Modeling the multi-state natural history of rare diseases with heterogeneous individual patient data: A simulation study.

Multi-state survival models are used to represent the natural history of a disease, forming the basis of a health technology assessment comparing a novel treatment to current practice. Constructing such models for rare diseases is problematic, since evidence sources are typically much sparser and more heterogeneous. This simulation study investigated different one-stage and two-stage approaches to meta-analyzing individual patient data (IPD) in a multi-state survival setting when the number and size of studies being meta-analyzed are small. The objective was to assess methods of different complexity to see when they are accurate, when they are inaccurate and when they struggle to converge due to the sparsity of data. Biologically plausible multi-state IPD were simulated from study- and transition-specific hazard functions. One-stage frailty and two-stage stratified models were estimated, and compared to a base case model that did not account for study heterogeneity. Convergence and the bias/coverage of population-level transition probabilities to, and lengths of stay in, each state were used to assess model performance. A real-world application to Duchenne Muscular Dystrophy, a neuromuscular rare disease, was conducted, and a software demonstration is provided. Models not accounting for study heterogeneity were consistently out-performed by two-stage models. Frailty models struggled to converge, particularly in scenarios of low heterogeneity, and predictions from models that did converge were also subject to bias. Stratified models may be better suited to meta-analyzing disparate sources of IPD in rare disease natural history/economic modeling, as they converge more consistently and produce less biased predictions of lengths of stay.

Hierarchical network meta-analysis models for synthesis of evidence from randomised and non-randomised studies.

BACKGROUND: With the increased interest in the inclusion of non-randomised data in network meta-analyses (NMAs) of randomised controlled trials (RCTs), analysts need to consider the implications of the differences in study designs as such data can be prone to increased bias due to the lack of randomisation and unmeasured confounding. This study aims to explore and extend a number of NMA models that account for the differences in the study designs, assessing their impact on the effect estimates and uncertainty. METHODS: Bayesian random-effects meta-analytic models, including naïve pooling and hierarchical models differentiating between the study designs, were extended to allow for the treatment class effect and accounting for bias, with further extensions allowing for bias terms to vary depending on the treatment class. Models were applied to an illustrative example in type 2 diabetes; using data from a systematic review of RCTs and non-randomised studies of two classes of glucose-lowering medications: sodium-glucose co-transporter 2 inhibitors and glucagon-like peptide-1 receptor agonists. RESULTS: Across all methods, the estimated mean differences in glycated haemoglobin after 24 and 52 weeks remained similar with the inclusion of observational data. The uncertainty around these estimates reduced when conducting naïve pooling, compared to NMA of RCT data alone, and remained similar when applying hierarchical model allowing for class effect. However, the uncertainty around these effect estimates increased when fitting hierarchical models allowing for the differences in study design. The impact on uncertainty varied between treatments when applying the bias adjustment models. Hierarchical models and bias adjustment models all provided a better fit in comparison to the naïve-pooling method. CONCLUSIONS: Hierarchical and bias adjustment NMA models accounting for study design may be more appropriate when conducting a NMA of RCTs and observational studies. The degree of uncertainty around the effectiveness estimates varied depending on the method but use of hierarchical models accounting for the study design resulted in increased uncertainty. Inclusion of non-randomised data may, however, result in inferences that are more generalisable and the models accounting for the differences in the study design allow for more detailed and appropriate modelling of complex data, preventing overly optimistic conclusions.

Risk factors associated with SARS-CoV-2 infection in a multiethnic cohort of United Kingdom healthcare workers (UK-REACH): A cross-sectional analysis.

