Assessing the consistency assumption by exploring treatment by covariate interactions in mixed treatment comparison meta‐analysis: individual patient‐level covariates versus aggregate trial‐level covariates

Mixed treatment comparison (MTC) meta‐analysis allows several treatments to be compared in a single analysis while utilising direct and indirect evidence. Treatment by covariate interactions can be included in MTC models to explore how the covariate modifies the treatment effects. If interactions exist, the assumptions underlying MTCs may be invalidated. For conventional pair‐wise meta‐analysis, important benefits regarding the investigation of such interactions, gained from using individual patient data (IPD) rather than aggregate data (AD), have been described. We aim to compare IPD MTC models including patient‐level covariates with AD MTC models including study‐level covariates. IPD and AD random‐effects MTC models for dichotomous outcomes are specified. Three assumptions are made regarding the interactions (i.e. independent, exchangeable and common interactions). The models are applied to a dataset to compare four drugs for treating malaria (i.e. amodiaquine‐artesunate, dihydroartemisinin‐piperaquine (DHAPQ), artemether‐lumefantrine and chlorproguanil‐dapsone plus artesunate) using the outcome unadjusted treatment success at day 28. The treatment effects and regression coefficients for interactions from the IPD models were more precise than those from AD models. Using IPD, assuming independent or exchangeable interactions, the regression coefficient for chlorproguanil‐dapsone plus artesunate versus DHAPQ was statistically significant and assuming common interactions, the common coefficient was significant; whereas using AD, no coefficients were significant. Using IPD, DHAPQ was the best drug; whereas using AD, the best drug varied. Using AD models, there was no evidence that the consistency assumption was invalid; whereas, the assumption was questionable based on the IPD models. The AD analyses were misleading. Copyright © 2012 John Wiley & Sons, Ltd.     

效应修饰作用分析方法与案例解析系列（三）：个体数据Meta分析中的效应修饰

简要介绍个体数据Meta分析在效应修饰作用方面的独特优势、整体分析思路及现有分析方法,除了常见的Meta回归和亚组分析外,还介绍了利用部分个体数据合并集合水平数据的分析方法,并总结以上方法的报告现状。以“钠-葡萄糖协同转运蛋白2抑制剂对2型糖尿病患者SBP的影响”作为案例,分别展示上述方法在个体数据Meta分析中的实际应用及结果解读,总结各方法的优势和局限性。     

Combining individual patient data and aggregate data in mixed treatment comparison meta‐analysis: Individual patient data may be beneficial if only for a subset of trials

Background Individual patient data (IPD) meta‐analysis is the gold standard. Aggregate data (AD) and IPD can be combined using conventional pairwise meta‐analysis when IPD cannot be obtained for all relevant studies. We extend the methodology to combine IPD and AD in a mixed treatment comparison (MTC) meta‐analysis. Methods The proposed random‐effects MTC models combine IPD and AD for a dichotomous outcome. We study the benefits of acquiring IPD for a subset of trials when assessing the underlying consistency assumption by including treatment‐by‐covariate interactions in the model. We describe three different model specifications that make increasingly stronger assumptions regarding the interactions. We illustrate the methodology through application to real data sets to compare drugs for treating malaria by using the outcome unadjusted treatment success at day 28. We compare results from AD alone, IPD alone and all data. Results When IPD contributed (i.e. either using IPD alone or combining IPD and AD), the chains converged, and we identified statistically significant regression coefficients for the interactions. Using IPD alone, we were able to compare only three of the six treatments of interest. When models were fitted to AD, the treatment effects and regression coefficients for the interactions were far more imprecise, and the chains did not converge. Conclusions The models combining IPD and AD encapsulated all available evidence. When exploring interactions, it can be beneficial to obtain IPD for a subset of trials and to combine IPD with additional AD. Copyright © 2012 John Wiley & Sons, Ltd.     

The detection of major loci by segregation and linkage analysis: A simulation study

Simulated multigenerational pedigrees were analyzed using the program GENPED and POINTER to examine the 1) limits of segregation analysis for detecting single locus, two-allele transmission of a dichotomous trait and 2) accuracy of the parameter estimates. Ten data sets of 30 pedigrees each (approximately 25 persons per pedigree) were simulated. The genotypic penetrance values were varied but the population prevalence of the trait was kept constant at 2%. For some data sets a linked marker locus was also simulated. Previous results had shown that a single major locus could be easily detected when the heterozygote penetrance (f₁) was high or midway between the two homozygote penetrances. In this study, we found a single major locus could not be consistently detected by either method of segregation analysis when f₁ was “low” to “intermediate.” Accuracy of the parameter estimates depended on assumptions about the population prevalence. In those cases where the major locus could not be detected by segregation analysis, linkage to a marker locus could be detected as long as the marker was closely linked and there were not phenocopies in the population. Owing to the limited number of simulations in this study, we cannot generalize these findings. However, they provide a basis for further testing of methods of segregation analysis when factors such as the parameter values, family structure, and ascertainment scheme are varied.     

