China
Africa
Explore chapters and articles related to this topic
Q
Published in Filomena Pereira-Maxwell, Medical Statistics, 2018
Also known as Cochran's Q test. A test that is used in the context of meta-analysis, to test for heterogeneity of treatment or exposure effect across the different primary studies included in a meta-analysis. This test is similar to the chi-squared test for heterogeneity that is used with Mantel-Haenszel stratification methods. The two tests differ in the way the stratum- or study-level components of the test statistic are weighted, using either Mantel-Haenszel or inverse variance weights. Degrees of freedom for the test are calculated as the number of studies minus 1. The larger the test statistic, the smaller the P-value and the greater the evidence against the null hypothesis that the underlying exposure or treatment effect is the same across the different studies. Due to the test's lack of power, in particular where a meta-analysis encompasses only a small number of studies, less stringent levels of significance are often suggested. The index of heterogeneity (I2) may offer a better assessment in these situations. See also forest plot, Galbraith plot, L'Abbe plot, meta-regression.
Neutrophil-lymphocyte ratio, monocyte-lymphocyte ratio, and platelet-lymphocyte ratio in stroke-associated pneumonia: a systematic review and meta-analysis
Published in Current Medical Research and Opinion, 2023
Mohammed Zawiah, Amer Hayat Khan, Rana Abu Farha, Abubakar Usman, Ahmad Naoras Bitar
Nine studies 19–25,34,35 with 4636 participants provided data suitable for evaluating the relationship between NLR level and SAP incidence. Data showed that the NLR levels of the SAP groups were significantly higher as compared with the non-SAP groups (SMD = 0.88; 95% CI: 0.70–1.07; p-value < .00001). Heterogeneity assessment suggests substantial heterogeneity between the studies included in the NLR subgroup (I2 = 77%). Thus, the random effect model was adopted in the meta-analysis of this subgroup (Figure 2). We conducted a subgroup analysis based on the stroke severity, sample size, and period between the stroke onset and blood sampling. We found no significant differences in the total effect size across different study characteristics (Figures S(1–3)). A Galbraith plot was used to test the heterogeneity source. Three studies21,22,35 were identified as contributors to the heterogeneity (Figures S(4,S5)). After excluding the outlier studies, the heterogeneity was significantly reduced, while the meta-analysis findings did not differ substantially. (SMD = 0.81, 95% CI = 0.68–0.93, p-value < .00001, I2 = 31%). Furthermore, we performed a sensitivity analysis to show that there was no substantial change in the overall effect size after removing each study at a time, suggesting that the meta-analysis result of this subset was robust and credible (Table S(1)).
Performance of adenosine deaminase in detecting paediatric pleural tuberculosis: a systematic review and meta-analysis
Published in Annals of Medicine, 2022
Feiyang Na, Yannan Wang, Hui Yang, Li Guo, Xuan Liang, Donghai Liu, Rongfang Zhang
We used a bivariate random model to compute the pooled sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), diagnostic odds ratio (DOR), and area under the summary receiver operating characteristic curve (AUC) for ADA in detecting paediatric pleural TB, with corresponding 95% confidence intervals (CI). Additionally, we constructed a hierarchical summary receiver operating characteristic (HSROC) plot to summarize the detection performance of ADA among paediatric participants [25]. We considered heterogeneity to be present if the I2 statistic was >75% [26]. The Galbraith plot was used to explore the heterogeneity, and the outlier studies which could be found in this plot showed possible heterogeneity. According to the general expert advice, we used the most common diagnostic threshold (cut-off: 40 IU/L) of ADA for the data analysis [11,15]. Other cut-off values were excluded. The possible existence of publication bias was assessed for using Deek’s funnel plot, and a p-value < 0.10 indicated a possible bias [27].
Cadmium exposure and DNA damage (genotoxicity): a systematic review and meta-analysis
Published in Critical Reviews in Toxicology, 2022
Raju Nagaraju, Ravibabu Kalahasthi, Rakesh Balachandar, Bhavani Shankara Bagepally
Thirteen studies analyzed UCd levels in Cd-exposed and unexposed groups, among which ten studies expressed results as µg/g creatinine (Wegner et al. 2004; Iarmarcovai et al. 2005; Botta et al. 2006; Wang et al. 2011, 2015; Lin et al. 2012; Hambach et al. 2013; Huang et al. 2013; Sciskalska et al. 2014; Su et al. 2019), while the remaining studies in the form of µg/L (Abrahim et al. 2011; Moitra et al. 2015; Yang et al. 2020). The majority of the study found considerably higher UCd among the exposures than the unexposed group. The standardized mean difference in Cd levels between the groups was 0.47 (95% CI, 0.10–0.85) with high heterogeneity (I2 = 94.63) (Figure 2(B)). Further, the funnel and contour-enhanced funnel plot exhibited asymmetry indicating publication bias (p = 0.271) (Supplementary Figure 4). Galbraith plot showed the heterogeneity between the studies as well as identified the presence of outliers (Supplementary Figure 5). Based on leave-one-out analysis, the exclusion of two studies (Abrahim et al. 2011; Moitra et al. 2015) revealed a trend of higher UCd among the Cd-exposed (Supplementary Figure 6).