Beyond sequential stages: asynchronous structural reorganisation in China’s mortality transition, 1981 ∼ 2020

China’s post-reform mortality transition has been widely interpreted, yet its sequencing, structural reorganization, and internal heterogeneity remain insufficiently examined. Using epidemiological transition theory as a heuristic rather than a deterministic model, this study evaluates whether China’s mortality transition conforms to or diverges from sequential transition assumptions. We re-estimated age-sex-year-specific mortality rates for 1981 ∼ 2020 using spline-Gaussian process regression, combined with cause-of-death estimates from GBD 2023. Decomposition techniques, joinpoint regression, and Gaussian mixture model were applied to examine age-cause-specific contributions, change pace, and pattern reorganization. The results show that early mortality improvements were driven by reductions in infant and infectious mortality; subsequent improvements shifted toward middle and older ages, dominated by cardiovascular and other chronic diseases. Infectious and degenerative causes coexisted for extended periods, and mortality patterns were reorganized through discrete structural regimes. Female adult mortality declined earlier and more consistently, widening and then stabilizing the gender gap. Our findings demonstrate that China’s mortality transition is consistent with the overall development described by epidemiological transition theory but deviates from the theory’s assumption that stages develop sequentially and uniformly with a nation. Instead, China’s scenario reflects an asynchronous transition influenced by institutional reforms and regional differences.