BACKGROUND: Healthcare workers (HCWs), particularly those from ethnic minority groups, have been shown to be at disproportionately higher risk of infection with Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) compared to the general population. However, there is insufficient evidence on how demographic and occupational factors influence infection risk among ethnic minority HCWs. METHODS AND FINDINGS: We conducted a cross-sectional analysis using data from the baseline questionnaire of the United Kingdom Research study into Ethnicity and Coronavirus Disease 2019 (COVID-19) Outcomes in Healthcare workers (UK-REACH) cohort study, administered between December 2020 and March 2021. We used logistic regression to examine associations of demographic, household, and occupational risk factors with SARS-CoV-2 infection (defined by polymerase chain reaction (PCR), serology, or suspected COVID-19) in a diverse group of HCWs. The primary exposure of interest was self-reported ethnicity. Among 10,772 HCWs who worked during the first UK national lockdown in March 2020, the median age was 45 (interquartile range [IQR] 35 to 54), 75.1% were female and 29.6% were from ethnic minority groups. A total of 2,496 (23.2%) reported previous SARS-CoV-2 infection. The fully adjusted model contained the following dependent variables: demographic factors (age, sex, ethnicity, migration status, deprivation, religiosity), household factors (living with key workers, shared spaces in accommodation, number of people in household), health factors (presence/absence of diabetes or immunosuppression, smoking history, shielding status, SARS-CoV-2 vaccination status), the extent of social mixing outside of the household, and occupational factors (job role, the area in which a participant worked, use of public transport to work, exposure to confirmed suspected COVID-19 patients, personal protective equipment [PPE] access, aerosol generating procedure exposure, night shift pattern, and the UK region of workplace). After adjustment, demographic and household factors associated with increased odds of infection included younger age, living with other key workers, and higher religiosity. Important occupational risk factors associated with increased odds of infection included attending to a higher number of COVID-19 positive patients (aOR 2.59, 95% CI 2.11 to 3.18 for ≥21 patients per week versus none), working in a nursing or midwifery role (1.30, 1.11 to 1.53, compared to doctors), reporting a lack of access to PPE (1.29, 1.17 to 1.43), and working in an ambulance (2.00, 1.56 to 2.58) or hospital inpatient setting (1.55, 1.38 to 1.75). Those who worked in intensive care units were less likely to have been infected (0.76, 0.64 to 0.92) than those who did not. Black HCWs were more likely to have been infected than their White colleagues, an effect which attenuated after adjustment for other known risk factors. This study is limited by self-selection bias and the cross sectional nature of the study means we cannot infer the direction of causality. CONCLUSIONS: We identified key sociodemographic and occupational risk factors associated with SARS-CoV-2 infection among UK HCWs, and have determined factors that might contribute to a disproportionate odds of infection in HCWs from Black ethnic groups. These findings demonstrate the importance of social and occupational factors in driving ethnic disparities in COVID-19 outcomes, and should inform policies, including targeted vaccination strategies and risk assessments aimed at protecting HCWs in future waves of the COVID-19 pandemic. TRIAL REGISTRATION: The study was prospectively registered at ISRCTN (reference number: ISRCTN11811602).

Use of copula to model within-study association in bivariate meta-analysis of binomial data at the aggregate level: A Bayesian approach and application to surrogate endpoint evaluation.

Bivariate meta-analysis provides a useful framework for combining information across related studies and has been utilized to combine evidence from clinical studies to evaluate treatment efficacy on two outcomes. It has also been used to investigate surrogacy patterns between treatment effects on the surrogate endpoint and the final outcome. Surrogate endpoints play an important role in drug development when they can be used to measure treatment effect early compared to the final outcome and to predict clinical benefit or harm. The standard bivariate meta-analytic approach models the observed treatment effects on the surrogate and the final outcome outcomes jointly, at both the within-study and between-studies levels, using a bivariate normal distribution. For binomial data, a normal approximation on log odds ratio scale can be used. However, this method may lead to biased results when the proportions of events are close to one or zero, affecting the validation of surrogate endpoints. In this article, we explore modeling the two outcomes on the original binomial scale. First, we present a method that uses independent binomial likelihoods to model the within-study variability avoiding to approximate the observed treatment effects. However, the method ignores the within-study association. To overcome this issue, we propose a method using a bivariate copula with binomial marginals, which allows the model to account for the within-study association. We applied the methods to an illustrative example in chronic myeloid leukemia to investigate the surrogate relationship between complete cytogenetic response and event-free-survival.

Bayesian network meta-analysis methods for combining individual participant data and aggregate data from single arm trials and randomised controlled trials.

