West China School of Public Health and West China Fourth Hospital, Sichuan University, Chengdu, China.
Xiamen Center for Disease Control and Prevention, Xiamen, China.
BMC Med Res Methodol. 2024 Feb 27;24(1):49. doi: 10.1186/s12874-024-02167-9.
Several approaches are commonly used to estimate the effect of diet on changes of various intermediate disease markers in prospective studies, including "change-score analysis", "concurrent change-change analysis" and "lagged change-change analysis". Although empirical evidence suggests that concurrent change-change analysis is most robust, consistent, and biologically plausible, in-depth dissection and comparison of these approaches from a causal inference perspective is lacking. We intend to explicitly elucidate and compare the underlying causal model, causal estimand and interpretation of these approaches, intuitively illustrate it with directed acyclic graph (DAG), and further clarify strengths and limitations of the recommended concurrent change-change analysis through simulations.
Causal model and DAG are deployed to clarify the causal estimand and interpretation of each approach theoretically. Monte Carlo simulation is used to explore the performance of distinct approaches under different extents of time-invariant heterogeneity and the performance of concurrent change-change analysis when its causal identification assumptions are violated.
Concurrent change-change analysis targets the contemporaneous effect of exposure on outcome (measured at the same survey wave), which is more relevant and plausible in studying the associations of diet and intermediate biomarkers in prospective studies, while change-score analysis and lagged change-change analysis target the effect of exposure on outcome after one-period timespan (typically several years). Concurrent change-change analysis always yields unbiased estimates even with severe unobserved time-invariant confounding, while the other two approaches are always biased even without time-invariant heterogeneity. However, concurrent change-change analysis produces almost linearly increasing estimation bias with violation of its causal identification assumptions becoming more serious.
Concurrent change-change analysis might be the most superior method in studying the diet and intermediate biomarkers in prospective studies, which targets the most plausible estimand and circumvents the bias from unobserved individual heterogeneity. Importantly, careful examination of the vital identification assumptions behind it should be underscored before applying this promising method.
在前瞻性研究中，有几种常用的方法来估计饮食对各种中间疾病标志物变化的影响，包括“变化评分分析”、“同期变化变化分析”和“滞后变化变化分析”。尽管经验证据表明同期变化变化分析最稳健、一致且具有生物学合理性，但从因果推理的角度深入剖析和比较这些方法还很缺乏。我们旨在从因果推理的角度明确阐明和比较这些方法的潜在因果模型、因果估计量和解释，并通过有向无环图（DAG）直观地说明，通过模拟进一步澄清推荐的同期变化变化分析的优缺点。
因果模型和 DAG 用于从理论上阐明每种方法的因果估计量和解释。蒙特卡罗模拟用于探索在不同程度的时不变异质性下不同方法的性能，以及在违反同期变化变化分析的因果识别假设时，同期变化变化分析的性能。
同期变化变化分析针对的是暴露对结果的同期影响（在同一调查波次测量），这在研究前瞻性研究中饮食与中间生物标志物的相关性时更相关且更合理，而变化评分分析和滞后变化变化分析针对的是暴露对结果的一个时段（通常为几年）后的影响。即使存在严重的未观察到时不变混杂，同期变化变化分析也始终产生无偏估计，而其他两种方法即使不存在时不变异质性也始终存在偏差。然而，同期变化变化分析随着其因果识别假设的违反变得更加严重，会产生几乎线性增加的估计偏差。
同期变化变化分析可能是研究前瞻性研究中饮食与中间生物标志物的最佳方法，它针对的是最合理的估计量，并避免了未观察到的个体异质性引起的偏差。重要的是，在应用这种很有前途的方法之前，应该强调仔细检查其背后的关键识别假设。
