce bias from measured confounding. The authors standardized 15-year risks using g-computation. Additionally, probabilistic bias analysis was used to address exposure misclassification, including scenarios stratified by income. For a prespecified high-risk subgroup, a Bayesian synthesis combined the subgroup survey estimate with a pooled randomized trial prior; the trial prior was discounted in sensitivity analyses (including a 50%-discounted prior reported in the abstract).
Primary outcomes were all-cause mortality and cardiovascular mortality, ascertained through linked mortality files. Across the pooled cohorts and follow-up period, there were 10,027 all-cause deaths and 3,040 cardiovascular deaths.
After doubly robust adjustment, the hazard ratio (HR) for all-cause mortality comparing those reporting meditation practice to non-practitioners was 1.02 (95% CI 0.90–1.16). For cardiovascular mortality the adjusted HR was 0.96 (95% CI 0.79–1.16). These point estimates and confidence intervals indicate no clear overall association between recent meditation practice and long-term all-cause or cardiovascular mortality in the full pooled sample.
The authors examined effect modification by family income, using 200% of the federal poverty level (FPL) as the threshold. Among individuals below 200% FPL, HR estimates suggested a possible protective association: all-cause HR 0.85 (95% CI 0.71–1.03) and cardiovascular HR 0.80 (95% CI 0.57–1.10). Among those at or above 200% FPL, HRs were 1.10 (all-cause; 95% CI 0.93–1.30) and 1.02 (cardiovascular; 95% CI 0.80–1.30). Tests for interaction returned p=0.19 for all-cause mortality and p=0.62 for cardiovascular mortality. The confidence intervals for stratified estimates overlap the null and are compatible with both modest harm and modest benefit; interaction tests were not statistically significant at conventional thresholds.
The investigators conducted probabilistic bias analyses to address potential misclassification of meditation exposure, including scenarios stratified by income. The abstract reports that such analyses were performed but does not provide detailed bias-adjusted numeric results in the summary. Therefore, specific corrected estimates from the bias analysis are not reported here because they were not detailed in the source abstract.
To integrate trial evidence with survey data for a subgroup resembling trial participants, the authors examined non-Hispanic Black adults aged ≥55 years with hypertension (n=2,786). They performed Bayesian synthesis combining the subgroup estimate with a pooled randomized trial prior. Under a 50%-discounted trial prior, posterior estimates were reported as 0.84 for all-cause mortality with Pr(HR<1)=0.89, and 0.70 for cardiovascular mortality with Pr(HR<1)=0.96. These posterior results suggest a higher probability of benefit—particularly for cardiovascular mortality—in this high-risk subgroup when trial evidence is incorporated, though the analysis is described as hypothesis-generating.
In this nationally representative pooled analysis, meditation practice was not associated with reduced all-cause or cardiovascular mortality after doubly robust adjustment in the overall sample. However, subgroup signals—most notably among adults below 200% of the federal poverty level and in a trial-comparable high-risk subgroup—suggest potential protective effects, particularly for cardiovascular mortality. The authors frame these findings as hypothesis-generating and recommend further research focused on effect modification and improved, more detailed exposure measurement of meditation practice.
All data used in the study are publicly available from the National Center for Health Statistics linked mortality files. The authors declared no competing interests and reported that ethical guidelines and appropriate approvals were followed; the analysis used openly available deidentified data.