BACKGROUND: Increasingly in network meta-analysis (NMA), there is a need to incorporate non-randomised evidence to estimate relative treatment effects, and in particular in cases with limited randomised evidence, sometimes resulting in disconnected networks of treatments. When combining different sources of data, complex NMA methods are required to address issues associated with participant selection bias, incorporating single-arm trials (SATs), and synthesising a mixture of individual participant data (IPD) and aggregate data (AD). We develop NMA methods which synthesise data from SATs and randomised controlled trials (RCTs), using a mixture of IPD and AD, for a dichotomous outcome. METHODS: We propose methods under both contrast-based (CB) and arm-based (AB) parametrisations, and extend the methods to allow for both within- and across-trial adjustments for covariate effects. To illustrate the methods, we use an applied example investigating the effectiveness of biologic disease-modifying anti-rheumatic drugs for rheumatoid arthritis (RA). We applied the methods to a dataset obtained from a literature review consisting of 14 RCTs and an artificial dataset consisting of IPD from two SATs and AD from 12 RCTs, where the artificial dataset was created by removing the control arms from the only two trials assessing tocilizumab in the original dataset. RESULTS: Without adjustment for covariates, the CB method with independent baseline response parameters (CBunadjInd) underestimated the effectiveness of tocilizumab when applied to the artificial dataset compared to the original dataset, albeit with significant overlap in posterior distributions for treatment effect parameters. The CB method with exchangeable baseline response parameters produced effectiveness estimates in agreement with CBunadjInd, when the predicted baseline response estimates were similar to the observed baseline response. After adjustment for RA duration, there was a reduction in across-trial heterogeneity in baseline response but little change in treatment effect estimates. CONCLUSIONS: Our findings suggest incorporating SATs in NMA may be useful in some situations where a treatment is disconnected from a network of comparator treatments, due to a lack of comparative evidence, to estimate relative treatment effects. The reliability of effect estimates based on data from SATs may depend on adjustment for covariate effects, although further research is required to understand this in more detail.

Double-counting of populations in evidence synthesis in public health: a call for awareness and future methodological development.

BACKGROUND: There is a growing interest in the inclusion of real-world and observational studies in evidence synthesis such as meta-analysis and network meta-analysis in public health. While this approach offers great epidemiological opportunities, use of such studies often introduce a significant issue of double-counting of participants and databases in a single analysis. Therefore, this study aims to introduce and illustrate the nuances of double-counting of individuals in evidence synthesis including real-world and observational data with a focus on public health. METHODS: The issues associated with double-counting of individuals in evidence synthesis are highlighted with a number of case studies. Further, double-counting of information in varying scenarios is discussed with potential solutions highlighted. RESULTS: Use of studies of real-world data and/or established cohort studies, for example studies evaluating the effectiveness of therapies using health record data, often introduce a significant issue of double-counting of individuals and databases. This refers to the inclusion of the same individuals multiple times in a single analysis. Double-counting can occur in a number of manners, such as, when multiple studies utilise the same database, when there is overlapping timeframes of analysis or common treatment arms across studies. Some common practices to address this include synthesis of data only from peer-reviewed studies, utilising the study that provides the greatest information (e.g. largest, newest, greater outcomes reported) or analysing outcomes at different time points. CONCLUSIONS: While common practices currently used can mitigate some of the impact of double-counting of participants in evidence synthesis including real-world and observational studies, there is a clear need for methodological and guideline development to address this increasingly significant issue.

Effect of Transcatheter Aortic Valve Implantation vs Surgical Aortic Valve Replacement on All-Cause Mortality in Patients With Aortic Stenosis: A Randomized Clinical Trial.

