The heterogeneity statistic I(2) can be biased in small meta-analyses.

The heterogeneity statistic I(2) can be biased in small meta-analyses.
复制标题

DOI:
10.1186/s12874-015-0024-z
复制
发表时间:
2015-04-14
影响因子:
4
通讯作者:
von Hippel PT
von Hippel PT
中科院分区:
医学3区
文献类型:
--
作者:
von Hippel PT

文献摘要

参考文献

被引文献

相似文献

不同研究的估计效果有所不同,部分原因是随机抽样误差,部分原因是异质性。在荟萃分析中,由异质性引起的方差分数由统计量 I2 估计。我们计算I2的偏倚,重点关注荟萃分析中研究数量较少的情况。小型荟萃分析很常见;在 Cochrane 图书馆中,每次荟萃分析的研究中位数为 7 项或更少。我们使用 Mathematica 软件来计算 I2 的期望和偏差。当研究数量较少时,I2 存在很大的偏差。当异质性的真实分数较小时,偏差为正,但当异质性的真实分数较大时,偏差通常为负。例如,如果有 7 项研究且没有真正的异质性,I2 会平均高估异质性 12 个百分点,但如果有 7 项研究和 80% 的真实异质性,I2 可能会平均低估异质性 28 个百分点。考虑到 Cochrane 图书馆中 I2 估计值的中位数为 21%,12-28 个百分点的偏差并非微不足道。当荟萃分析的研究很少时,应谨慎解释点估计 I2。在小型荟萃分析中,置信区间应补充或取代有偏差的点估计 I2。
Estimated effects vary across studies, partly because of random sampling error and partly because of heterogeneity. In meta-analysis, the fraction of variance that is due to heterogeneity is estimated by the statistic I2. We calculate the bias of I2, focusing on the situation where the number of studies in the meta-analysis is small. Small meta-analyses are common; in the Cochrane Library, the median number of studies per meta-analysis is 7 or fewer. We use Mathematica software to calculate the expectation and bias of I2. I2 has a substantial bias when the number of studies is small. The bias is positive when the true fraction of heterogeneity is small, but the bias is typically negative when the true fraction of heterogeneity is large. For example, with 7 studies and no true heterogeneity, I2 will overestimate heterogeneity by an average of 12 percentage points, but with 7 studies and 80 percent true heterogeneity, I2 can underestimate heterogeneity by an average of 28 percentage points. Biases of 12–28 percentage points are not trivial when one considers that, in the Cochrane Library, the median I2 estimate is 21 percent. The point estimate I2 should be interpreted cautiously when a meta-analysis has few studies. In small meta-analyses, confidence intervals should supplement or replace the biased point estimate I2.
DOI: 10.1093/ije/dys041
发表时间: 2012-06
影响因子: 7.7
作者:
Turner RM;Davey J;Clarke MJ;Thompson SG;Higgins JP
通讯作者: Higgins JP
DOI: 10.1002/sim.1186
发表时间: 2002-06-15
影响因子: 2
作者:
Higgins, JPT;Thompson, SG
通讯作者: Thompson, SG
DOI: 10.1016/j.jspi.2003.09.032
发表时间: 2005-01-01
影响因子: 0.9
作者:
Hartung, J;Knapp, G
通讯作者: Knapp, G
DOI: 10.1037/1082-989x.3.4.486
发表时间: 1998-12-01
影响因子: 7
作者:
Hedges, LV;Vevea, JL
通讯作者: Vevea, JL
DOI: 10.1016/0304-4076(84)90048-4
发表时间: 1984-01-01
影响因子: 6.3
作者:
BOCK, ME;JUDGE, GG;YANCEY, TA
通讯作者: YANCEY, TA