Testing small study effects in multivariate meta-analysis

摘要: Small study effects occur when smaller studies show different, often larger, treatment than large ones, which may threaten the validity of systematic reviews and meta-analyses. The most well-known reasons for small include publication bias, outcome reporting clinical heterogeneity. Methods to account in univariate meta-analysis have been extensively studied. However, detecting a multivariate setting remains an untouched research area. One complications is that different types selection processes can be involved outcomes. For example, some completely unpublished while others selectively report multiple In this paper, we propose score test as overall meta-analysis. Two detailed case are given demonstrate advantage proposed over various naive applications tests practice. Through simulation studies, found retain nominal Type I error rates with considerable power moderate sample size settings. Finally, also evaluate concordance between application by evaluating 44 outcomes from Cochrane Database.