Multivariate meta‐analysis with an increasing number of parameters

Meta‐analysis can average estimates of multiple parameters, such as a treatment's effect on multiple outcomes, across studies. Univariate meta‐analysis (UVMA) considers each parameter individually, while multivariate meta‐analysis (MVMA) considers the parameters jointly and accounts for the correlation between their estimates. The performance of MVMA and UVMA has been extensively compared in scenarios with two parameters. Our objective is to compare the performance of MVMA and UVMA as the number of parameters, p, increases. Specifically, we show that (i) for fixed‐effect (FE) meta‐analysis, the benefit from using MVMA can substantially increase as p increases; (ii) for random effects (RE) meta‐analysis, the benefit from MVMA can increase as p increases, but the potential improvement is modest in the presence of high between‐study variability and the actual improvement is further reduced by the need to estimate an increasingly large between study covariance matrix; and (iii) when there is little to no between‐study variability, the loss of efficiency due to choosing RE MVMA over FE MVMA increases as p increases. We demonstrate these three features through theory, simulation, and a meta‐analysis of risk factors for non‐Hodgkin lymphoma.     

Nonindependence and sensitivity analyses in ecological and evolutionary meta‐analyses

Meta‐analysis is an important tool for synthesizing research on a variety of topics in ecology and evolution, including molecular ecology, but can be susceptible to nonindependence. Nonindependence can affect two major interrelated components of a meta‐analysis: (i) the calculation of effect size statistics and (ii) the estimation of overall meta‐analytic estimates and their uncertainty. While some solutions to nonindependence exist at the statistical analysis stages, there is little advice on what to do when complex analyses are not possible, or when studies with nonindependent experimental designs exist in the data. Here we argue that exploring the effects of procedural decisions in a meta‐analysis (e.g. inclusion of different quality data, choice of effect size) and statistical assumptions (e.g. assuming no phylogenetic covariance) using sensitivity analyses are extremely important in assessing the impact of nonindependence. Sensitivity analyses can provide greater confidence in results and highlight important limitations of empirical work (e.g. impact of study design on overall effects). Despite their importance, sensitivity analyses are seldom applied to problems of nonindependence. To encourage better practice for dealing with nonindependence in meta‐analytic studies, we present accessible examples demonstrating the impact that ignoring nonindependence can have on meta‐analytic estimates. We also provide pragmatic solutions for dealing with nonindependent study designs, and for analysing dependent effect sizes. Additionally, we offer reporting guidelines that will facilitate disclosure of the sources of nonindependence in meta‐analyses, leading to greater transparency and more robust conclusions.     

Transgenerational effects benefit offspring across diverse environments: a meta‐analysis in plants and animals

The adaptive value of transgenerational effects (the ancestor environmental effects on offspring) in changing environments has received much attention in recent years, but the related empirical evidence remains equivocal. Here, we conducted a meta‐analysis summarising 139 experimental studies in plants and animals with 1170 effect sizes to investigate the generality of transgenerational effects across taxa, traits, and environmental contexts. It was found that transgenerational effects generally enhanced offspring performance in response to both stressful and benign conditions. The strongest effects are in annual plants and invertebrates, whereas vertebrates appear to benefit mostly under benign conditions, and perennial plants show hardly any transgenerational responses at all. These differences among taxonomic/life‐history groups possibly reflect that vertebrates can avoid stressful conditions through their mobility, and longer‐lived plants have alternative strategies. In addition to environmental contexts and taxonomic/life‐history groups, transgenerational effects also varied among traits and developmental stages of ancestors and offspring, but the effects were similarly strong across three generations of offspring. By way of a more comprehensive data set and a different effect size, our results differ from those of a recent meta‐analysis, suggesting that transgenerational effects are widespread, strong and persistent and can substantially impact the responses of plants and animals to changing environments.     

Bivariate random‐effects meta‐analysis models for diagnostic test accuracy studies using arcsine‐based transformations

Diagnostic or screening tests are widely used in medical fields to classify patients according to their disease status. Several statistical models for meta‐analysis of diagnostic test accuracy studies have been developed to synthesize test sensitivity and specificity of a diagnostic test of interest. Because of the correlation between test sensitivity and specificity, modeling the two measures using a bivariate model is recommended. In this paper, we extend the current standard bivariate linear mixed model (LMM) by proposing two variance‐stabilizing transformations: the arcsine square root and the Freeman–Tukey double arcsine transformation. We compared the performance of the proposed methods with the standard method through simulations using several performance measures. The simulation results showed that our proposed methods performed better than the standard LMM in terms of bias, root mean square error, and coverage probability in most of the scenarios, even when data were generated assuming the standard LMM. We also illustrated the methods using two real data sets.     

