Undue reliance on I
2 in assessing heterogeneity may mislead

Rücker, Gerta; Schwarzer, Guido; Carpenter, James; Schumacher, Martin

doi:10.1186/1471-2288-8-79

Cited by 890 publications

(662 citation statements)

References 24 publications

Supporting

Mentioning

641

Contrasting

Unclassified

Order By: Relevance

“…21 We assessed the heterogeneity of study results by the use of t 2 metric. 19,22 Possible sources of heterogeneity were explored through a subgroup and meta-regression technique. The effect of small studies in publication bias was explored by the Egger' s test.…”

Section: Discussionmentioning

confidence: 99%

MRSA Colonization and Risk of Infection in the Neonatal and Pediatric ICU: A Meta-analysis

Zervou

Zacharioudakis

Ziakas³

et al. 2014

Pediatrics

101

View full text Add to dashboard Cite

BACKGROUND AND OBJECTIVE: Methicillin-resistant Staphylococcus aureus (MRSA) is a significant cause of morbidity and mortality in NICUs and PICUs. Our objective was to assess the burden of MRSA colonization on admission, study the time trends, and examine the significance of MRSA colonization in this population. METHODS:PubMed and Embase databases were consulted. Studies that reported prevalence of MRSA colonization on ICU admission were selected. Two authors independently extracted data on MRSA colonization and infection. RESULTS:We identified 18 suitable articles and found an overall prevalence of MRSA colonization of 1.9% (95% confidence interval [CI] 1.3%-2.6%) on admission to the NICU or PICU, with a stable trend over the past 12 years. Interestingly, 5.8% (95% CI 1.9%-11.4%) of outborn neonates were colonized with MRSA on admission to NICU, compared with just 0.2% (95% CI 0.0%-0.9%) of inborn neonates (P = .01). The pooled acquisition rate of MRSA colonization was 4.1% (95% CI 1.2%-8.6%) during the NICU and PICU stay and was as high as 6.1% (95% CI 2.8%-10.6%) when the NICU population was studied alone. There was a relative risk of 24.2 (95% CI 8.9-66.0) for colonized patients to develop a MRSA infection during hospitalization. CONCLUSIONS:In the NICU and PICU, there are carriers of MRSA on admission, and MRSA colonization in the NICU is almost exclusively associated with outborn neonates. Importantly, despite infection control measures, the acquisition rate is high, and patients colonized with MRSA on admission are more likely to suffer a MRSA infection during hospitalization.

show abstract

Section: Discussionmentioning

confidence: 99%

MRSA Colonization and Risk of Infection in the Neonatal and Pediatric ICU: A Meta-analysis

Zervou

Zacharioudakis

Ziakas³

et al. 2014

Pediatrics

101

View full text Add to dashboard Cite

show abstract

“…17 We ran a sensitivity analysis on our primary outcome using a Bayesian random-eff ects model with Monte Carlo Markov chain simulations of variability. 18 We calculated the τ² statistic to assess the proportion of overall variation attributable to betweenstudy heterogeneity, because this measure is less sensitive to the number of studies, 19 and explored potential sources of heterogeneity with univariate sensitivity analyses with χ² to estabish the potential eff ect of the following covariates: DST confi rmatory testing (≤50% or >50% of the cohort), treatment duration (≤ 18 months or >18 months), use of injectable drugs (≤80% or >80% of the cohort), age (≤ 5 years or >5 years), and HIV status (≤30% or >30% of the cohort). All p values are two-sided, and a p value of less than 0·05 was regarded as signifi cant.…”

Section: Discussionmentioning

confidence: 99%

Treatment outcomes for children with multidrug-resistant tuberculosis: a systematic review and meta-analysis

Ettehad

Schaaf

Seddon

et al. 2012

The Lancet Infectious Diseases

155

114

View full text Add to dashboard Cite

show abstract

“…We measured statistical heterogeneity using the between-study variance t 2 (8,9) and performed subgroup and sensitivity analysis to account for potential sources of heterogeneity. We used t 2 statistic to describe the underlying between-study variability, given that it does not systematically increase with the number or size of studies included in the analysis, whereas I 2 may increase artificially with accumulating evidence (9). The effect of carrier state on ensuing infections was reported as risk ratio (RR) estimate and random effects confidence intervals (95% CI).…”

Section: Methodsmentioning

confidence: 99%

MRSA and VRE Colonization in Solid Organ Transplantation: A Meta-Analysis of Published Studies

Ziakas

Pliakos

Zervou

et al. 2014

American Journal of Transplantation

View full text Add to dashboard Cite

The burden of methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant enterococcus (VRE) colonization among the increasing number of solid organ transplant patients has not been systematically explored. We searched PubMed and EMBASE for pertinent articles, performed a metaanalysis of prevalence across eligible studies and estimated the risk of ensuing MRSA or VRE infections relative to colonization status. We stratified effects in the pretransplant and posttransplant period. Twentythree studies were considered eligible. Seventeen out of 23 (74%) referred to liver transplants. Before transplantation, the pooled prevalence estimate for MRSA and VRE was 8.5% (95% confidence interval [CI] 3.2-15.8) and 11.9% (95% CI 6.8-18.2), respectively. MRSA estimate was influenced by small studies and was lower (4.0%; 95% CI 0.4-10.2) across large studies (>200 patients). After transplantation, the prevalence estimates were 9.4% (95% CI 3.0-18.5) for MRSA and 16.2% (95% CI 10.7-22.6) for VRE. Pretransplant as well as posttransplant MRSA colonization significantly increased the risk for MRSA infections (pooled risk ratio [RR] 5.51; 95% CI 2. respectively). Pretransplant and posttransplant VRE colonization were also associated with significant risk of VRE infection (RR 6.65; respectively). Solid organ transplantation is a high-risk setting for MRSA and VRE colonization, and carrier state is associated with infection. Upgraded focus in prevention and eradication strategies is warranted.

show abstract

Undue reliance on I 2 in assessing heterogeneity may mislead

Abstract: Background: The heterogeneity statistic I 2 , interpreted as the percentage of variability due to heterogeneity between studies rather than sampling error, depends on precision, that is, the size of the studies included.

Cited by 890 publications

References 24 publications

MRSA Colonization and Risk of Infection in the Neonatal and Pediatric ICU: A Meta-analysis

MRSA Colonization and Risk of Infection in the Neonatal and Pediatric ICU: A Meta-analysis

Treatment outcomes for children with multidrug-resistant tuberculosis: a systematic review and meta-analysis

MRSA and VRE Colonization in Solid Organ Transplantation: A Meta-Analysis of Published Studies

Contact Info

Product

Resources

About