Importance of meta-analysis and practical obstacles in oncological and epidemiological studies: statistics very close but also far!

Studies of epidemiological and prognostic factors are very important for oncology practice. There is a rapidly increasing amount of research and resultant knowledge in the scientific literature. This means that health professionals have major challenges in accessing relevant information and they increasingly require best available evidence to make their clinical decisions. Meta-analyses of prognostic and other epidemiological factors are very practical statistical approaches to define clinically important parameters. However, they also feature many obstacles in terms of data collection, standardization of results from multiple centers, bias, and commentary for intepretation. In this paper, the obstacles of meta-analysis are briefly reviewed, and potential problems with this statistical method are discussed.