Statistical approaches to identify subgroups in meta-analysis of individual participant data: a simulation study.
Ontology highlight
ABSTRACT:
SUBMITTER: Belias M
PROVIDER: S-EPMC6720416 | biostudies-literature | 2019 Sep
REPOSITORIES: biostudies-literature
Publications
Statistical approaches to identify subgroups in meta-analysis of individual participant data: a simulation study.
Belias Michail M Rovers Maroeska M MM Reitsma Johannes B JB Debray Thomas P A TPA IntHout Joanna J
BMC medical research methodology 20190902 1
<h4>Background</h4>Individual participant data meta-analysis (IPD-MA) is considered the gold standard for investigating subgroup effects. Frequently used regression-based approaches to detect subgroups in IPD-MA are: meta-regression, per-subgroup meta-analysis (PS-MA), meta-analysis of interaction terms (MA-IT), naive one-stage IPD-MA (ignoring potential study-level confounding), and centred one-stage IPD-MA (accounting for potential study-level confounding). Clear guidance on the analyses is la ...[more]
Similar Datasets
Project description:BACKGROUND:Researchers and funders should consider the statistical power of planned Individual Participant Data (IPD) meta-analysis projects, as they are often time-consuming and costly. We propose simulation-based power calculations utilising a two-stage framework, and illustrate the approach for a planned IPD meta-analysis of randomised trials with continuous outcomes where the aim is to identify treatment-covariate interactions. METHODS:The simulation approach has four steps: (i) specify an underlying (data generating) statistical model for trials in the IPD meta-analysis; (ii) use readily available information (e.g. from publications) and prior knowledge (e.g. number of studies promising IPD) to specify model parameter values (e.g. control group mean, intervention effect, treatment-covariate interaction); (iii) simulate an IPD meta-analysis dataset of a particular size from the model, and apply a two-stage IPD meta-analysis to obtain the summary estimate of interest (e.g. interaction effect) and its associated p-value; (iv) repeat the previous step (e.g. thousands of times), then estimate the power to detect a genuine effect by the proportion of summary estimates with a significant p-value. RESULTS:In a planned IPD meta-analysis of lifestyle interventions to reduce weight gain in pregnancy, 14 trials (1183 patients) promised their IPD to examine a treatment-BMI interaction (i.e. whether baseline BMI modifies intervention effect on weight gain). Using our simulation-based approach, a two-stage IPD meta-analysis has <?60% power to detect a reduction of 1 kg weight gain for a 10-unit increase in BMI. Additional IPD from ten other published trials (containing 1761 patients) would improve power to over 80%, but only if a fixed-effect meta-analysis was appropriate. Pre-specified adjustment for prognostic factors would increase power further. Incorrect dichotomisation of BMI would reduce power by over 20%, similar to immediately throwing away IPD from ten trials. CONCLUSIONS:Simulation-based power calculations could inform the planning and funding of IPD projects, and should be used routinely.
| S-EPMC5960205 | biostudies-literature
Project description:When combining results across related studies, a multivariate meta-analysis allows the joint synthesis of correlated effect estimates from multiple outcomes. Joint synthesis can improve efficiency over separate univariate syntheses, may reduce selective outcome reporting biases, and enables joint inferences across the outcomes. A common issue is that within-study correlations needed to fit the multivariate model are unknown from published reports. However, provision of individual participant data (IPD) allows them to be calculated directly. Here, we illustrate how to use IPD to estimate within-study correlations, using a joint linear regression for multiple continuous outcomes and bootstrapping methods for binary, survival and mixed outcomes. In a meta-analysis of 10 hypertension trials, we then show how these methods enable multivariate meta-analysis to address novel clinical questions about continuous, survival and binary outcomes; treatment-covariate interactions; adjusted risk/prognostic factor effects; longitudinal data; prognostic and multiparameter models; and multiple treatment comparisons. Both frequentist and Bayesian approaches are applied, with example software code provided to derive within-study correlations and to fit the models.
