Department of Biostatistics & Data Science, University of Texas Health Science Center, Houston, TX, USA.
Department of Epidemiology, Human Genetics & Environmental Sciences, University of Texas Health Science Center, Houston, TX, USA.
Am J Drug Alcohol Abuse. 2022 Sep 3;48(5):529-537. doi: 10.1080/00952990.2022.2027955. Epub 2022 Jan 31.
Existing studies of dual use of electronic and combustible cigarettes either collected longitudinal data with long gaps in between waves or conducted ecological momentary assessment (EMA) over a short period of time. In recent years, the measurement burst design that embeds an EMA protocol in each wave assessment of a traditional longitudinal study has become more popular and yet conventional generalized linear mixed models (GLMM) have important limitations for handling data from this design. This study proposed a new statistical method to analyze data from the measurement burst design. This new statistical method was designed to model the short-term (within-wave) as well as long-term (between-wave) changes and was validated by a simulation study. Secondary analysis was conducted to analyze data from 205 dual users (52% male) and 146 exclusive smokers (50% male) who participated in a recent study using the measurement burst design. The simulation study shows that the proposed method can handle the gap between waves well and is also robust to nonlinear changes across waves. Although no short-term change in smoking was found, dual users reported a long-term reduction in cigarette use that was more rapid compared to exclusive smokers (). Vaping more was associated with smoking less (). The proposed method is highly applicable as it can be easily implemented by substance use researchers and the results can be straightforwardly interpreted. The results suggest that e-cigarette use may play a role in promoting a long-term reduction in smoking among dual users.
现有的关于电子烟和可燃香烟双重使用的研究要么在波与波之间存在很长的时间间隔，要么在短时间内进行生态瞬时评估（EMA）。近年来，在传统纵向研究的每一波评估中嵌入 EMA 方案的测量爆发设计变得越来越流行，但传统的广义线性混合模型（GLMM）对于处理来自这种设计的数据存在重要的局限性。本研究提出了一种新的统计方法来分析测量爆发设计的数据。这种新的统计方法旨在对短期（波内）和长期（波间）变化进行建模，并通过模拟研究进行了验证。对最近使用测量爆发设计的 205 名双重使用者（52%为男性）和 146 名单一吸烟者（50%为男性）进行了二次分析。模拟研究表明，所提出的方法可以很好地处理波之间的差距，并且对波之间的非线性变化也具有稳健性。虽然没有发现吸烟的短期变化，但双重使用者报告称，与单一吸烟者相比，他们的吸烟量长期减少更快（）。吸更多电子烟与吸更少香烟有关（）。由于该方法具有高度的适用性，因此可以被药物使用研究人员轻松实施，并且结果可以直接解释。研究结果表明，电子烟的使用可能在促进双重使用者长期减少吸烟方面发挥作用。
