Search

Background: Joint modeling of longitudinal and time-to-event data is often advantageous over separate longitudinal or time-to-event analyses as it can account for study dropout, error in longitudinally measured covariates, and correlation between longitudinal and time-to-event outcomes. The current literature on joint modeling focuses mainly on the analysis of single studies with a lack of methods available for the meta-analysis of joint data from multiple studies., Methods: We investigate a variety of one-stage methods for the meta-analysis of joint longitudinal and time-to-event outcome data. These methods are applied to the INDANA dataset to investigate longitudinally measured systolic blood pressure, with each of time to death, time to myocardial infarction, and time to stroke. Results are compared to separate longitudinal or time-to-event meta-analyses. A simulation study is conducted to contrast separate versus joint analyses over a range of scenarios., Results: The performance of the examined one-stage joint meta-analytic models varied. Models that accounted for between study heterogeneity performed better than models that ignored it. Of the examined methods to account for between study heterogeneity, under the examined association structure, fixed effect approaches appeared preferable, whereas methods involving a baseline hazard stratified by study were least time intensive., Conclusions: One-stage joint meta-analytic models that accounted for between study heterogeneity using a mix of fixed effects or a stratified baseline hazard were reliable; however, models examined that included study level random effects in the association structure were less reliable., (© 2018 The Authors. Statistics in Medicine Published by John Wiley & Sons Ltd.)

Background: Joint modelling of longitudinal and time-to-event data is often preferred over separate longitudinal or time-to-event analyses as it can account for study dropout, error in longitudinally measured covariates, and correlation between longitudinal and time-to-event outcomes. The joint modelling literature focuses mainly on the analysis of single studies with no methods currently available for the meta-analysis of joint model estimates from multiple studies., Methods: We propose a 2-stage method for meta-analysis of joint model estimates. These methods are applied to the INDANA dataset to combine joint model estimates of systolic blood pressure with time to death, time to myocardial infarction, and time to stroke. Results are compared to meta-analyses of separate longitudinal or time-to-event models. A simulation study is conducted to contrast separate versus joint analyses over a range of scenarios., Results: Using the real dataset, similar results were obtained by using the separate and joint analyses. However, the simulation study indicated a benefit of use of joint rather than separate methods in a meta-analytic setting where association exists between the longitudinal and time-to-event outcomes., Conclusions: Where evidence of association between longitudinal and time-to-event outcomes exists, results from joint models over standalone analyses should be pooled in 2-stage meta-analyses., (© 2017 The Authors. Statistics in Medicine Published by John Wiley & Sons Ltd.)

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., (© 2014 The Authors. Research Synthesis Methods published by John Wiley & Sons, Ltd.)

Objective: To establish the extent to which systematic reviews and meta-analyses of individual participant data (IPD) are being used to inform the recommendations included in published clinical guidelines., Design: Descriptive study., Setting: Database maintained by the Cochrane IPD Meta-analysis Methods Group, supplemented by records of published IPD meta-analyses held in a separate database., Population: A test sample of systematic reviews of randomised controlled trials that included a meta-analysis of IPD, and a separate sample of clinical guidelines, matched to the IPD meta-analyses according to medical condition, interventions, populations, and dates of publication., Data Extraction: Descriptive information on each guideline was extracted along with evidence showing use or critical appraisal, or both, of the IPD meta-analysis within the guideline; recommendations based directly on its findings and the use of other systematic reviews in the guideline., Results: Based on 33 IPD meta-analyses and 177 eligible, matched clinical guidelines there was evidence that IPD meta-analyses were being under-utilised. Only 66 guidelines (37%) cited a matched IPD meta-analysis. Around a third of these (n=22, 34%) had critically appraised the IPD meta-analysis. Recommendations based directly on the matched IPD meta-analyses were identified for only 18 of the 66 guidelines (27%). For the guidelines that did not cite a matched IPD meta-analysis (n=111, 63%), search dates had preceded the publication of the IPD meta-analysis in 23 cases (21%); however, for the remainder, there was no obvious reasons why the IPD meta-analysis had not been cited., Conclusions: Our results indicate that systematic reviews and meta-analyses based on IPD are being under-utilised. Guideline developers should routinely seek good quality and up to date IPD meta-analyses to inform guidelines. Increased use of IPD meta-analyses could lead to improved guidelines ensuring that routine patient care is based on the most reliable evidence available., (© Vale et al 2015.)