No Study Left Behind: A Network Meta-Analysis in Non–Small-Cell Lung Cancer Demonstrating the Importance of Considering All Relevant Data

Objective:  To demonstrate the importance of considering all relevant indirect data in a network meta-analysis of treatments for non–small-cell lung cancer (NSCLC).
Methods:  A recent National Institute for Health and Clinical Excellence appraisal focussed on the indirect comparison of docetaxel with erlotinib in second-line treatment of NSCLC based on trials including a common comparator. We compared the results of this analysis to a network meta-analysis including other trials that formed a network of evidence. We also examined the importance of allowing for the correlations between the estimated treatment effects that can arise when analysing such networks.
Results:  The analysis of the restricted network including only trials of docetaxel and erlotinib linked via the common placebo comparator produced an estimated mean hazard ratio (HR) for erlotinib compared with docetaxel of 1.55 (95% confidence interval [CI] 0.72–2.97). In contrast, the network meta-analysis produced an estimated HR for erlotinib compared with docetaxel of 0.83 (95% CI 0.65–1.06). Analyzing the wider network improved the precision of estimated treatment effects, altered their rankings and also allowed further treatments to be compared. Some of the estimated treatment effects from the wider network were highly correlated.
Conclusions:  This empirical example shows the importance of considering all potentially relevant data when comparing treatments. Care should therefore be taken to consider all relevant information, including correlations induced by the network of trial data, when comparing treatments.     

Searching for Indirect Evidence and Extending the Network of Studies for Network Meta-Analysis: Case Study in Venous Thromboembolic Events Prevention Following Elective Total Knee Replacement Surgery

ObjectiveTo evaluate the effect of study identification methods and network size on the relative effectiveness and cost-effectiveness of recommended pharmacological venous thromboembolic events (VTEs) prophylaxis for adult patients undergoing elective total knee replacement surgery in the United Kingdom.MethodsA stepwise literature search specifically designed to identify indirect evidence was conducted to extend the original clinical review from the latest National Institute for Health and Care Excellence (NICE) VTE technology appraisal. Different network sizes or network orders, based on the successive searches, informed three network meta-analyses (NMAs), which were compared with a replicated base case. The resulting comparative estimates were inputted in an economic model to investigate the effect of network size on cost-effectiveness probabilities.ResultsSearches increased the number of indirect comparisons between VTE interventions, progressively widening the relevant network of studies for NMA. Precision around mean relative treatment effects was increased as the network was extended from the base case to first-order NMA, but further extensions had limited effect. Cost-effectiveness analysis results were largely insensitive to variation in clinical inputs from the different NMA orders.ConclusionsNo standard methodology is currently recommended by NICE to identify the most relevant network of studies for NMA. Our study showed that optimizing the identification of studies for NMA can extend the evidence base for analysis and reduce the uncertainty in relative effectiveness estimates. Although in our example network extensions did not affect the acceptability of available treatments in VTE prevention based on cost-effectiveness results, it may in other applications.     

Much ado about nothing: a comparison of the performance of meta-analytical methods with rare events

For rare outcomes, meta-analysis of randomized trials may be the only way to obtain reliable evidence of the effects of healthcare interventions. However, many methods of meta-analysis are based on large sample approximations, and may be unsuitable when events are rare. Through simulation, we evaluated the performance of 12 methods for pooling rare events, considering estimability, bias, coverage and statistical power. Simulations were based on data sets from three case studies with between five and 19 trials, using baseline event rates between 0.1 and 10 per cent and risk ratios of 1, 0.75, 0.5 and 0.2. We found that most of the commonly used meta-analytical methods were biased when data were sparse. The bias was greatest in inverse variance and DerSimonian and Laird odds ratio and risk difference methods, and the Mantel-Haenszel (MH) odds ratio method using a 0.5 zero-cell correction. Risk difference meta-analytical methods tended to show conservative confidence interval coverage and low statistical power at low event rates. At event rates below 1 per cent the Peto one-step odds ratio method was the least biased and most powerful method, and provided the best confidence interval coverage, provided there was no substantial imbalance between treatment and control group sizes within trials, and treatment effects were not exceptionally large. In other circumstances the MH OR without zero-cell corrections, logistic regression and the exact method performed similarly to each other, and were less biased than the Peto method.     

