Meta-Analysis: Some Clinical and Statistical Contributions in Several Medical Disciplines

摘要

Meta-analysis in its present form of statistically integrating information from several studies all with a common underlying theme has been around for over 25 years. The medical field has seen many attempts by many investigators to pull summary data together from various sources within a discipline with the goal of making some definitive statement about the state of the science in that discipline. Likewise authors of manuscripts in the background and rationale section of their paper always summarize what they believe to be the state of affairs up to the time of the presentation of their own results in that particular paper. The new data and results they present in their current publication is an attempt to update the progress in that field. Thus in a sense they have performed a partial meta-analysis of summarizing information from the past, presenting their added contribution and thus updating the knowledge base. They have not quite integrated past data in a rigorous statistical way with their new data, but have merely used the data history to justify their current research which pretty much stands on its own. Thus meta-analysis is an after the fact attempt to pull together the current knowledge base whether it be publications or raw data and present a statistical synthesis of all the information and reach a conclusion as to the best treatment or intervention strategy based on all these past contributions. Now it's time to look back at some of these meta-analyses and determine what contributions, if any, they have made to the knowledge base within certain medical disciplines. Many disciplines including psychiatry have been visited by meta-analysis. One now examines some of these studies in the areas of oncology, orthopedics, psychiatry, pediatrics and cardiology. The purpose is to determine, given the information presented, what contributions, statistical challenges and peripheral issues in these disciplines have been brought to light in these meta-analyses.