Importance: Transcatheter aortic valve implantation (TAVI) is a less invasive alternative to surgical aortic valve replacement and is the treatment of choice for patients at high operative risk. The role of TAVI in patients at lower risk is unclear. Objective: To determine whether TAVI is noninferior to surgery in patients at moderately increased operative risk. Design, Setting, and Participants: In this randomized clinical trial conducted at 34 UK centers, 913 patients aged 70 years or older with severe, symptomatic aortic stenosis and moderately increased operative risk due to age or comorbidity were enrolled between April 2014 and April 2018 and followed up through April 2019. Interventions: TAVI using any valve with a CE mark (indicating conformity of the valve with all legal and safety requirements for sale throughout the European Economic Area) and any access route (n = 458) or surgical aortic valve replacement (surgery; n = 455). Main Outcomes and Measures: The primary outcome was all-cause mortality at 1 year. The primary hypothesis was that TAVI was noninferior to surgery, with a noninferiority margin of 5% for the upper limit of the 1-sided 97.5% CI for the absolute between-group difference in mortality. There were 36 secondary outcomes (30 reported herein), including duration of hospital stay, major bleeding events, vascular complications, conduction disturbance requiring pacemaker implantation, and aortic regurgitation. Results: Among 913 patients randomized (median age, 81 years [IQR, 78 to 84 years]; 424 [46%] were female; median Society of Thoracic Surgeons mortality risk score, 2.6% [IQR, 2.0% to 3.4%]), 912 (99.9%) completed follow-up and were included in the noninferiority analysis. At 1 year, there were 21 deaths (4.6%) in the TAVI group and 30 deaths (6.6%) in the surgery group, with an adjusted absolute risk difference of -2.0% (1-sided 97.5% CI, -∞ to 1.2%; P < .001 for noninferiority). Of 30 prespecified secondary outcomes reported herein, 24 showed no significant difference at 1 year. TAVI was associated with significantly shorter postprocedural hospitalization (median of 3 days [IQR, 2 to 5 days] vs 8 days [IQR, 6 to 13 days] in the surgery group). At 1 year, there were significantly fewer major bleeding events after TAVI compared with surgery (7.2% vs 20.2%, respectively; adjusted hazard ratio [HR], 0.33 [95% CI, 0.24 to 0.45]) but significantly more vascular complications (10.3% vs 2.4%; adjusted HR, 4.42 [95% CI, 2.54 to 7.71]), conduction disturbances requiring pacemaker implantation (14.2% vs 7.3%; adjusted HR, 2.05 [95% CI, 1.43 to 2.94]), and mild (38.3% vs 11.7%) or moderate (2.3% vs 0.6%) aortic regurgitation (adjusted odds ratio for mild, moderate, or severe [no instance of severe reported] aortic regurgitation combined vs none, 4.89 [95% CI, 3.08 to 7.75]). Conclusions and Relevance: Among patients aged 70 years or older with severe, symptomatic aortic stenosis and moderately increased operative risk, TAVI was noninferior to surgery with respect to all-cause mortality at 1 year. Trial Registration: isrctn.com Identifier: ISRCTN57819173.

Incorporating single-arm studies in meta-analysis of randomised controlled trials: a simulation study.

BACKGROUND: Use of real world data (RWD) from non-randomised studies (e.g. single-arm studies) is increasingly being explored to overcome issues associated with data from randomised controlled trials (RCTs). We aimed to compare methods for pairwise meta-analysis of RCTs and single-arm studies using aggregate data, via a simulation study and application to an illustrative example. METHODS: We considered contrast-based methods proposed by Begg & Pilote (1991) and arm-based methods by Zhang et al (2019). We performed a simulation study with scenarios varying (i) the proportion of RCTs and single-arm studies in the synthesis (ii) the magnitude of bias, and (iii) between-study heterogeneity. We also applied methods to data from a published health technology assessment (HTA), including three RCTs and 11 single-arm studies. RESULTS: Our simulation study showed that the hierarchical power and commensurate prior methods by Zhang et al provided a consistent reduction in uncertainty, whilst maintaining over-coverage and small error in scenarios where there was limited RCT data, bias and differences in between-study heterogeneity between the two sets of data. The contrast-based methods provided a reduction in uncertainty, but performed worse in terms of coverage and error, unless there was no marked difference in heterogeneity between the two sets of data. CONCLUSIONS: The hierarchical power and commensurate prior methods provide the most robust approach to synthesising aggregate data from RCTs and single-arm studies, balancing the need to account for bias and differences in between-study heterogeneity, whilst reducing uncertainty in estimates. This work was restricted to considering a pairwise meta-analysis using aggregate data.

