Handbook of Epidemiology
Description
This 3-volume reference covers the entire field of epidemiology, from statistical methods and study design, to specialized areas such as molecular epidemiology, and applications in clinical medicine and health services research. This updated edition of the Handbook of Epidemiology adds 22 new chapters on: History of Epidemiological Methods and Concepts, Cluster Randomized Trials, Internet-Based Epidemiology, Misclassification, Sensitivity Analysis and Bias Analysis, Emergency and Disaster Health Surveillance, Statistical Inference, Data Management in Epidemiology, Visual Display of Quantitative Information, Bayesian Methods in Epidemiology, Generalized Estimating Equations, Directed Acyclic Graphs, Life Course Epidemiology, Molecular Epidemiology, Physical Activity, Radiation Epidemiology, Epidemiology of Obesity, Epidemiology of Respiratory Allergies and Asthma, Epidemiology of Dental Diseases, Epidemiology of Digestive Diseases, Psychiatric Disorders, Epidemiology of Diabetes.
All other chapters are extensively revised from the 1st edition. This is a reference for epidemiological researchers and graduate students in public health.
Reviews / Votes
From the reviews:
"The book represents a comprehensive reference source on practical epidemiology. . the handbook covers a very wide spectrum of problems and is a very good reference material." (Tanya Kostova-Vassilevska, Zentralblatt MATH, Vol. 1094 (20), 2006)
More details
Other editions
Additional editions
Persons
Iris Pigeot is Director of the Bremen Institute for Prevention Research and Social Medicine, Germany.
Content
II. 1 Sample Size Determination in Epidemiologic Studies
Introduction
When planning a research project an epidemiologist must consider how many subjects should be studied.While factors such as available budget certainly present constraints on the maximumnumber of subjects that might actually be included in a study, statistical considerations are extremelyimportant.Toaddress the statistical questions about appropriate sample size, the researchermust .rst specify the study design, the nature of the outcome variable, the aims of the study, the planned analysis method, and the expected results of the study. Is the goal of the study to distinguish between hypotheses about the value of a parameter or function of parameters, or is the goal to provide a con.dence interval estimate of a parameter such as the odds ratio or relative risk? This chapter is organized as follows.
We introduce the issue of how to choose sample size for estimation of a parameter or for a hypothesis test regarding a parameter in the context of one-sample studies in which it is desired to estimate or test a population proportion. We continue on to two-sample studies involving comparisons between two proportions, and one and two-sample studies involving estimation or testing of population means.We conclude with a section on sample size for logistic regression. In this chapter we will provide a brief introduction to power and sample size computation and only address sample size issues for a few of the procedures that are most commonly used in epidemiologic research.
However, we do hope that the reader will gain a sense for what one can accomplish by planning a study with appropriate attention to sample size considerations. A focus on sample size considerations when the study is .rst being planned is critical for the ultimate likelihood that a study proposal is accepted for funding and that the .nal manuscript will be accepted for publication. To ignore the issue of sample size would greatly increase the likelihood of embarking on a costly and time-consuming epidemiologic studywith little likelihood of .nding any de.nitive results.
One Group Designs, Inferences About Proportions
The simplest study design is one in which interest focuses on results for a single group. One is often interested in making inferences about the value of a population proportion. In this section we will illustrate how to choose sample size for the following examples:
Example 1 . A district medical of.cer seeks to estimate the proportion of children in the district receiving appropriate childhood vaccinations. Assuming a simple randomsample is to be selected froma community, how many children must be studied if the resulting estimate is to fall within 10 percentage points of the true proportion with 95% con.dence?
Example 2 . Consider the information given in Example 1, only this time we will determine the sample size necessary to estimate the proportion vaccinated in the population to within 10% (not 10 percentage points) of the true value. ,
Example 3 . During a virulent outbreak of neonatal tetanus, health workers wish to determine whether the rate is decreasing after a period during which it had risen to a level of 150 cases per thousand live births. What sample size is necessary to test the null hypothesis that the population proportion is 0.15 at the 0.05 level if it is desired to have a 90% probability of detecting a decrease to a rate of 100 per thousand if that were the true proportion?