| S-EPMC4847645 | biostudies-literature
Project description:BackgroundIndividual participant data (IPD) meta-analyses that obtain "raw" data from studies rather than summary data typically adopt a "two-stage" approach to analysis whereby IPD within trials generate summary measures, which are combined using standard meta-analytical methods. Recently, a range of "one-stage" approaches which combine all individual participant data in a single meta-analysis have been suggested as providing a more powerful and flexible approach. However, they are more complex to implement and require statistical support. This study uses a dataset to compare "two-stage" and "one-stage" models of varying complexity, to ascertain whether results obtained from the approaches differ in a clinically meaningful way.Methods and findingsWe included data from 24 randomised controlled trials, evaluating antiplatelet agents, for the prevention of pre-eclampsia in pregnancy. We performed two-stage and one-stage IPD meta-analyses to estimate overall treatment effect and to explore potential treatment interactions whereby particular types of women and their babies might benefit differentially from receiving antiplatelets. Two-stage and one-stage approaches gave similar results, showing a benefit of using anti-platelets (Relative risk 0.90, 95% CI 0.84 to 0.97). Neither approach suggested that any particular type of women benefited more or less from antiplatelets. There were no material differences in results between different types of one-stage model.ConclusionsFor these data, two-stage and one-stage approaches to analysis produce similar results. Although one-stage models offer a flexible environment for exploring model structure and are useful where across study patterns relating to types of participant, intervention and outcome mask similar relationships within trials, the additional insights provided by their usage may not outweigh the costs of statistical support for routine application in syntheses of randomised controlled trials. Researchers considering undertaking an IPD meta-analysis should not necessarily be deterred by a perceived need for sophisticated statistical methods when combining information from large randomised trials.
| S-EPMC3463584 | biostudies-literature
Project description:BackgroundIt remains unclear when standard systematic reviews and meta-analyses that rely on published aggregate data (AD) can provide robust clinical conclusions. We aimed to compare the results from a large cohort of systematic reviews and meta-analyses based on individual participant data (IPD) with meta-analyses of published AD, to establish when the latter are most likely to be reliable and when the IPD approach might be required.Methods and findingsWe used 18 cancer systematic reviews that included IPD meta-analyses: all of those completed and published by the Meta-analysis Group of the MRC Clinical Trials Unit from 1991 to 2010. We extracted or estimated hazard ratios (HRs) and standard errors (SEs) for survival from trial reports and compared these with IPD equivalents at both the trial and meta-analysis level. We also extracted or estimated the number of events. We used paired t tests to assess whether HRs and SEs from published AD differed on average from those from IPD. We assessed agreement, and whether this was associated with trial or meta-analysis characteristics, using the approach of Bland and Altman. The 18 systematic reviews comprised 238 unique trials or trial comparisons, including 37,082 participants. A HR and SE could be generated for 127 trials, representing 53% of the trials and approximately 79% of eligible participants. On average, trial HRs derived from published AD were slightly more in favour of the research interventions than those from IPD (HRAD to HRIPD ratio = 0.95, p = 0.007), but the limits of agreement show that for individual trials, the HRs could deviate substantially. These limits narrowed with an increasing number of participants (p < 0.001) or a greater number (p < 0.001) or proportion (p < 0.001) of events in the AD. On average, meta-analysis HRs from published AD slightly tended to favour the research interventions whether based on fixed-effect (HRAD to HRIPD ratio = 0.97, p = 0.088) or random-effects (HRAD to HRIPD ratio = 0.96, p = 0.044) models, but the limits of agreement show that for individual meta-analyses, agreement was much more variable. These limits tended to narrow with an increasing number (p = 0.077) or proportion of events (p = 0.11) in the AD. However, even when the information size of the AD was large, individual meta-analysis HRs could still differ from their IPD equivalents by a relative 10% in favour of the research intervention to 5% in favour of control. We utilised the results to construct a decision tree for assessing whether an AD meta-analysis includes sufficient information, and when estimates of effects are most likely to be reliable. A lack of power at the meta-analysis level may have prevented us identifying additional factors associated with the reliability of AD meta-analyses, and we cannot be sure that our results are generalisable to all outcomes and effect measures.ConclusionsIn this study we found that HRs from published AD were most likely to agree with those from IPD when the information size was large. Based on these findings, we provide guidance for determining systematically when standard AD meta-analysis will likely generate robust clinical conclusions, and when the IPD approach will add considerable value.
