| pubmed-article:19390938 | pubmed:abstractText | Meta-analysis of randomized controlled trials is considered to be the methodology that provides the most solid scientific basis for constructing clinical guidelines. It involves systematically collecting the results of similar studies that were conducted to verify similar medical hypotheses and combining these results statistically. In meta-analysis, targeting only those randomized controlled trials with good comparability also provides the meta-analysis with comparability. With the combining of multiple studies and the increased sample size, meta-analysis provides results with higher clarity than those obtained from a single study. In conventional meta-analyses, in addition to estimating the combined effect, the cause of heterogeneity of the effects among studies is usually explored. If multiple studies reveal homogeneous effects, the overall effect is interpretable and generalizability can be suggested; that is, the results can be reproducible even when the study conditions are slightly modified. On the other hand, if the effect cannot be viewed as homogeneous among the studies, it is difficult to interpret the overall effect obtained from a meta-analysis. From the viewpoints of clarity, comparability, and generalizability, meta-analysis and large-scale clinical trials can provide the most valuable evidence among several possible study designs. In this article, the role of meta-analysis in cancer clinical trials is illustrated with the example of adjuvant therapy with UFT in patients with curatively resected rectal cancer, compared with the example of a large-scale clinical trial using oxaliplatin, fluorouracil, and leucovorin as adjuvant treatment for colon cancer. | lld:pubmed |