When some trials provide individual patient data (IPD) and the others provide only aggregate data (AD), meta-analysis methods for combining IPD and AD are required. We propose a method that reconstructs the missing IPD for AD trials by a Bayesian sampling procedure and then applies an IPD meta-analysis model to the mixture of simulated IPD and collected IPD. The method is applicable when a treatment effect can be assumed fixed across trials. We focus on situations of a single continuous outcome and covariate and aim to estimate treatment-covariate interactions separated into within-trial and across-trial effect. An illustration with hypertension data which has similar mean covariates across trials indicates that the method substantially reduces mean square error of the pooled within-trial interaction estimate in comparison with existing approaches. A simulation study supposing there exists one IPD trial and nine AD trials suggests that the method has suitable type I error rate and approximately zero bias as long as the available IPD contains at least 10% of total patients, where the average gain in mean square error is up to about 40%. However, the method is currently restricted by the fixed effect assumption, and extension to random effects to allow heterogeneity is required., (Copyright © 2014 John Wiley & Sons, Ltd.)

We describe methods for meta-analysis of randomised trials where a continuous outcome is of interest, such as blood pressure, recorded at both baseline (pre treatment) and follow-up (post treatment). We used four examples for illustration, covering situations with and without individual participant data (IPD) and with and without baseline imbalance between treatment groups in each trial. Given IPD, meta-analysts can choose to synthesise treatment effect estimates derived using analysis of covariance (ANCOVA), a regression of just final scores, or a regression of the change scores. When there is baseline balance in each trial, treatment effect estimates derived using ANCOVA are more precise and thus preferred. However, we show that meta-analysis results for the summary treatment effect are similar regardless of the approach taken. Thus, without IPD, if trials are balanced, reviewers can happily utilise treatment effect estimates derived from any of the approaches. However, when some trials have baseline imbalance, meta-analysts should use treatment effect estimates derived from ANCOVA, as this adjusts for imbalance and accounts for the correlation between baseline and follow-up; we show that the other approaches can give substantially different meta-analysis results. Without IPD and with unavailable ANCOVA estimates, reviewers should limit meta-analyses to those trials with baseline balance. Trowman's method to adjust for baseline imbalance without IPD performs poorly in our examples and so is not recommended. Finally, we extend the ANCOVA model to estimate the interaction between treatment effect and baseline values and compare options for estimating this interaction given only aggregate data., (Copyright © 2013 John Wiley & Sons, Ltd.)

Background: An Individual Patient Data (IPD) meta-analysis is often considered the gold-standard for synthesising survival data from clinical trials. An IPD meta-analysis can be achieved by either a two-stage or a one-stage approach, depending on whether the trials are analysed separately or simultaneously. A range of one-stage hierarchical Cox models have been previously proposed, but these are known to be computationally intensive and are not currently available in all standard statistical software. We describe an alternative approach using Poisson based Generalised Linear Models (GLMs)., Methods: We illustrate, through application and simulation, the Poisson approach both classically and in a Bayesian framework, in two-stage and one-stage approaches. We outline the benefits of our one-stage approach through extension to modelling treatment-covariate interactions and non-proportional hazards. Ten trials of hypertension treatment, with all-cause death the outcome of interest, are used to apply and assess the approach., Results: We show that the Poisson approach obtains almost identical estimates to the Cox model, is additionally computationally efficient and directly estimates the baseline hazard. Some downward bias is observed in classical estimates of the heterogeneity in the treatment effect, with improved performance from the Bayesian approach., Conclusion: Our approach provides a highly flexible and computationally efficient framework, available in all standard statistical software, to the investigation of not only heterogeneity, but the presence of non-proportional hazards and treatment effect modifiers.

