Chains-of-Risk Model

Both describe a sequence of linked exposures, but they differ in how the links cause disease. In an additive chain, each link independently adds to final risk, so the early exposures retain effects even after later ones are accounted for, and intervening at any link yields some benefit. In a trigger chain, the early links matter only because they raise the probability of a final, decisive exposure that is the true cause; once that final exposure is conditioned on, the early effects vanish. The practical stakes are large: additive chains reward intervention anywhere, while trigger chains demand breaking the chain early or blocking the final trigger.

They are closely linked. The accumulation-of-risk model says total dose of adverse exposure drives disease regardless of timing. A chain of risk adds explicit temporal ordering and linkage between exposures. An additive chain is essentially a temporally sequenced accumulation model — the cumulative burden builds through ordered links — whereas a trigger chain departs from pure accumulation because only the final link causes disease. In practice analysts treat chains as the structured, sequence-aware version of accumulation, using the ordering of exposures to learn not just how much adversity mattered but how it traveled through life.

Because the intermediate exposures in a chain are typically affected by earlier exposures and also share confounders with the outcome. Adjusting for such an intermediate to estimate an early exposure's effect both blocks the very pathway of interest and can open spurious paths through the confounders, biasing the result — the classic problem of a mediator that is also a time-varying confounder. Correctly decomposing a chain therefore requires causal-mediation or g-methods (such as marginal structural models or g-computation) designed for time-varying confounding affected by prior exposure, rather than naive regression adjustment for the chain's intermediate links.