Plant traits and extinction in urban areas: a meta‐analysis of 11 cities

Aim Urban environments around the world share many features in common, including the local extinction of native plant species. We tested the hypothesis that similarity in environmental conditions among urban areas should select for plant species with a particular suite of traits suited to those conditions, and lead to the selective extinction of species lacking those traits. Location Eleven cities with data on the plant species that persisted and those that went locally extinct within at least the last 100 years following urbanization. Methods We compiled data on 11 plant traits for 8269 native species in the 11 cities and used hierarchical logistic regression models to identify the degree to which traits could distinguish species that persisted from those that went locally extinct in each city. The trait effects from each city were then combined in a meta‐analysis. Results The cities fell into two groups: those with relatively low rates of extinction (less than 0.05% species per year – Adelaide, Hong Kong, Los Angeles, San Diego and San Francisco), for which no traits reliably predicted the pattern of extinction, and those with higher rates of extinction (> 0.08% species per year – Auckland, Chicago, Melbourne, New York, Singapore and Worcester, MA), where short‐statured, small‐seeded plants were more likely to go extinct. Main conclusions Our analysis reveals patterns in trait selectivity consistent with local studies, suggesting some consistency in trait selection by urbanization. Overall, however, few traits reliably predicted the pattern of plant extinction across cities, making it difficult to identify a priori the extinction‐prone species most likely to be affected by urban expansion.     

The association of six non‐synonymous variants in three DNA repair genes with hepatocellular carcinoma risk: a meta‐analysis

Hepatocellular carcinoma is a complex polygenic disease. Despite the huge advances in genetic epidemiology, it still remains a challenge to unveil the genetic architecture of hepatocellular carcinoma. We, therefore, decided to meta‐analytically assess the association of six non‐synonymous coding variants from XRCC1, XRCC3 and XPD genes with hepatocellular carcinoma risk by pooling the results of 20 English articles. This meta‐analysis was conducted according to the PRISMA statement, and data collection was independently completed in duplicate. In overall analyses, the minor alleles of four variants, Arg280His (odds ratio, 95% confidence interval, P: 1.37, 1.13–1.66, 0.001), Thr241Met (1.93, 1.17–3.20, 0.011), Asp312Asn (1.22, 1.08–1.38, 0.001) and Lys751Gln (1.42, 1.02–1.97, 0.038), were associated with the significant risk for hepatocellular carcinoma. There were low probabilities of publication bias for all variants. Subgroup analyses revealed significant association of XRCC1 gene Arg399Gln with hepatocellular carcinoma in Chinese especially from south China (odds ratio, 95% confidence interval, P: 1.57, 1.16–2.14, 0.004), in larger studies (1.48, 1.11–1.98, 0.007) and in studies with population‐based controls (1.33, 1.06–1.68, 0.016). Taken together, our findings demonstrated that XPD gene Asp312Asn and XRCC1 gene Arg399Gln might be candidate susceptibility loci for hepatocellular carcinoma. Considering the ubiquity of genetic heterogeneity, further validation in a broad range of ethnic populations is warranted.     

Interactions between climate and habitat loss effects on biodiversity: a systematic review and meta‐analysis

Climate change and habitat loss are both key threatening processes driving the global loss in biodiversity. Yet little is known about their synergistic effects on biological populations due to the complexity underlying both processes. If the combined effects of habitat loss and climate change are greater than the effects of each threat individually, current conservation management strategies may be inefficient and at worst ineffective. Therefore, there is a pressing need to identify whether interacting effects between climate change and habitat loss exist and, if so, quantify the magnitude of their impact. In this article, we present a meta‐analysis of studies that quantify the effect of habitat loss on biological populations and examine whether the magnitude of these effects depends on current climatic conditions and historical rates of climate change. We examined 1319 papers on habitat loss and fragmentation, identified from the past 20 years, representing a range of taxa, landscapes, land‐uses, geographic locations and climatic conditions. We find that current climate and climate change are important factors determining the negative effects of habitat loss on species density and/or diversity. The most important determinant of habitat loss and fragmentation effects, averaged across species and geographic regions, was current maximum temperature, with mean precipitation change over the last 100 years of secondary importance. Habitat loss and fragmentation effects were greatest in areas with high maximum temperatures. Conversely, they were lowest in areas where average rainfall has increased over time. To our knowledge, this is the first study to conduct a global terrestrial analysis of existing data to quantify and test for interacting effects between current climate, climatic change and habitat loss on biological populations. Understanding the synergistic effects between climate change and other threatening processes has critical implications for our ability to support and incorporate climate change adaptation measures into policy development and management response.     

