We investigate long-term cognitive effects of an intervention, where systolic blood pressure (sBP) is monitored at more optimal levels, in a large representative sample. A limitation with previous research on the potential risk reduction of such interventions is that they do not properly account for the reduction of mortality rates. Hence, one can only speculate whether the effect is a result from changes in cognition or changes in mortality. As such, we extend previous research by providing both an etiological and a prognostic effect estimate. To do this we propose a Bayesian semi-parametric estimation approach for an incremental intervention, using the extended G-formula. We also introduce a novel sparsity-inducing Dirichlet hyperprior for longitudinal data, demonstrate the usefulness of our approach in simulations, and compare the performance relative to other Bayesian decision tree ensemble approaches. In our study, there were no significant prognostic- or etiological effects across all ages, indicating that sBP interventions likely do not have a strong effect on memory neither at the population level nor at the individual level.

翻译：我们基于大规模代表性样本，探究将收缩压(sBP)维持在更优水平这一干预措施的长期认知效应。既往关于此类干预措施降低潜在风险的研究存在局限性——未能充分考虑干预对死亡率变化的修正作用。因此，我们仅能推测认知效应源于认知功能改变还是死亡率变化。为此，我们通过提供病因学效应与预后效应两种估计值来拓展既往研究。我们提出一种基于贝叶斯半参数估计方法的增量干预方案，并采用扩展G公式实现。同时引入一种适用于纵向数据的新型稀疏诱导狄利克雷超先验，通过仿真验证该方法的有效性，并与其他贝叶斯决策树集成方法进行性能对比。研究结果表明，所有年龄组均未观察到显著的预后效应或病因学效应，提示无论群体层面还是个体层面，收缩压干预可能均不会对记忆功能产生显著影响。