Meta-analysis of individual patient data (IPD) is the gold-standard for synthesizing evidence across clinical studies. However, for some studies IPD may not be available and only aggregate data (AD), such as a treatment effect estimate and its standard error, may be obtained. In this situation, methods for combining IPD and AD are important to utilize all the available evidence. In this paper, we develop and assess a range of statistical methods for combining IPD and AD in meta-analysis of continuous outcomes from randomized controlled trials. The methods take either a one-step or a two-step approach. The latter is simple, with IPD reduced to AD so that standard AD meta-analysis techniques can be employed. The one-step approach is more complex but offers a flexible framework to include both patient-level and trial-level parameters. It uses a dummy variable to distinguish IPD trials from AD trials and to constrain which parameters the AD trials estimate. We show that this is important when assessing how patient-level covariates modify treatment effect, as aggregate-level relationships across trials are subject to ecological bias and confounding. We thus develop models to separate within-trial and across-trials treatment-covariate interactions; this ensures that only IPD trials estimate the former, whilst both IPD and AD trials estimate the latter in addition to the pooled treatment effect and any between-study heterogeneity. Extension to multiple correlated outcomes is also considered. Ten IPD trials in hypertension, with blood pressure the continuous outcome of interest, are used to assess the models and identify the benefits of utilizing AD alongside IPD., (Copyright (c) 2007 John Wiley & Sons, Ltd.)

Exploration of the variation of treatment effect over time in randomized clinical trials with low event rates is limited by lack of power. A meta-analysis on individual patient data from such trials can partly solve the problem, but brings other computational difficulties. Using an example in hypertension, we describe appropriate methods for graphical description and statistical modelling of treatment-time interactions in large data sets. Also, a method is developed for determining the total number of events required to detect treatment-period interactions of plausible magnitude. We conclude that trialists tend to overinterpret the observed data when looking for potential treatment-time interactions by visual comparisons of survival curves, failing to realize the substantial amounts of data that are needed for their detection and estimation.

Reliable syntheses are more and more needed by physicians. In the therapeutic field, this need is illustrated by the growing number of syntheses using meta-analysis tools. Some of these syntheses are based on individual patient data, and this offers three kinds of advantages: the favourable consequences of a collaborative approach, the enhancement of data reliability, and a more complete extraction of useful information thanks to wider analysis potential. Most often, achieving the first two advantages does not require the individual patient data to be available, and a wider analysis potential seems to be the major rationale of such an approach. For instance, the assessment of the variation of treatment effect with along time is more accurate, but may also raise doubts about the validity of some hypotheses, for example, those underlying classical analysis techniques, or at the basis of even acknowledged pathophysiological theory. Ultimately, the concepts of identifying responders (here defined in terms of quantity or quality of life) and identifying the therapy target population are concretely applied in a way that will affect classical medical practice. Tomorrow, drugs will not be prescribed to a hypertensive individual because of the blood pressure level, but considering the size of the predicted therapeutic benefit.

The efficacy of a drug is a quantitative concept rather than a qualitative one. This quantity is expressed by several efficacy indices. None of them meet all the requirements. However, that of absolute benefit is especially suitable for the patients because it tells them the exact gain they can expect from taking the treatment. The absolute benefit varies according to patients' profiles because it interacts with some components of these profiles. In theory, such interactions can be used to predict the size of the absolute benefit for each patient, as well to describe better than with the current tools the therapy target population. We explain why meta-analysis and effect models are means of improving the prediction of the size of the effect and the definition of the therapy target population.

The overall effect of antihypertensive drug treatment has been well documented. The proportion of patients who benefit varies according to their baseline cardiovascular risk, and is small for the majority of people treated. Some investigators propose limiting the treatment target population to patients at high cardiovascular risk, but several assumptions must be made to justify this procedure. The INDANA project is a meta-analysis based on individual patient data, and thus offers the opportunity to check the validity of these assumptions. Its main objective is to identify responders (and non-responders) in the drug treatment of hypertension. The rationale and methods for such an approach are presented here, with the solution for some technical problems. The conclusion of the data collection has shown that the project is feasible. The results of the main analysis should be available in 1996, and should contribute to the selection of responders and to the individualization of the treatment of hypertension.