| S-EPMC6993967 | biostudies-literature
Project description:Meta-analytic methods for combining data from multiple intervention trials are commonly used to estimate the effectiveness of an intervention. They can also be extended to study comparative effectiveness, testing which of several alternative interventions is expected to have the strongest effect. This often requires network meta-analysis (NMA), which combines trials involving direct comparison of two interventions within the same trial and indirect comparisons across trials. In this paper, we extend existing network methods for main effects to examining moderator effects, allowing for tests of whether intervention effects vary for different populations or when employed in different contexts. In addition, we study how the use of individual participant data may increase the sensitivity of NMA for detecting moderator effects, as compared with aggregate data NMA that employs study-level effect sizes in a meta-regression framework. A new NMA diagram is proposed. We also develop a generalized multilevel model for NMA that takes into account within-trial and between-trial heterogeneity and can include participant-level covariates. Within this framework, we present definitions of homogeneity and consistency across trials. A simulation study based on this model is used to assess effects on power to detect both main and moderator effects. Results show that power to detect moderation is substantially greater when applied to individual participant data as compared with study-level effects. We illustrate the use of this method by applying it to data from a classroom-based randomized study that involved two sub-trials, each comparing interventions that were contrasted with separate control groups. Copyright © 2016 John Wiley & Sons, Ltd.
| S-EPMC4893003 | biostudies-literature
Project description:Multiple imputation is a strategy for the analysis of incomplete data such that the impact of the missingness on the power and bias of estimates is mitigated. When data from multiple studies are collated, we can propose both within-study and multilevel imputation models to impute missing data on covariates. It is not clear how to choose between imputation models or how to combine imputation and inverse-variance weighted meta-analysis methods. This is especially important as often different studies measure data on different variables, meaning that we may need to impute data on a variable which is systematically missing in a particular study. In this paper, we consider a simulation analysis of sporadically missing data in a single covariate with a linear analysis model and discuss how the results would be applicable to the case of systematically missing data. We find in this context that ensuring the congeniality of the imputation and analysis models is important to give correct standard errors and confidence intervals. For example, if the analysis model allows between-study heterogeneity of a parameter, then we should incorporate this heterogeneity into the imputation model to maintain the congeniality of the two models. In an inverse-variance weighted meta-analysis, we should impute missing data and apply Rubin's rules at the study level prior to meta-analysis, rather than meta-analyzing each of the multiple imputations and then combining the meta-analysis estimates using Rubin's rules. We illustrate the results using data from the Emerging Risk Factors Collaboration.