Methods for the inclusion of real-world evidence in network meta-analysis.

BACKGROUND: Network Meta-Analysis (NMA) is a key component of submissions to reimbursement agencies world-wide, especially when there is limited direct head-to-head evidence for multiple technologies from randomised controlled trials (RCTs). Many NMAs include only data from RCTs. However, real-world evidence (RWE) is also becoming widely recognised as a valuable source of clinical data. This study aims to investigate methods for the inclusion of RWE in NMA and its impact on the level of uncertainty around the effectiveness estimates, with particular interest in effectiveness of fingolimod. METHODS: A range of methods for inclusion of RWE in evidence synthesis were investigated by applying them to an illustrative example in relapsing remitting multiple sclerosis (RRMS). A literature search to identify RCTs and RWE evaluating treatments in RRMS was conducted. To assess the impact of inclusion of RWE on the effectiveness estimates, Bayesian hierarchical and adapted power prior models were applied. The effect of the inclusion of RWE was investigated by varying the degree of down weighting of this part of evidence by the use of a power prior. RESULTS: Whilst the inclusion of the RWE led to an increase in the level of uncertainty surrounding effect estimates in this example, this depended on the method of inclusion adopted for the RWE. 'Power prior' NMA model resulted in stable effect estimates for fingolimod yet increasing the width of the credible intervals with increasing weight given to RWE data. The hierarchical NMA models were effective in allowing for heterogeneity between study designs, however, this also increased the level of uncertainty. CONCLUSION: The 'power prior' method for the inclusion of RWE in NMAs indicates that the degree to which RWE is taken into account can have a significant impact on the overall level of uncertainty. The hierarchical modelling approach further allowed for accommodating differences between study types. Consequently, further work investigating both empirical evidence for biases associated with individual RWE studies and methods of elicitation from experts on the extent of such biases is warranted.

Comparing associations between frailty and mortality in hospitalised older adults with or without COVID-19 infection: a retrospective observational study using electronic health records.

BACKGROUND: The aim of this study was to describe outcomes in hospitalised older people with different levels of frailty and COVID-19 infection. METHODS: We undertook a single-centre, retrospective cohort study examining COVID-19-related mortality using electronic health records, for older people (65 and over) with frailty, hospitalised with or without COVID-19 infection. Baseline covariates included demographics, early warning scores, Charlson Comorbidity Indices and frailty (Clinical Frailty Scale, CFS), linked to COVID-19 status. FINDINGS: We analysed outcomes on 1,071 patients with COVID-19 test results (285 (27%) were positive for COVID-19). The mean age at ED arrival was 79.7 and 49.4% were female. All-cause mortality (by 30 days) rose from 9 (not frail) to 33% (severely frail) in the COVID-negative cohort but was around 60% for all frailty categories in the COVID-positive cohort. In adjusted analyses, the hazard ratio for death in those with COVID-19 compared to those without COVID-19 was 7.3 (95% CI: 3.00, 18.0) with age, comorbidities and illness severity making small additional contributions. INTERPRETATION: In this study, frailty measured using the CFS appeared to make little incremental contribution to the hazard of dying in older people hospitalised with COVID-19 infection; illness severity and comorbidity had a modest association with the overall adjusted hazard of death, whereas confirmed COVID-19 infection dominated, with a sevenfold hazard for death.

A review of the quantitative effectiveness evidence synthesis methods used in public health intervention guidelines.

