Measuring Cigarette Smoking Behavior Progression with Cross-Sectional Data

描述（申请人提供）：美国数十年的烟草研究和控制留下了大量国家和地方层面的横断面调查数据。典型的数据来源包括1975年以来的青少年风险行为调查、1984年以来的行为风险因素调查、1971年以来的全国健康和营养检查调查、1957年以来的全国健康访谈调查、1975年以来的监测未来研究、1971年以来的全国药物使用和健康调查以及加州青少年烟草调查。这些包含数百万随机抽样受试者的多源多年数据构成了宝贵的资产，对这些数据的进一步探索将有利于烟草控制，以实现《2010年健康人民》提出的目标。烟草控制计划的有效规划和评估将受益于吸烟行为变化的数据（例如，从不吸烟者到吸烟者，从吸烟者到戒烟者，从戒烟者再次回到吸烟者），因为这些数据比传统使用的终生吸烟者或吸烟者流行率提供更多信息。 30天吸烟者。衡量吸烟行为的变化需要纵向数据，而逐年收集州和地方层面的纵向数据往往成本过高且过于复杂。我们的分析表明，横截面数据包含所需的纵向信息；但剩下的挑战是如何从数据中提取信息。通过长期的合作，该项目的研究人员认识到，工程和系统控制领域的概率离散事件系统（PDES）可以克服这一挑战。遵循香农的采样原则，PDES 技术已成功用于利用横截面数据对复杂的人机交互进行建模。本研究的目标是采用这种方法来模拟吸烟行为。本研究的四个目标是：（1）建立PDES吸烟行为模型，能够利用节省成本的横截面数据评估吸烟行为进展； (2) 分析常见的烟草控制计划（例如税收、限制烟草销售和使用、基于学校​​的计划和戒烟）对吸烟行为进展的影响； (3) 预测吸烟趋势并模拟与《健康人民2010》设定的控烟目标直接相关的不同控烟情景； (4) 开发用于烟草研究和烟草控制实践的 PDES 方法计算机化软件。来自全国药物使用和健康调查的十年数据（1995-2004）将用于测试所提出的方法。然后，该方法将通过纵向数据进一步验证（例如 1997 年全国青年纵向调查）。一个在烟草研究和系统控制方面具有专业知识的合作研究团队已经成立，有能力开展该项目。这项拟议的研究与公共卫生完全相关，因为预期结果将为推进现有烟草研究和加强国家烟草控制战略提供所需的技术和数据。

DESCRIPTION (provided by applicant): Several decades of tobacco research and control in the U.S. have left a large amount of cross-sectional survey data at the national and the local levels. Typical data sources include the Youth Risk Behavioral Survey since 1975, the Behavioral Risk Factor Survey since 1984, the National Health and Nutrition Examination Survey since 1971, the National Health Interview Survey since 1957, the Monitoring the Future Studies since 1975, the National Survey on Drug Use and Health since 1971, and the California Youth Tobacco Survey. These multisource multi-year data with millions of randomly sampled subjects consist of an invaluable asset and further exploration of the data would benefit tobacco control to reach the objectives set forth by the Healthy People 2010. Effective planning and evaluation of tobacco control programs will benefit from data on smoking behavior change (e.g., from nonsmokers to smokers, from smokers to quitters and from quitters backward to smokers again), because such data are more informative than the conventionally used prevalence rates of lifetime smokers or 30-day smokers. Measuring smoking behavior change requires longitudinal data and collecting longitudinal data at the state and local levels year by year are often too costly and too complex. Our analyses indicate that cross-sectional data contain the needed longitudinal information; but a challenge remaining is how to extract the information out of the data. Through a long-term of collaboration, the investigators of this project have recognized that the Probability Discrete Event System (PDES) from the engineering and system control field can be adapted to overcome this challenge. Following Shannon's sampling principles, the PDES technique has been successfully used in modeling sophisticated machine-human interactions with cross-sectional data. The goal of this study is to adapt this method for modeling smoking behavior. The four aims of this research are: (1) to establish a PDES smoking behavior model capable of assessing smoking behavior progression with cost-saving cross-sectional data; (2) to analyze effect of common tobacco control programs (e.g., taxation, restriction of sales and use of tobacco, school-based programs and tobacco cessation) on smoking behavior progression; (3) to predict smoking trends and to simulate different tobacco control scenarios that are directly related to the tobacco control objectives set by the Healthy People 2010; and (4) to develop software that computerizes the PDES method for tobacco research and tobacco control practice. Ten-year data (1995-2004) from the National Survey on Drug Use and Health will be used to test the proposed method. The method will then be further validated with longitudinal data (e.g. National Longitudinal Survey of Youth 1997). A collaborative research team with expertise in tobacco research and system control has been established is well positioned to carry out this project. This proposed research is completely relevant to public health, because the expected results will provide needed techniques and data for advancing the existing tobacco research and strengthening the tobacco control strategies in the nation.