Roads and bats: a meta‐analysis and review of the evidence on vehicle collisions and barrier effects

相似文献   

A matrix‐based method of moments for fitting the multivariate random effects model for meta‐analysis and meta‐regression

Multivariate meta‐analysis is becoming more commonly used. Methods for fitting the multivariate random effects model include maximum likelihood, restricted maximum likelihood, Bayesian estimation and multivariate generalisations of the standard univariate method of moments. Here, we provide a new multivariate method of moments for estimating the between‐study covariance matrix with the properties that (1) it allows for either complete or incomplete outcomes and (2) it allows for covariates through meta‐regression. Further, for complete data, it is invariant to linear transformations. Our method reduces to the usual univariate method of moments, proposed by DerSimonian and Laird, in a single dimension. We illustrate our method and compare it with some of the alternatives using a simulation study and a real example.     

A network meta‐analysis of the effects of psychotherapies,pharmacotherapies and their combination in the treatment of adult depression

No network meta‐analysis has examined the relative effects of psychotherapies, pharmacotherapies and their combination in the treatment of adult depression, while this is a very important clinical issue. We conducted systematic searches in bibliographical databases to identify randomized trials in which a psychotherapy and a pharmacotherapy for the acute or long‐term treatment of depression were compared with each other, or in which the combination of a psychotherapy and a pharmacotherapy was compared with either one alone. The main outcome was treatment response (50% improvement between baseline and endpoint). Remission and acceptability (defined as study drop‐out for any reason) were also examined. Possible moderators that were assessed included chronic and treatment‐resistant depression and baseline severity of depression. Data were pooled as relative risk (RR) using a random‐effects model. A total of 101 studies with 11,910 patients were included. Depression in most studies was moderate to severe. In the network meta‐analysis, combined treatment was more effective than psychotherapy alone (RR=1.27; 95% CI: 1.14‐1.39) and pharmacotherapy alone (RR=1.25; 95% CI: 1.14‐1.37) in achieving response at the end of treatment. No significant difference was found between psychotherapy alone and pharmacotherapy alone (RR=0.99; 95% CI: 0.92‐1.08). Similar results were found for remission. Combined treatment (RR=1.23; 95% CI: 1.05‐1.45) and psychotherapy alone (RR=1.17; 95% CI: 1.02‐1.32) were more acceptable than pharmacotherapy. Results were similar for chronic and treatment‐resistant depression. The combination of psychotherapy and pharmacotherapy seems to be the best choice for patients with moderate depression. More research is needed on long‐term effects of treatments (including cost‐effectiveness), on the impact of specific pharmacological and non‐pharmacological approaches, and on the effects in specific populations of patients.     

Psychological interventions to reduce positive symptoms in schizophrenia: systematic review and network meta‐analysis

Psychological treatments are increasingly regarded as useful interventions for schizophrenia. However, a comprehensive evaluation of the available evidence is lacking and the benefit of psychological interventions for patients with current positive symptoms is still debated. The present study aimed to evaluate the efficacy, acceptability and tolerability of psychological treatments for positive symptoms of schizophrenia by applying a network meta‐analysis approach, that can integrate direct and indirect comparisons. We searched EMBASE, MEDLINE, PsycINFO, PubMed, BIOSIS, Cochrane Library, World Health Organization's International Clinical Trials Registry Platform and ClinicalTrials.gov for randomized controlled trials of psychological treatments for positive symptoms of schizophrenia, published up to January 10, 2018. We included studies on adults with a diagnosis of schizophrenia or a related disorder presenting positive symptoms. The primary outcome was change in positive symptoms measured with validated rating scales. We included 53 randomized controlled trials of seven psychological interventions, for a total of 4,068 participants receiving the psychological treatment as add‐on to antipsychotics. On average, patients were moderately ill at baseline. The network meta‐analysis showed that cognitive behavioural therapy (40 studies) reduced positive symptoms more than inactive control (standardized mean difference, SMD=?0.29; 95% CI: –0.55 to ?0.03), treatment as usual (SMD=?0.30; 95% CI: –0.45 to ?0.14) and supportive therapy (SMD=?0.47; 95% CI: –0.91 to ?0.03). Cognitive behavioural therapy was associated with a higher dropout rate compared with treatment as usual (risk ratio, RR=0.74; 95% CI: 0.58 to 0.95). Confidence in the estimates ranged from moderate to very low. The other treatments contributed to the network with a lower number of studies. Results were overall consistent in sensitivity analyses controlling for several factors, including the role of researchers’ allegiance and blinding of outcome assessor. Cognitive behavior therapy seems to be effective on positive symptoms in moderately ill patients with schizophrenia, with effect sizes in the lower to medium range, depending on the control condition.