BACKGROUND: The complexity of public health interventions create challenges in evaluating their effectiveness. There have been huge advancements in quantitative evidence synthesis methods development (including meta-analysis) for dealing with heterogeneity of intervention effects, inappropriate 'lumping' of interventions, adjusting for different populations and outcomes and the inclusion of various study types. Growing awareness of the importance of using all available evidence has led to the publication of guidance documents for implementing methods to improve decision making by answering policy relevant questions. METHODS: The first part of this paper reviews the methods used to synthesise quantitative effectiveness evidence in public health guidelines by the National Institute for Health and Care Excellence (NICE) that had been published or updated since the previous review in 2012 until the 19th August 2019.The second part of this paper provides an update of the statistical methods and explains how they address issues related to evaluating effectiveness evidence of public health interventions. RESULTS: The proportion of NICE public health guidelines that used a meta-analysis as part of the synthesis of effectiveness evidence has increased since the previous review in 2012 from 23% (9 out of 39) to 31% (14 out of 45). The proportion of NICE guidelines that synthesised the evidence using only a narrative review decreased from 74% (29 out of 39) to 60% (27 out of 45).An application in the prevention of accidents in children at home illustrated how the choice of synthesis methods can enable more informed decision making by defining and estimating the effectiveness of more distinct interventions, including combinations of intervention components, and identifying subgroups in which interventions are most effective. CONCLUSIONS: Despite methodology development and the publication of guidance documents to address issues in public health intervention evaluation since the original review, NICE public health guidelines are not making full use of meta-analysis and other tools that would provide decision makers with fuller information with which to develop policy. There is an evident need to facilitate the translation of the synthesis methods into a public health context and encourage the use of methods to improve decision making.

Bayesian hierarchical meta-analytic methods for modeling surrogate relationships that vary across treatment classes using aggregate data.

Surrogate endpoints play an important role in drug development when they can be used to measure treatment effect early compared to the final clinical outcome and to predict clinical benefit or harm. Such endpoints are assessed for their predictive value of clinical benefit by investigating the surrogate relationship between treatment effects on the surrogate and final outcomes using meta-analytic methods. When surrogate relationships vary across treatment classes, such validation may fail due to limited data within each treatment class. In this paper, two alternative Bayesian meta-analytic methods are introduced which allow for borrowing of information from other treatment classes when exploring the surrogacy in a particular class. The first approach extends a standard model for the evaluation of surrogate endpoints to a hierarchical meta-analysis model assuming full exchangeability of surrogate relationships across all the treatment classes, thus facilitating borrowing of information across the classes. The second method is able to relax this assumption by allowing for partial exchangeability of surrogate relationships across treatment classes to avoid excessive borrowing of information from distinctly different classes. We carried out a simulation study to assess the proposed methods in nine data scenarios and compared them with subgroup analysis using the standard model within each treatment class. We also applied the methods to an illustrative example in colorectal cancer which led to obtaining the parameters describing the surrogate relationships with higher precision.

Multivariate network meta-analysis incorporating class effects.

BACKGROUND: Network meta-analysis synthesises data from a number of clinical trials in order to assess the comparative efficacy of multiple healthcare interventions in similar patient populations. In situations where clinical trial data are heterogeneously reported i.e. data are missing for one or more outcomes of interest, synthesising such data can lead to disconnected networks of evidence, increased uncertainty, and potentially biased estimates which can have severe implications for decision-making. To overcome this issue, strength can be borrowed between outcomes of interest in multivariate network meta-analyses. Furthermore, in situations where there are relatively few trials informing each treatment comparison, there is a potential issue with the sparsity of data in the treatment networks, which can lead to substantial parameter uncertainty. A multivariate network meta-analysis approach can be further extended to borrow strength between interventions of the same class using hierarchical models. METHODS: We extend the trivariate network meta-analysis model to incorporate the exchangeability between treatment effects belonging to the same class of intervention to increase precision in treatment effect estimates. We further incorporate a missing data framework to estimate uncertainty in trials that did not report measures of variability in order to maximise the use of all available information for healthcare decision-making. The methods are applied to a motivating dataset in overactive bladder syndrome. The outcomes of interest were mean change from baseline in incontinence, voiding and urgency episodes. All models were fitted using Bayesian Markov Chain Monte Carlo (MCMC) methods in WinBUGS. RESULTS: All models (univariate, multivariate, and multivariate models incorporating class effects) produced similar point estimates for all treatment effects. Incorporating class effects in multivariate models often increased precision in treatment effect estimates. CONCLUSIONS: Multivariate network meta-analysis incorporating class effects allowed for the comparison of all interventions across all outcome measures to ameliorate the potential impact of outcome reporting bias, and further borrowed strength between interventions belonging to the same class of treatment to increase the precision in treatment effect estimates for healthcare policy and decision-making.