Alternative methods for testing treatment effects on the basis of multiple outcomes: simulation and case study

In clinical trials multiple outcomes are often used to assess treatment interventions. This paper presents an evaluation of likelihood-based methods for jointly testing treatment effects in clinical trials with multiple continuous outcomes. Specifically, we compare the power of joint tests of treatment effects obtained from joint models for the multiple outcomes with univariate tests based on modeling the outcomes separately. We also consider the power and bias of tests when data are missing, a common feature of many trials, especially in psychiatry. Our results suggest that joint tests capitalize on the correlation of multiple outcomes and are more powerful than standard univariate methods, especially when outcomes are missing completely at random. When outcomes are missing at random, test procedures based on correctly specified joint models are unbiased, while standard univariate procedures are not. Results of a simulation study are reported, and the methods are illustrated in an example from the Clinical Antipsychotic Trials of Intervention Effectiveness for schizophrenia.     

One-stage individual participant data meta-analysis models for continuous and binary outcomes: Comparison of treatment coding options and estimation methods

A one-stage individual participant data (IPD) meta-analysis synthesizes IPD from multiple studies using a general or generalized linear mixed model. This produces summary results (eg, about treatment effect) in a single step, whilst accounting for clustering of participants within studies (via a stratified study intercept, or random study intercepts) and between-study heterogeneity (via random treatment effects). We use simulation to evaluate the performance of restricted maximum likelihood (REML) and maximum likelihood (ML) estimation of one-stage IPD meta-analysis models for synthesizing randomized trials with continuous or binary outcomes. Three key findings are identified. First, for ML or REML estimation of stratified intercept or random intercepts models, a t-distribution based approach generally improves coverage of confidence intervals for the summary treatment effect, compared with a z-based approach. Second, when using ML estimation of a one-stage model with a stratified intercept, the treatment variable should be coded using “study-specific centering” (ie, 1/0 minus the study-specific proportion of participants in the treatment group), as this reduces the bias in the between-study variance estimate (compared with 1/0 and other coding options). Third, REML estimation reduces downward bias in between-study variance estimates compared with ML estimation, and does not depend on the treatment variable coding; for binary outcomes, this requires REML estimation of the pseudo-likelihood, although this may not be stable in some situations (eg, when data are sparse). Two applied examples are used to illustrate the findings.     

A graphical method for exploring heterogeneity in meta-analyses: application to a meta-analysis of 65 trials

Heterogeneity can be a major component of meta-analyses and by virtue of that fact warrants investigation. Classic analysis methods, such as meta-regression, are used to explore the sources of heterogeneity. However, it may be difficult to apply such a method in complex cases or in the absence of an a priori hypothesis. This paper presents a graphical method to identify trials, groups of trials or groups of patients that are sources of heterogeneity. The contribution of these trials to the overall result can also be evaluated with this method. Each trial is represented by a dot on a 2D graph. The X-axis represents the contribution of the trial to the overall Cochran Q-test for heterogeneity. The Y-axis represents the influence of the trial, defined as the standardized squared difference between the treatment effects estimated with and without the trial. This approach has been applied to data from the Meta-Analysis of Chemotherapy in Head and Neck Cancer (MACH-NC) comprising 10,850 patients in 65 randomized trials. The graphical method allowed us to identify trials that contributed considerably to the overall heterogeneity and had a strong influence on the overall result. It also provided useful information for the interpretation of heterogeneity in this meta-analysis. The proposed graphical method identifies trials that account for most of the heterogeneity without having to explore all possible sources of heterogeneity by subgroup analyses. This method can also be applied to identify types of patients that explain heterogeneity in the treatment effect.     

A principal stratification approach for evaluating natural direct and indirect effects in the presence of treatment‐induced intermediate confounding

Recently, several authors have shown that natural direct and indirect effects (NDEs and NIEs) can be identified under the sequential ignorability assumptions, as long as there is no mediator–outcome confounder that is affected by the treatment. However, if such a confounder exists, NDEs and NIEs will generally not be identified without making additional identifying assumptions. In this article, we propose novel identification assumptions and estimators for evaluating NDEs and NIEs under the usual sequential ignorability assumptions, using the principal stratification framework. It is assumed that the treatment and the mediator are dichotomous. We must impose strong assumptions for identification. However, even if these assumptions were violated, the bias of our estimator would be small under typical conditions, which can be easily evaluated from the observed data. This conjecture is confirmed for binary outcomes by deriving the bounds of the bias terms. In addition, the advantage of our estimator is illustrated through a simulation study. We also propose a method of sensitivity analysis that examines what happens when our assumptions are violated. We apply the proposed method to data from the National Center for Health Statistics. Copyright © 2014 John Wiley & Sons, Ltd.     

