Arman Oganisian, Anthony Girard, Jon A Steingrimsson, Patience Moyo
{"title":"对具有时间错位的重复事件进行因果分析的贝叶斯框架。","authors":"Arman Oganisian, Anthony Girard, Jon A Steingrimsson, Patience Moyo","doi":"10.1093/biomtc/ujae145","DOIUrl":null,"url":null,"abstract":"<p><p>Observational studies of recurrent event rates are common in biomedical statistics. Broadly, the goal is to estimate differences in event rates under 2 treatments within a defined target population over a specified follow-up window. Estimation with observational data is challenging because, while membership in the target population is defined in terms of eligibility criteria, treatment is rarely observed exactly at the time of eligibility. Ad hoc solutions to this timing misalignment can induce bias by incorrectly attributing prior event counts and person-time to treatment. Even if eligibility and treatment are aligned, a terminal event process (eg, death) often stops the recurrent event process of interest. In practice, both processes can be censored so that events are not observed over the entire follow-up window. Our approach addresses misalignment by casting it as a time-varying treatment problem: some patients are on treatment at eligibility while others are off treatment but may switch to treatment at a specified time-if they survive long enough. We define and identify an average causal effect estimand under right-censoring. Estimation is done using a g-computation procedure with a joint semiparametric Bayesian model for the death and recurrent event processes. We apply the method to contrast hospitalization rates among patients with different opioid treatments using Medicare insurance claims data.</p>","PeriodicalId":8930,"journal":{"name":"Biometrics","volume":"80 4","pages":""},"PeriodicalIF":1.4000,"publicationDate":"2024-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"A Bayesian framework for causal analysis of recurrent events with timing misalignment.\",\"authors\":\"Arman Oganisian, Anthony Girard, Jon A Steingrimsson, Patience Moyo\",\"doi\":\"10.1093/biomtc/ujae145\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><p>Observational studies of recurrent event rates are common in biomedical statistics. Broadly, the goal is to estimate differences in event rates under 2 treatments within a defined target population over a specified follow-up window. Estimation with observational data is challenging because, while membership in the target population is defined in terms of eligibility criteria, treatment is rarely observed exactly at the time of eligibility. Ad hoc solutions to this timing misalignment can induce bias by incorrectly attributing prior event counts and person-time to treatment. Even if eligibility and treatment are aligned, a terminal event process (eg, death) often stops the recurrent event process of interest. In practice, both processes can be censored so that events are not observed over the entire follow-up window. Our approach addresses misalignment by casting it as a time-varying treatment problem: some patients are on treatment at eligibility while others are off treatment but may switch to treatment at a specified time-if they survive long enough. We define and identify an average causal effect estimand under right-censoring. Estimation is done using a g-computation procedure with a joint semiparametric Bayesian model for the death and recurrent event processes. We apply the method to contrast hospitalization rates among patients with different opioid treatments using Medicare insurance claims data.</p>\",\"PeriodicalId\":8930,\"journal\":{\"name\":\"Biometrics\",\"volume\":\"80 4\",\"pages\":\"\"},\"PeriodicalIF\":1.4000,\"publicationDate\":\"2024-10-03\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Biometrics\",\"FirstCategoryId\":\"100\",\"ListUrlMain\":\"https://doi.org/10.1093/biomtc/ujae145\",\"RegionNum\":4,\"RegionCategory\":\"数学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q3\",\"JCRName\":\"BIOLOGY\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Biometrics","FirstCategoryId":"100","ListUrlMain":"https://doi.org/10.1093/biomtc/ujae145","RegionNum":4,"RegionCategory":"数学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"BIOLOGY","Score":null,"Total":0}
引用次数: 0
摘要
在生物医学统计中,对复发事件发生率的观察研究很常见。从广义上讲,其目的是估算在指定的随访时间窗口内,目标人群在两种治疗方法下的事件发生率差异。利用观察数据进行估算具有挑战性,因为目标人群的成员资格是根据资格标准确定的,但很少能在获得资格时准确观察到治疗情况。临时解决这种时间错位的方法可能会将先前的事件计数和人时错误地归因于治疗,从而导致偏差。即使治疗资格和治疗时间一致,终末事件过程(如死亡)往往也会停止相关的重复事件过程。在实践中,可以对这两个过程进行删减,这样就不会在整个随访窗口中观察到事件。我们的方法是将错配作为一个时变治疗问题来解决:一些患者在获得治疗资格时接受治疗,而另一些患者则没有接受治疗,但如果他们存活足够长的时间,则可能在特定时间转为接受治疗。我们定义并确定了右删减下的平均因果效应估计值。我们使用 g 计算程序和半参数贝叶斯联合模型对死亡和复发事件过程进行估计。我们将该方法应用于使用医疗保险理赔数据对比不同阿片类药物治疗患者的住院率。
A Bayesian framework for causal analysis of recurrent events with timing misalignment.
Observational studies of recurrent event rates are common in biomedical statistics. Broadly, the goal is to estimate differences in event rates under 2 treatments within a defined target population over a specified follow-up window. Estimation with observational data is challenging because, while membership in the target population is defined in terms of eligibility criteria, treatment is rarely observed exactly at the time of eligibility. Ad hoc solutions to this timing misalignment can induce bias by incorrectly attributing prior event counts and person-time to treatment. Even if eligibility and treatment are aligned, a terminal event process (eg, death) often stops the recurrent event process of interest. In practice, both processes can be censored so that events are not observed over the entire follow-up window. Our approach addresses misalignment by casting it as a time-varying treatment problem: some patients are on treatment at eligibility while others are off treatment but may switch to treatment at a specified time-if they survive long enough. We define and identify an average causal effect estimand under right-censoring. Estimation is done using a g-computation procedure with a joint semiparametric Bayesian model for the death and recurrent event processes. We apply the method to contrast hospitalization rates among patients with different opioid treatments using Medicare insurance claims data.
期刊介绍:
The International Biometric Society is an international society promoting the development and application of statistical and mathematical theory and methods in the biosciences, including agriculture, biomedical science and public health, ecology, environmental sciences, forestry, and allied disciplines. The Society welcomes as members statisticians, mathematicians, biological scientists, and others devoted to interdisciplinary efforts in advancing the collection and interpretation of information in the biosciences. The Society sponsors the biennial International Biometric Conference, held in sites throughout the world; through its National Groups and Regions, it also Society sponsors regional and local meetings.