Meta-analysis collects the results of primary studies to answer specific research questions in a quantitative manner. By taking account of information on the same topic from many sources, a meta-analysis increases the statistical and scientific credibility of a result. The project will enhance the ability of clinical research to produce, organize, and disseminate better information to guide health and medical practice. Using data bases already available and results acquired through new literature searches and analyses, the project will evaluate diagnostic and therapeutic methods, assess the quality and cost of both primary studies and meta- analyses, use empirical analyses to appraise the impact of studies on medical practice, and develop, improve, and standardize methods for carrying out, analyzing, and combining information from clinical and other studies. The project assesses and improves the quality of clinical trials in health and medicine by empirical studies that show how features of research designs, including randomization, placebos, blinding, stopping rules and multifactorial designs influence reliability of study results. Locating reports of trials in the literature requires strong search methods. The project studies a new approach using citation analysis that may be cost-effective in producing more comprehensive results than other methods. By carrying out new and updating previous meta-analyses for diagnostic and therapeutic maneuvers, the project will contribute directly to our knowledge of specific diagnostic procedures as well as methods for their evaluation. To explore the impact of published reports of randomized control trials and meta-analyses on clinical practice, the investigators will use a variety of databases to appraise the effect of publications of assessments on changes in specific therapies and diagnostic procedures. Extending meta-analysis to the dental field brings assessment of quality of experiments to a new important area, and thus gives an opportunity to strengthen primary studies. New statistical Methods are required specifically for meta- analysis in the areas of health and medicine. Their development will improve the ability of investigators to combine information on rates and proportions, data from follow-up studies, and measures of sensitivity and specificity, as well as to deal with problems of heterogeneity of studies, missing data, and outlying values.
Showing the most recent 10 out of 15 publications