Modeling Interventions for Lung Cancer Mortality

对癌症发病率和死亡率的时间趋势的研究可以对疾病对人口的影响提供有价值的见解。将建立一个模型，其中吸烟（这种疾病的一个众所周知的原因）对以人口为基础的肺癌发病率的影响。年龄-时期-队列模型提供了一种对时间趋势进行统计总结的有用方法。在这种情况下，年龄代表了衰老过程对疾病风险的影响。另一方面，时期和队列可能反映了接触进口风险因素或监测系统的变化。虽然分析性流行病学研究是估计假定的风险因素对疾病风险的影响的最佳方法，但定量描述暴露变化对人口比率的影响可能更具挑战性。本研究的目的是建立一个模型，用肺癌发病率危险因素的趋势来描述观察到的该病发病率和死亡率的趋势。这些数据将用于估计旨在减少吸烟的干预措施对肺癌死亡率的影响。这种方法将扩展到其他风险因素，如石棉暴露、遗传学和肺癌筛查的使用。本研究的具体目的是：(1)完成SEER登记中肺癌发病率趋势模型的开发，并确定现有吸烟趋势数据可以用作解释变量的程度；(2)利用SEER登记处的现有数据，完成描述肺癌发病率和死亡率之间关系的隔室模型的开发；(3)建立一个模型，利用现有的州吸烟趋势信息来解释相邻州之间癌症死亡率趋势的变化；(4)从目标1-3验证模型；(5)利用目标1-3中建立的模型，估计各种癌症控制策略对未来肺癌死亡率趋势的群体效应；(6)积极与CISNET成员合作；(7)开发开源软件代码和文档，使人们能够应用来自其他来源的数据，并将这种方法扩展到分析来自其他癌症部位的数据；(8)开发一个网站，将模型的假设提供给感兴趣的建模者，提供相当详细的信息。

The study of time trends in cancer incidence and mortality can provide valuable insights into the effect that a disease is having on the population, A model will be developed in which the effect that smoking cigarettes, a well known cause of this disease, has on population based lung cancer rates. Age-period-cohort models have offered one useful way of developing a statistical summary of temporal trends. In this case, age represents the effect of the aging process on a disease risk. Period and cohort, on the other hand, are likely to reflect changes in the exposure to import risk factors or in the surveillance system. While analytical epidemiologic studies offer the best way to estimate the effect of putative risk factors on disease risk, quantitative descriptions of the way in which changes in exposure can affect population rates can be much more challenging. The purpose of this research is to develop a model in which trends in risk factors for lung cancer incidence are used to describe observed trends in incidence and mortality for the disease. These will then be used to estimate the effect on lung cancer mortality of interventions designed to reduce cigarette smoking. This approach will be extended to other risk factors, such as asbestos exposure, genetics and use of lung cancer screening. The specific aims of this research are to: (1) Complete development of a model for lung cancer incidence trends among SEER registries and determine the extent to which available data on smoking trends can be used as explanatory variables; (2) Complete development of a compartment model that describes the relationship between lung cancer incidence and mortality using available data from SEER registries; (3) Develop a model that uses available state information on cigarette smoking trends to explain the variation in cancer mortality trends among contiguous states; (4) Validate the model from aims 1-3; (5) Use the model developed in aims 1-3 to estimate the population effect of various cancer control strategies on future lung cancer mortality trends; (6) Actively collaborate with members of CISNET; (7) Develop open source software code and documentation that will allow one to apply data from other sources, as well as to extend this approach to the analysis of data from other cancer sites; and (8) Develop a web site that will make the assumptions of the model available to interested modelers in considerable detail.