Mixed-effects models for health care longitudinal data with an informative visiting process: A Monte Carlo simulation study.

Electronic health records are being increasingly used in medical research to answer more relevant and detailed clinical questions; however, they pose new and significant methodological challenges. For instance, observation times are likely correlated with the underlying disease severity: Patients with worse conditions utilise health care more and may have worse biomarker values recorded. Traditional methods for analysing longitudinal data assume independence between observation times and disease severity; yet, with health care data, such assumptions unlikely hold. Through Monte Carlo simulation, we compare different analytical approaches proposed to account for an informative visiting process to assess whether they lead to unbiased results. Furthermore, we formalise a joint model for the observation process and the longitudinal outcome within an extended joint modelling framework. We illustrate our results using data from a pragmatic trial on enhanced care for individuals with chronic kidney disease, and we introduce user-friendly software that can be used to fit the joint model for the observation process and a longitudinal outcome.

Impact of model misspecification in shared frailty survival models.

Survival models incorporating random effects to account for unmeasured heterogeneity are being increasingly used in biostatistical and applied research. Specifically, unmeasured covariates whose lack of inclusion in the model would lead to biased, inefficient results are commonly modeled by including a subject-specific (or cluster-specific) frailty term that follows a given distribution (eg, gamma or lognormal). Despite that, in the context of parametric frailty models, little is known about the impact of misspecifying the baseline hazard or the frailty distribution or both. Therefore, our aim is to quantify the impact of such misspecification in a wide variety of clinically plausible scenarios via Monte Carlo simulation, using open-source software readily available to applied researchers. We generate clustered survival data assuming various baseline hazard functions, including mixture distributions with turning points, and assess the impact of sample size, variance of the frailty, baseline hazard function, and frailty distribution. Models compared include standard parametric distributions and more flexible spline-based approaches; we also included semiparametric Cox models. The resulting bias can be clinically relevant. In conclusion, we highlight the importance of fitting models that are flexible enough and the importance of assessing model fit. We illustrate our conclusions with two applications using data on diabetic retinopathy and bladder cancer. Our results show the importance of assessing model fit with respect to the baseline hazard function and the distribution of the frailty: misspecifying the former leads to biased relative and absolute risk estimates, whereas misspecifying the latter affects absolute risk estimates and measures of heterogeneity.

Bivariate network meta-analysis for surrogate endpoint evaluation.

Surrogate endpoints are very important in regulatory decision making in healthcare, in particular if they can be measured early compared to the long-term final clinical outcome and act as good predictors of clinical benefit. Bivariate meta-analysis methods can be used to evaluate surrogate endpoints and to predict the treatment effect on the final outcome from the treatment effect measured on a surrogate endpoint. However, candidate surrogate endpoints are often imperfect, and the level of association between the treatment effects on the surrogate and final outcomes may vary between treatments. This imposes a limitation on methods which do not differentiate between the treatments. We develop bivariate network meta-analysis (bvNMA) methods, which combine data on treatment effects on the surrogate and final outcomes, from trials investigating multiple treatment contrasts. The bvNMA methods estimate the effects on both outcomes for all treatment contrasts individually in a single analysis. At the same time, they allow us to model the trial-level surrogacy patterns within each treatment contrast and treatment-level surrogacy, thus enabling predictions of the treatment effect on the final outcome either for a new study in a new population or for a new treatment. Modelling assumptions about the between-studies heterogeneity and the network consistency, and their impact on predictions, are investigated using an illustrative example in advanced colorectal cancer and in a simulation study. When the strength of the surrogate relationships varies across treatment contrasts, bvNMA has the advantage of identifying treatment comparisons for which surrogacy holds, thus leading to better predictions.