内蒙古扬沙天气与人群健康效应关系的初步研究

目的初步研究扬沙天气对人群可能造成的急性健康危害。方法根据整群抽样的原则选取内蒙古自治区包头市昆区2所小学校的3~5年级的1362名儿童及2618名成人为研究对象,采用随访研究的方法,在2004年3月~4月扬沙天气发生期间进行健康状况问卷调查,同期的SO2、NO2、CO、PM10的浓度由包头市环保局提供。问卷项目主要包括:人群一般情况;环境暴露情况:咳嗽咳痰等呼吸系统自觉症状及急性呼吸道刺激症状,医院就诊情况,心情是否焦虑或压抑等。用SPSS11·5统计软件进行单因素统计分析和多元回归分析。结果大气PM10的浓度在扬沙发生当天明显升高,扬沙过后随之下降;人群以呼吸系统为主的自觉症状发生率在扬沙发生当天明显升高,扬沙过后开始下降,这些症状包括咳嗽、咳痰、胸闷气短、咽干口苦、眼睛干涩、流鼻涕、流泪、打喷嚏等,还包括焦虑或压抑的感觉以及医院就诊率(均P<0·01);大气PM10的浓度和人群咳嗽症状以及急性刺激症状的发生率呈显著正相关(P<0·05)。结论扬沙天气造成大气PM10的浓度明显升高,对暴露人群可造成急性呼吸系统健康危害。其带来的空气污染加剧和人群呼吸系统自觉症状等发生率的升高主要为短期急性效应。     

Performance of variable and function selection methods for estimating the nonlinear health effects of correlated chemical mixtures: A simulation study

Statistical methods for identifying harmful chemicals in a correlated mixture often assume linearity in exposure-response relationships. Nonmonotonic relationships are increasingly recognized (eg, for endocrine-disrupting chemicals); however, the impact of nonmonotonicity on exposure selection has not been evaluated. In a simulation study, we assessed the performance of Bayesian kernel machine regression (BKMR), Bayesian additive regression trees (BART), Bayesian structured additive regression with spike-slab priors (BSTARSS), generalized additive models with double penalty (GAMDP) and thin plate shrinkage smoothers (GAMTS), multivariate adaptive regression splines (MARS), and lasso penalized regression. We simulated realistic exposure data based on pregnancy exposure to 17 phthalates and phenols in the US National Health and Nutrition Examination Survey using a multivariate copula. We simulated data sets of size N = 250 and compared methods across 32 scenarios, varying by model size and sparsity, signal-to-noise ratio, correlation structure, and exposure-response relationship shapes. We compared methods in terms of their sensitivity, specificity, and estimation accuracy. In most scenarios, BKMR, BSTARSS, GAMDP, and GAMTS achieved moderate to high sensitivity (0.52-0.98) and specificity (0.21-0.99). BART and MARS achieved high specificity (≥0.90), but low sensitivity in low signal-to-noise ratio scenarios (0.20-0.51). Lasso was highly sensitive (0.71-0.99), except for quadratic relationships (≤0.27). Penalized regression methods that assume linearity, such as lasso, may not be suitable for studies of environmental chemicals hypothesized to have nonmonotonic relationships with outcomes. Instead, BKMR, BSTARSS, GAMDP, and GAMTS are attractive methods for flexibly estimating the shapes of exposure-response relationships and selecting among correlated exposures.     

Effects of stage-matched repeated individual counseling on smoking cessation: A randomized controlled trial for the high-risk strategy by lifestyle modification (HISLIM) study

Objective  The purpose of this study was to evaluate the effects of stage-matched repeated individual behavioral counseling as an intervention for the cessation of smoking. Methods  We conducted a multisite randomized controlled trial that enrolled smokers unselected for their readiness to quit. There were 979 smokers with hypertension or hypercholesterolemia recruited from 72 study sites and randomly allocated to the intervention or control group. Smokers in the intervention group received stage-matched individual counseling consisting of a 40 minute initial session and four 20–30 minute follow-up sessions. Smokers in the control group received individual behavioral counseling for hypertension or hypercholesterolemia. Results  The point prevalence abstinence rate at 6 months, validated by carbon monoxide testing, in the intervention group (13.6%) was 5.4 times higher (p<0.001) than that in the control group (2.5%). When the data were analyzed based on the baseline stage of change, there were significant differences in the abstinence rates at 6 months in smokers versus controls with each stage of change except in immotives. The odds ratio was 6.4 (p<0.001) in precontemplators, 6.7 (p<0.001) in contemplators, and 6.2 (p<0.01) in preparators. There was a positive, consistent effect of the intervention regardless of study site (worksite or community) or the presence of hypertension or hypercholesterolemia. Conclusions  We showed the effects of an intervention with repeated individual behavioral counseling on the cessation of smoking in smokers unselected for their readiness to quit. This result suggests that stage-matched individual counseling, based on the transtheoretical model, is effective in smokers with a lower motivation to quit as well as those ready to quit. Investigators of the research group are listed in the final report of the Research on Long-Term Chronic Disease “Seikatsu Syukanbyo han” 1998, granted from the Ministry of Health and Welfare, Japan