[論文レビュー] Novel g-computation algorithms for time-varying actions with recurrent and semi-competing events

Alena Sorensen D'Alessio, Lucas M. Neuroth+3 more|arXiv (Cornell University)|Mar 10, 2026

Advanced Causal Inference Techniques被引用数 0

ひとこと要約

The paper introduces two novel g-computation algorithms to estimate causal effects with time-varying actions in the presence of semi-competing and recurrent events, validated by simulations and an application to smoking prevention and hypertension.

ABSTRACT

Background: A core aspect of epidemiology is determining the impacts of potential public health interventions over time. With long follow-up periods, epidemiologists may need to consider semi-competing events, in which a terminal event, like death, precludes a non-terminal event, like hypertension. Time-varying confounding poses an additional challenge when studying time-varying interventions or actions. Existing methods do not simultaneously address semi- competing events and time-varying confounding. Methods: We propose two novel g-computation algorithms for causal effects with semi- competing events and time-varying actions. To explore performance of our novel g-computation estimators, we conducted a Monte Carlo simulation study. We then applied our estimator to investigate how cigarette smoking prevention throughout young and middle adulthood might impact prevalent hypertension using data from Waves III (aged 18-26 years) - VI (aged 39-51 years) of the National Longitudinal Study of Adolescent to Adult Health. Results: Our simulations show that the novel g-computation estimators had little bias and appropriate confidence interval coverage. They outperformed existing alternative estimators across sample sizes. In the illustrative application, the novel estimator identified a small reduction in prevalence of hypertension and risk of death in midlife had all cigarette smoking been prevented across follow-up compared to the observed smoking patterns. Conclusion: As long-running cohorts progress in age, death within the study sample will become an increasing concern for studies of aging-related outcomes, life course analyses, and investigations into chronic disease development. Our novel g-computation estimators provide a simultaneous solution.

研究の動機と目的

• Address how semi-competing events (e.g., death) influence causal effect estimation in long-term action interventions.
• Develop two novel g-computation algorithms that accommodate time-varying confounding and semi-competing events.
• Evaluate estimator performance via Monte Carlo simulations across sample sizes.
• Apply the methods to a longitudinal cohort to assess how cigarette smoking prevention affects hypertension and mortality.

提案手法

• Develop two novel g-computation algorithms for causal effect estimation under semi-competing events and time-varying actions.
• Conduct Monte Carlo simulations to assess bias and confidence interval coverage.
• Compare the novel estimators against existing alternative estimators under various sample sizes.
• Apply the methods to Waves III–VI of the National Longitudinal Study of Adolescent to Adult Health to study smoking prevention effects on hypertension and death.

実験結果

リサーチクエスチョン

• RQ1 How do semi-competing events and time-varying actions affect causal effect estimation in longitudinal studies?
• RQ2 Do the proposed g-computation estimators exhibit low bias and proper confidence interval coverage across sample sizes?
• RQ3 How do the novel estimators perform relative to existing methods in handling time-varying confounding and semi-competing events?
• RQ4 What is the estimated impact of comprehensive cigarette smoking prevention on hypertension prevalence and mortality in midlife in the studied cohort?

主な発見

• Simulations showed the novel estimators had little bias and appropriate confidence interval coverage.
• The novel estimators outperformed existing alternative estimators across different sample sizes.
• In the illustrative application, full cigarette smoking prevention was associated with a small reduction in hypertension prevalence and a reduced risk of death in midlife compared to observed smoking patterns.