In this Master Class timely topics in study design of epidemiologic and clinical studies will be addressed. Four renowned faculty members will address advanced study design issues in a seminar format.
Monday August 11, 2014
The Growing Rift Between Epidemiologists and Their Data
In former times, epidemiologists customarily published their data in a way that showed the distributions of key variables and their associations. Those days are mostly gone. From earlier papers, readers could see enough detail in the data to conduct their own analyses. This approach kept everyone informed and careful. Nowadays, it is usual for epidemiologists to present only derived results, typically from regression models, that cannot be checked from the published data. This gulf between epidemiologists and their data is conducive to publishing mistakes and obscures important information that readers should have.
Prof. Kenneth J. Rothman
Distinguished Fellow and VP for Epidemiology Research, RTI Health Solutions, Research Triangle Park, North Carolina, USA and Professor of Epidemiology Boston University School of Public Health, Boston, MA, USA
Tuesday August 12, 2014
Big epidemiology: too big to fail?
Prof. Matthias Egger
Professor of Epidemiology and Public Health and Director, Inst. Social and Preventive Medicine (ISPM), Univ. Bern, Switzerland; visiting Professor of Clinical Epidemiology, Dept. of Social Medicine, Univ. Bristol, UK; visiting Professor of Epidemiology, School of Public Health, Univ. Cape Town, South Africa
Wednesday August 13, 2014
Change as a determinant, determinants of change.
Studying whether change depends on the initial value and the effect of change is of general interest in epidemiological studies with repeated measurements. Simple analyses of these data are often biased because of regression-to-the-mean.
The phenomenon will be presented together with typical examples from the literature. Simple and more advanced methods to correct for the effect of regression-to-the-mean will be proposed.”
Dr. Zoltán Vokó
Associate professor, deputy head of institute
Department of Health Policy & Health Economics
Institute of Economics, Faculty of Social Sciences
Eötvös Loránd University, Budapest, Hungary
Thursday August 14, 2014
The Case-Crossover Design: Applications in Perinatal Epidemiology Research
The case-crossover study is designed to answer the question “Was event X triggered by something unusual that happened just before?” The key feature of the design is that each case serves as its own control. The method is analogous to a crossover experiment viewed retrospectively, except that the investigator does not control when a patient starts being exposed to a potential trigger. The case-crossover study can also be compared with a traditional matched-pair case-control study. In both types of studies, each case has a matched control. In a traditional matched-pair case-control study, the control is a different individual at a similar time. In the matched-pair case-crossover design, the control is the same person at a different time. Decades of applied research have shown that the study design applies best when the exposure is intermittent, the effect of risk is immediate and transient, and the outcome is abrupt. Using the case-crossover design, investigators have been able to assess the extent to which episodes of physical exertion, anger, sexual activity, cocaine use, and alcohol consumption are triggers of myocardial infarctions. Others have applied the method to study triggers of other acute events including car crashes, injury, and adverse drug reactions. A placental abruption, for the most part, is an abrupt event that typically occurs in mid-to-late pregnancy; and is amenable for study using the case-crossover design.
Prof. Michelle A. Williams
Professor and Chair of Epidemiology
Harvard School of Public Health
Boston, MA, USA