%0 Journal Article
%T Profile Likelihood Tests for Common Risk Ratios in Meta-Analysis Studies
%A Chukiat Viwatwongkasem
%A Khanokporn Donjdee
%A Tantanut Poodphraw
%J Open Journal of Statistics
%P 915-930
%@ 2161-7198
%D 2018
%I Scientific Research Publishing
%R 10.4236/ojs.2018.86061
%X <span style=\"font-family:Verdana;\">It is well-known that the
power of Cochran¡¯s </span><i><span style=\"font-family:Verdana;\">Q</span></i><span style=\"font-family:Verdana;\"> test to assess the presence of heterogeneity among
treatment effects in a clinical meta-analysis is low due to the small number of
studies combined. Two modified tests (</span><i><span style=\"font-family:Verdana;\">PL</span></i><span style=\"font-family:Verdana;\">1</span><i><span style=\"font-family:Verdana;\">, PL</span></i><span style=\"font-family:Verdana;\">2) were proposed by replacing the profile maximum
likelihood estimator (PMLE) into the variance formula of logarithm of risk
ratio in the standard chi-square test statistic for testing the null common
risk ratios across all </span><i><span style=\"font-family:Verdana;\">k</span></i><span><span style=\"font-family:Verdana;\"> studies (</span><em><span style=\"font-family:Verdana;\">i</span></em><span style=\"font-family:Verdana;\"> = 1, L, </span><em><span style=\"font-family:Verdana;\">k</span></em></span><span><span><span style=\"font-family:Verdana;\">). The simply naive test (</span><i><span style=\"font-family:Verdana;\">SIM</span></i><span><span style=\"font-family:Verdana;\">)
as another comparative candidate has considerably arisen. The performance of
tests in terms of type I error rate under the null hypothesis and power of test
under the random effects hypothesis was done via a simulation plan with various
combinations of significance levels, numbers of studies, sample sizes in
treatment and control arms, and true risk ratios as effect sizes of interest.
The results indicated that for moderate to large study sizes (</span><em><span style=\"font-family:Verdana;\">k&nbsp;</span></em><span style=\"white-space:nowrap;font-family:Verdana;\">¡Ý 16</span><span style=\"font-family:Verdana;\">)&nbsp;</span></span></span><span><span style=\"font-family:Verdana;\">in combination with
moderate to large sample sizes&nbsp;</span><span style=\"white-space:normal;\"><span style=\"font-family:Verdana;\">(</span><img src=\"Edit_4424cc50-e4d1-4fb1-87f5-f260059ec34f.bmp\" alt=\"\" /></span><em style=\"font-family:Verdana;white-space:normal;\"><span style=\"font-family:Verdana;\">&nbsp;</span></em><span style=\"font-family:Verdana;\">¡Ý 50</span><span style=\"white-space:normal;font-family:Verdana;\">)</span></span><span><span style=\"font-family:Verdana;\">, three tests (</span><i><span style=\"font-family:Verdana;\">PL</span></i><span style=\"font-family:Verdana;\">1</span><i><span style=\"font-family:Verdana;\">,
%K Profile Likelihood Test
%K Cochran Q Test
%K Meta-Analysis
%K Heterogeneity
%K Risk Ratios
%U http://www.scirp.org/journal/PaperInformation.aspx?PaperID=89141