Interpretation of longitudinal studies. An overview.
Interpretation of longitudinal studies. An overview.
复制标题
纵向研究的解释。
DOI：
--
复制
发表时间：
1996
影响因子：
24.7
通讯作者：
Ira B. Tager
中科院分区：
文献类型：
--
作者：
Jan P. Schouten;Ira B. Tager
在他们的介绍性概述中，Weiss和Ware(1)列出了纵向研究的六个优势。其中三项似乎在解释方面特别重要，在某种程度上，这三项包含了其他优势：(1)了解事件的时间顺序；(2)观察个别变化模式；(3)评估水平和变化。三者中的后一者是前两者的变体，因为过程(变化)和状态(水平)之间的联系依赖于对时间序列的清楚理解。指出纵向研究的所有这些特征都是相互关联的，这可能是陈词滥调。然而，对这些因素中的一种或另一种的重视通常决定了研究的设计和分析，并确定了一种疾病或过程的自然病程中适当推断的持续时间和位置。前一点被Sherrill和Viegi的(2)时间序列程序(关注个体随时间的反应)和当关注群体平均趋势时的纵向程序之间的区别所强调。Dockery和Brunekreef(3)在他们的陈述中特别清楚地说明了后一点。在本文中，我们简要总结了对这三个要素进行研究的各种方式。我们讨论了这些方法在解释设计的“纵向”特征和预期推理方面的优势，并讨论了纵向研究的推理与横断面设计的推理的好处。由于人们普遍同意，在流行病学(观察性)研究(4，5)中，适当的时间序列的指定是有效因果推断的必要条件，因此我们仅根据与特定类型的研究相关的推断的时间框架来讨论时间序列。最后，我们的讨论仅限于连续数据的应用，因为肺功能的纵向流行病学分析的主旨是在这个方向上(与离散的分类，如“有病的”/“无病的”形成对比)。
In their introductory overview, Weiss and Ware (1) list six advantages of longitudinal research. Three of these seem particularly important in relation to interpretation and, to some extent, these three subsume the other strengths: (1) understanding the temporal order of events; (2) observation of individual patterns of change; and (3) assessment of both level and change. The latter of the three is a variation of the first two, since the linkage between process (change) and state (level)depends upon a clear understanding of the temporal sequence. It is perhaps trite to note that all of these characteristics of longitudinal studies are interrelated. However, the emphasis on one or another of these elements usually determines the study design and analysis and defines the duration of time and the place in the natural history of a disease or a process over which inference is appropriate. The former point is highlighted by Sherrill and Viegi's(2) distinction between time seriesprocedures (focus on the individual's response over time) and longitudinal procedures when group mean trends are of interest. Dockery and Brunekreef (3) make the latter point particularly clear in their presentation. In this paper, we briefly summarize the various ways that studies have focused on these three elements. We discuss the strengths of the approaches with regard to interpretation of the "longitudinal" characteristics of the design and the intended inference, and wediscuss the benefits of inference from longitudinal studies versus those that can be made from cross-sectional designs. Since there is general agreement that specification of proper temporal sequence is a sine qua non as a criterion for valid causal inference in epidemiologic (observational) studies (4, 5), we discuss temporal sequence only in terms of the time frame over which inference is relevant for specific types of studies. Finally, we confine our discussion to applications for continuous data, since the major thrust of longitudinal epidemiologic analysis of lung function is in this direction (in contrast to discrete classifications such as "diseased"/"nondiseased").
登录
查看更多内容
使用自回归模型分析流行病学研究中的纵向数据。
影响因子：
2
作者：
Rosner,B;Muñoz,A;Tager,I;Speizer,F;Weiss,S
通讯作者：
Weiss,S
纵向和横向确定的肺活量测定年度变化的不可比性。
DOI：
10.1164/arrd.1982.125.5.544
发表时间：
1982
期刊：
The American review of respiratory disease
影响因子：
--
作者：
Glindmeyer,HW;Diem,JE;Jones,RN;Weill,H
通讯作者：
Weill,H
用于拟合用于分析肺功能数据的灵活模型的样条和平滑方法。
DOI：
10.1164/ajrccm/154.6_pt_2.s223
发表时间：
1996
期刊：
American journal of respiratory and critical care medicine.
影响因子：
--
作者：
Wypij,D
通讯作者：
Wypij,D
DOI：
10.1164/arrd.1984.130.3.380
发表时间：
1984
期刊：
The American review of respiratory disease
影响因子：
--
作者：
Higgins,MW;Keller,JB;Landis,JR;Beaty,TH;Burrows,B;Demets,D;Diem,JE;Higgins,IT;Lakatos,E;Lebowitz,MD
通讯作者：
Lebowitz,MD
对纵向肺功能数据进行建模。
DOI：
10.1164/ajrccm/154.6_pt_2.s217
发表时间：
1996
期刊：
American journal of respiratory and critical care medicine.
影响因子：
--
作者：
Sherrill,D;Viegi,G
通讯作者：
Viegi,G