| S-EPMC3963448 | biostudies-literature
Project description:BackgroundFour studies previously indicated that the effect of malaria infection during pregnancy on the risk of low birthweight (LBW; <2,500 g) may depend upon maternal nutritional status. We investigated this dependence further using a large, diverse study population.Methods and findingsWe evaluated the interaction between maternal malaria infection and maternal anthropometric status on the risk of LBW using pooled data from 14,633 pregnancies from 13 studies (6 cohort studies and 7 randomized controlled trials) conducted in Africa and the Western Pacific from 1996-2015. Studies were identified by the Maternal Malaria and Malnutrition (M3) initiative using a convenience sampling approach and were eligible for pooling given adequate ethical approval and availability of essential variables. Study-specific adjusted effect estimates were calculated using inverse probability of treatment-weighted linear and log-binomial regression models and pooled using a random-effects model. The adjusted risk of delivering a baby with LBW was 8.8% among women with malaria infection at antenatal enrollment compared to 7.7% among uninfected women (adjusted risk ratio [aRR] 1.14 [95% confidence interval (CI): 0.91, 1.42]; N = 13,613), 10.5% among women with malaria infection at delivery compared to 7.9% among uninfected women (aRR 1.32 [95% CI: 1.08, 1.62]; N = 11,826), and 15.3% among women with low mid-upper arm circumference (MUAC <23 cm) at enrollment compared to 9.5% among women with MUAC ? 23 cm (aRR 1.60 [95% CI: 1.36, 1.87]; N = 9,008). The risk of delivering a baby with LBW was 17.8% among women with both malaria infection and low MUAC at enrollment compared to 8.4% among uninfected women with MUAC ? 23 cm (joint aRR 2.13 [95% CI: 1.21, 3.73]; N = 8,152). There was no evidence of synergism (i.e., excess risk due to interaction) between malaria infection and MUAC on the multiplicative (p = 0.5) or additive scale (p = 0.9). Results were similar using body mass index (BMI) as an anthropometric indicator of nutritional status. Meta-regression results indicated that there may be multiplicative interaction between malaria infection at enrollment and low MUAC within studies conducted in Africa; however, this finding was not consistent on the additive scale, when accounting for multiple comparisons, or when using other definitions of malaria and malnutrition. The major limitations of the study included availability of only 2 cross-sectional measurements of malaria and the limited availability of ultrasound-based pregnancy dating to assess impacts on preterm birth and fetal growth in all studies.ConclusionsPregnant women with malnutrition and malaria infection are at increased risk of LBW compared to women with only 1 risk factor or none, but malaria and malnutrition do not act synergistically.
| S-EPMC5549702 | biostudies-literature
Project description:One-stage individual participant data meta-analysis models should account for within-trial clustering, but it is currently debated how to do this. For continuous outcomes modeled using a linear regression framework, two competing approaches are a stratified intercept or a random intercept. The stratified approach involves estimating a separate intercept term for each trial, whereas the random intercept approach assumes that trial intercepts are drawn from a normal distribution. Here, through an extensive simulation study for continuous outcomes, we evaluate the impact of using the stratified and random intercept approaches on statistical properties of the summary treatment effect estimate. Further aims are to compare (i) competing estimation options for the one-stage models, including maximum likelihood and restricted maximum likelihood, and (ii) competing options for deriving confidence intervals (CI) for the summary treatment effect, including the standard normal-based 95% CI, and more conservative approaches of Kenward-Roger and Satterthwaite, which inflate CIs to account for uncertainty in variance estimates. The findings reveal that, for an individual participant data meta-analysis of randomized trials with a 1:1 treatment:control allocation ratio and heterogeneity in the treatment effect, (i) bias and coverage of the summary treatment effect estimate are very similar when using stratified or random intercept models with restricted maximum likelihood, and thus either approach could be taken in practice, (ii) CIs are generally best derived using either a Kenward-Roger or Satterthwaite correction, although occasionally overly conservative, and (iii) if maximum likelihood is required, a random intercept performs better than a stratified intercept model. An illustrative example is provided.
