Modern Meta-Analysis

Modern meta-analyses do more than combine the effect sizes of a series of similar studies. Meta-analyses are currently increasingly applied for any analysis beyond the primary analysis of studies, and for the analysis of big data. This 26-chapter book was written for nonmathematical professionals of medical and health care, in the first place, but, in addition, for anyone involved in any field involving scientific research. The authors have published over twenty innovative meta-analyses from the turn of the century till now. This edition will review the current state of the art, and will use for that purpose the methodological aspects of the authors' own publications, in addition to other relevant methodological issues from the literature.

Are there alternative works in the field? Yes, there are, particularly in the field of psychology. Psychologists have invented meta-analyses in 1970, and have continuously written and updated methodologies. Although very interesting, their work, just like the whole discipline of psychology, is rather explorative in nature, and so is their focus to meta-analysis. As such, they are not particularly involved in confirmatory placebo controlled double blind therapeutic clinical trials, and, despite their overwhelming productions and sometimes expensive software, they never address clinically important subjects like the meta-analysis with diagnostic tests, contrast coefficients, tetrachoric correlations, Bayesian networks, quasi-likelihood modeling, correlation coefficients to z transformations, confounding and interaction assessments, and other clinically relevant subjects. Then, there is the field of epidemiologists. Many of them are from the school of angry young men, who publish shocking news all the time, and JAMA and other publishers are happy to publish it. The reality is, of course, that things are usually not as bad as they seem. Finally, some textbooks, written by professional statisticians, tend to use software programs with miserable menu programs and requiring lots of syntax to be learnt. This is prohibitive to clinical and other health professionals.

The current edition is a must-read textbook written by a very experienced mathematical statistician, and an internist / clinical pharmacologist, and consists, like the authors' previous books, of many examples and step by step analyses using user friendly software like the free META XL from Excel , the SPSS' work bench for automatic data mining entitled SPSS Modeler, the free Konstanz information miner KNIME, and pocket calculator methods, if more convenient. In order for readers to perform their own analyses, SPSS data files are given in extras.springer.com. The book is available as an e-book and chapters can be studied without the need to check the others.   

 

• ISBN: 978-3-319-55894-3
• Editorial: Springer
• Encuadernacion: Cartoné
• Páginas: 270
• Fecha Publicación: 04/05/2017
• Nº Volúmenes: 1
• Idioma: Inglés