Towards regulatory endorsement of drug development tools to promote the application of model-informed drug development in Duchenne muscular dystrophy.

Drug development for rare diseases is challenged by small populations and limited data. This makes development of clinical trial protocols difficult and contributes to the uncertainty around whether or not a potential therapy is efficacious. The use of data standards to aggregate data from multiple sources, and the use of such integrated databases to develop statistical models can inform protocol development and reduce the risks in developing new therapies. Achieving regulatory endorsement of such models through defined pathways at the US Food and Drug Administration and European Medicines Authority allows such tools to be used by the drug development community for defined contexts of use without further need for discussion of the underlying model(s). The Duchenne Regulatory Science Consortium (D-RSC) has brought together multiple stakeholders to develop a clinical trial simulation tool for Duchenne muscular dystrophy using such an approach. Here we describe the work of D-RSC as an example of how such an approach may be effective at reducing uncertainty in drug development for rare diseases, and thus bringing effective therapies to patients faster.

Accounting for Uncertainty in Decision Analytic Models Using Rank Preserving Structural Failure Time Modeling: Application to Parametric Survival Models.

OBJECTIVES: Rank Preserving Structural Failure Time models are one of the most commonly used statistical methods to adjust for treatment switching in oncology clinical trials. The method is often applied in a decision analytic model without appropriately accounting for additional uncertainty when determining the allocation of health care resources. The aim of the study is to describe novel approaches to adequately account for uncertainty when using a Rank Preserving Structural Failure Time model in a decision analytic model. METHODS: Using two examples, we tested and compared the performance of the novel Test-based method with the resampling bootstrap method and with the conventional approach of no adjustment. In the first example, we simulated life expectancy using a simple decision analytic model based on a hypothetical oncology trial with treatment switching. In the second example, we applied the adjustment method on published data when no individual patient data were available. RESULTS: Mean estimates of overall and incremental life expectancy were similar across methods. However, the bootstrapped and test-based estimates consistently produced greater estimates of uncertainty compared with the estimate without any adjustment applied. Similar results were observed when using the test based approach on a published data showing that failing to adjust for uncertainty led to smaller confidence intervals. CONCLUSIONS: Both the bootstrapping and test-based approaches provide a solution to appropriately incorporate uncertainty, with the benefit that the latter can implemented by researchers in the absence of individual patient data.

Comparison of the effectiveness of repeated injections of onabotulinum toxin A for refractory idiopathic detrusor overactivity: analysis of an open label extension of a randomized trial (the RELAX study).

AIMS: To assess effects of repeat treatment with onabotulinumtoxin A (onaBoNT-A) in women with refractory idiopathic detrusor overactivity (DO). METHODS: Analysis of an open-label extension study of a large randomized placebo controlled trial of onaBoNT-A. Participants had been randomized to receive 200 IU onaBoNTA or placebo and were offered up to two further onaBoNTA injections over a 5-year period. For this analysis, the primary outcome was duration of treatment effect by patient-reported symptom return. Weibull proportional hazards regression models were fitted in a Bayesian framework to estimate missing times. Multivariable hazard regression analysis (hazard ratio, 95% credible intervals (HR, 95% CrI) compared repeated injections adjusting for differences in baseline symptom severity. Secondary outcomes included inter-injection interval, incontinence, urgency, and voiding episodes 6 weeks after injection. RESULTS: Four hundred and forty-two active injections were administered: 228 patients had one, 155 had two, and 59 had three injections. Time to symptom return for injection number 1 and 2 was 84 (95%CI: 63, 112) and 180 (95%CI: 135, 223) days, respectively. Median inter-injection intervals for receiving second and third injection were 266 days (range: 130, 1400) and 372 days (range: 134, 1283). No statistically significant differences in symptom outcomes or time to symptom return (HR 0.88, 95% CrI 0.37, 2.07 for injection 2, HR 0.33, 95% CrI 0.09, 1.03 for injection 3) were observed. CONCLUSIONS: Repeated onaBoNT-A injections have consistent efficacy and duration of action. There appears to be long-term placebo effects in both groups of randomized patients, with implications for open-label extension studies.