Department of Community and Allied Health, La Trobe Rural Health School, La Trobe University, Bendigo, Australia.
Violet Vines Marshman Centre For Rural Health Research, La Trobe Rural Health School, Melbourne, VIC, Australia.
BMC Public Health. 2023 Oct 18;23(1):2038. doi: 10.1186/s12889-023-16914-y.
To date only a limited number of reviews have focused on how exposure and outcome measures are defined in the existing literature on associations between tobacco retailer density ('density') and smoking behaviour ('smoking'). Therefore this systematic review classified and summarised how both density and smoking variables are operationalised in the existing literature, and provides several methodological recommendations for future density and smoking research.
Two literature searches between March and April 2018 and April 2022 were conducted across 10 databases. Inclusion and exclusion criteria were developed and keyword database searches were undertaken. Studies were imported into Covidence. Cross-sectional studies that met the inclusion criteria were extracted and a quality assessment was undertaken. Studies were categorised according to the density measure used, and smoking was re-categorised using a modified classification tool.
Large heterogeneity was found in the operationalisation of both measures in the 47 studies included for analysis. Density was most commonly measured directly from geocoded locations using circular buffers at various distances (n = 14). After smoking was reclassified using a smoking classification tool, past-month smoking was the most common smoking type reported (n = 26).
It is recommended that density is measured through length-distance and travel time using the street network and weighted (e.g. by the size of an area), or by using Kernel Density Estimates as these methods provide a more accurate measure of geographical to tobacco and e-cigarette retailer density. The consistent application of a smoking measures classification tool, such as the one developed for this systematic review, would enable better comparisons between studies. Future research should measure exposure and outcome measures in a way that makes them comparable with other studies.
This systematic review provides a strong case for improving data collection and analysis methodologies in studies assessing tobacco retailer density and smoking behaviour to ensure that both exposure and outcome measures are clearly defined and captured. As large heterogeneity was found in the operationalisation of both density and smoking behaviour measures in the studies included for analysis, there is a need for future studies to capture, measure and classify exposure measures accurately, and to define outcome measures in a manner that makes them comparable with other studies.
迄今为止，只有少数评论集中讨论了现有关于烟草零售商密度（“密度”）与吸烟行为（“吸烟”）之间关联的文献中，暴露和结果测量是如何定义的。因此，本系统综述对现有文献中密度和吸烟变量的操作化方法进行了分类和总结，并为未来的密度和吸烟研究提供了一些方法学建议。
2018 年 3 月至 4 月和 2022 年 4 月期间，在 10 个数据库中进行了两次文献检索。制定了纳入和排除标准，并进行了关键字数据库搜索。研究结果被导入 Covidence。纳入的符合标准的横断面研究被提取出来，并进行了质量评估。研究根据使用的密度测量方法进行分类，并使用改良的分类工具重新分类吸烟。
在纳入分析的 47 项研究中，两种测量方法的操作化都存在很大的异质性。密度最常直接从地理编码位置使用各种距离的圆形缓冲区进行测量（n=14）。使用吸烟分类工具对吸烟进行重新分类后，过去一个月的吸烟是报告最常见的吸烟类型（n=26）。
建议使用街道网络和加权（例如，通过区域大小）的长度-距离和旅行时间测量密度，或者使用核密度估计，因为这些方法可以更准确地测量地理到烟草和电子烟零售商的密度。一致应用吸烟测量分类工具，例如本系统综述中开发的工具，可以使研究之间的比较更加容易。未来的研究应在评估烟草零售商密度和吸烟行为的研究中以可比较的方式测量暴露和结果测量。
本系统综述强烈建议改进评估烟草零售商密度和吸烟行为的研究中的数据收集和分析方法，以确保明确和捕捉暴露和结果测量。由于纳入分析的研究中密度和吸烟行为测量的操作化存在很大的异质性，因此未来的研究需要准确地捕捉、测量和分类暴露测量，并以与其他研究可比的方式定义结果测量。