| S-EPMC6283045 | biostudies-literature
Project description:BackgroundNetwork meta-analyses using individual participant data (IPD-NMAs) have been increasingly used to compare the effects of multiple interventions. Although there have been many studies on statistical methods for IPD-NMAs, it is unclear whether there are statistical defects in published IPD-NMAs and whether the reporting of statistical analyses has improved. This study aimed to investigate statistical methods used and assess the reporting and methodological quality of IPD-NMAs.MethodsWe searched four bibliographic databases to identify published IPD-NMAs. The methodological quality was assessed using AMSTAR-2 and reporting quality assessed based on PRISMA-IPD and PRISMA-NMA. We performed stratified analyses and correlation analyses to explore the factors that might affect quality.ResultsWe identified 21 IPD-NMAs. Only 23.8% of the included IPD-NMAs reported statistical techniques used for missing participant data, 42.9% assessed the consistency, and none assessed the transitivity. None of the included IPD-NMAs reported sources of funding for trials included, only 9.5% stated pre-registration of protocols, and 28.6% assessed the risk of bias in individual studies. For reporting quality, compliance rates were lower than 50.0% for more than half of the items. Less than 15.0% of the IPD-NMAs reported data integrity, presented the network geometry, or clarified risk of bias across studies. IPD-NMAs with statistical or epidemiological authors often better assessed the inconsistency (P?=?0.017). IPD-NMAs with a priori protocol were associated with higher reporting quality in terms of search (P?=?0.046), data collection process (P?=?0.031), and syntheses of results (P?=?0.006).ConclusionsThe reporting of statistical methods and compliance rates of methodological and reporting items of IPD-NMAs were suboptimal. Authors of future IPD-NMAs should address the identified flaws and strictly adhere to methodological and reporting guidelines.
| S-EPMC7262764 | biostudies-literature
Project description:BackgroundSome studies have suggested a link between antihypertensive medication and cancer, but the evidence is so far inconclusive. Thus, we aimed to investigate this association in a large individual patient data meta-analysis of randomised clinical trials.MethodsWe searched PubMed, MEDLINE, The Cochrane Central Register of Controlled Trials, and ClinicalTrials.gov from Jan 1, 1966, to Sept 1, 2019, to identify potentially eligible randomised controlled trials. Eligible studies were randomised controlled trials comparing one blood pressure lowering drug class with a placebo, inactive control, or other blood pressure lowering drug. We also required that trials had at least 1000 participant years of follow-up in each treatment group. Trials without cancer event information were excluded. We requested individual participant data from the authors of eligible trials. We pooled individual participant-level data from eligible trials and assessed the effects of angiotensin-converting enzyme inhibitors (ACEIs), angiotensin II receptor blockers (ARBs), β blockers, calcium channel blockers, and thiazide diuretics on cancer risk in one-stage individual participant data and network meta-analyses. Cause-specific fixed-effects Cox regression models, stratified by trial, were used to calculate hazard ratios (HRs). The primary outcome was any cancer event, defined as the first occurrence of any cancer diagnosed after randomisation. This study is registered with PROSPERO (CRD42018099283).Findings33 trials met the inclusion criteria, and included 260 447 participants with 15 012 cancer events. Median follow-up of included participants was 4·2 years (IQR 3·0-5·0). In the individual participant data meta-analysis comparing each drug class with all other comparators, no associations were identified between any antihypertensive drug class and risk of any cancer (HR 0·99 [95% CI 0·95-1·04] for ACEIs; 0·96 [0·92-1·01] for ARBs; 0·98 [0·89-1·07] for β blockers; 1·01 [0·95-1·07] for thiazides), with the exception of calcium channel blockers (1·06 [1·01-1·11]). In the network meta-analysis comparing drug classes against placebo, we found no excess cancer risk with any drug class (HR 1·00 [95% CI 0·93-1·09] for ACEIs; 0·99 [0·92-1·06] for ARBs; 0·99 [0·89-1·11] for β blockers; 1·04 [0·96-1·13] for calcium channel blockers; 1·00 [0·90-1·10] for thiazides).InterpretationWe found no consistent evidence that antihypertensive medication use had any effect on cancer risk. Although such findings are reassuring, evidence for some comparisons was insufficient to entirely rule out excess risk, in particular for calcium channel blockers.FundingBritish Heart Foundation, National Institute for Health Research, Oxford Martin School.
| S-EPMC8024901 | biostudies-literature