Background: Evidence on mortality benefits of general health checks comes mainly from higher-resource settings. We examined variation by resource setting, baseline risk, modifiable risk-factor change, and allocation strategy. Methods: We analysed six longitudinal cohorts from 17 countries and a real-world and population-based cohort in northwestern China. Health-check receipt was time varying and the outcome was all-cause mortality. Cohort-specific Cox models were fitted; the northwestern China analysis also included instrumental-variable models, mediation analysis, absolute-benefit estimation, and six-year allocation simulations. Findings: The multinational cohorts included 149,620 participants and 2,803,872 person-years; the the Northwest China Real-world and Population-based Cohort (NCRP) included 14,505,733 participants and 89,829,304 person-years. Health checks were not associated with mortality in the US or European cohorts, but were associated with lower mortality in cohorts from Mexico, Indonesia, and China (hazard ratios 0.726–0.898). In northwestern China, annual health checks were associated with lower subsequent-year mortality (hazard ratio 0.828 [95% CI 0.822–0.835]), with directionally consistent instrumental-variable estimates. Absolute benefit increased with risk burden, and modifiable risk-factor improvement mediated 16.3% (95% CI 15.59–16.96) of the association. With 10 million annual examinations, risk-targeted allocation was estimated to avert 10,284 deaths per year versus 7,717 with population-wide allocation. Interpretation: Associations varied by resource context and baseline risk. Risk-targeted implementation might improve efficiency in lower-resource settings, although the observational and modelled findings should not be interpreted as causal effects.