Multilevel mixed effects parametric survival models using adaptive Gauss-Hermite quadrature with application to recurrent events and individual participant data meta-analysis

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.

A flexible parametric accelerated failure time model and the extension to time-dependent acceleration factors

Accelerated failure time (AFT) models are used widely in medical research, though to a much lesser extent than proportional hazards models. In an AFT model, the effect of covariates act to accelerate or decelerate the time to event of interest, that is, shorten or extend the time to event. Commonly used parametric AFT models are limited in the underlying shapes that they can capture. In this article, we propose a general parametric AFT model, and in particular concentrate on using restricted cubic splines to model the baseline to provide substantial flexibility. We then extend the model to accommodate time-dependent acceleration factors. Delayed entry is also allowed, and hence, time-dependent covariates. We evaluate the proposed model through simulation, showing substantial improvements compared to standard parametric AFT models. We also show analytically and through simulations that the AFT models are collapsible, suggesting that this model class will be well suited to causal inference. We illustrate the methods with a data set of patients with breast cancer. Finally, we provide highly efficient, user-friendly Stata, and R software packages.

Correcting for heterogeneity and non-comparability bias in multicenter clinical trials with a rescaled random-effect excess hazard model

In the presence of competing causes of event occurrence (e.g., death), the interest might not only be in the overall survival but also in the so-called net survival, that is, the hypothetical survival that would be observed if the disease under study were the only possible cause of death. Net survival estimation is commonly based on the excess hazard approach in which the hazard rate of individuals is assumed to be the sum of a disease-specific and expected hazard rate, supposed to be correctly approximated by the mortality rates obtained from general population life tables. However, this assumption might not be realistic if the study participants are not comparable with the general population. Also, the hierarchical structure of the data can induces a correlation between the outcomes of individuals coming from the same clusters (e.g., hospital, registry). We proposed an excess hazard model that corrects simultaneously for these two sources of bias, instead of dealing with them independently as before. We assessed the performance of this new model and compared it with three similar models, using extensive simulation study, as well as an application to breast cancer data from a multicenter clinical trial. The new model performed better than the others in terms of bias, root mean square error, and empirical coverage rate. The proposed approach might be useful to account simultaneously for the hierarchical structure of the data and the non-comparability bias in studies such as long-term multicenter clinical trials, when there is interest in the estimation of net survival.

Using individual participant data to improve network meta-analysis projects

A network meta-analysis combines the evidence from existing randomised trials about the comparative efficacy of multiple treatments. It allows direct and indirect evidence about each comparison to be included in the same analysis, and provides a coherent framework to compare and rank treatments. A traditional network meta-analysis uses aggregate data (eg, treatment effect estimates and standard errors) obtained from publications or trial investigators. An alternative approach is to obtain, check, harmonise and meta-analyse the individual participant data (IPD) from each trial. In this article, we describe potential advantages of IPD for network meta-analysis projects, emphasising five key benefits: (1) improving the quality and scope of information available for inclusion in the meta-analysis, (2) examining and plotting distributions of covariates across trials (eg, for potential effect modifiers), (3) standardising and improving the analysis of each trial, (4) adjusting for prognostic factors to allow a network meta-analysis of conditional treatment effects and (5) including treatment-covariate interactions (effect modifiers) to allow relative treatment effects to vary by participant-level covariate values (eg, age, baseline depression score). A running theme of all these benefits is that they help examine and reduce heterogeneity (differences in the true treatment effect between trials) and inconsistency (differences in the true treatment effect between direct and indirect evidence) in the network. As a consequence, an IPD network meta-analysis has the potential for more precise, reliable and informative results for clinical practice and even allows treatment comparisons to be made for individual patients and targeted populations conditional on their particular characteristics.

MEGH: A parametric class of general hazard models for clustered survival data

In many applications of survival data analysis, the individuals are treated in different medical centres or belong to different clusters defined by geographical or administrative regions. The analysis of such data requires accounting for between-cluster variability. Ignoring such variability would impose unrealistic assumptions in the analysis and could affect the inference on the statistical models. We develop a novel parametric mixed-effects general hazard (MEGH) model that is particularly suitable for the analysis of clustered survival data. The proposed structure generalises the mixed-effects proportional hazards and mixed-effects accelerated failure time structures, among other structures, which are obtained as special cases of the MEGH structure. We develop a likelihood-based algorithm for parameter estimation in general subclasses of the MEGH model, which is implemented in our R package MEGH. We propose diagnostic tools for assessing the random effects and their distributional assumption in the proposed MEGH model. We investigate the performance of the MEGH model using theoretical and simulation studies, as well as a real data application on leukaemia.

Analyzing factors associated with time to age at first marriage among women in Ethiopia: log logistic-gamma shared frailty model

OBJECTIVE

The main objective of this study is to fit Log logistic-Gamma shared frailty model for the determinant of time to age at first marriage among women in Ethiopia.

METHODS

The data set in this study were obtained from Demography and Health survey conducted in Ethiopia in 2016. In this study, we used Log logistic-Gamma shared frailty model to account for the loss of independence that arises from the clustering of women in region of Ethiopia. A total of 12,066 women aged 15-49 in Ethiopia were included in this study.

RESULTS

Of all 12,066 women aged 15-49, 9466 (78.45%) were married and the median & mean age at first marriage for women living in Ethiopia were 17.2 years and 17.5 years respectively, while the minimum and maximum age at first marriage observed were 8 years and 49 years respectively.

CONCLUSION

The most significant contributing factors to delaying time to age at first marriage of women aged 15-49 in Ethiopia were increased education level of women, increased education level of the head, increased income, residing in urban and being followers of religion other than orthodox, catholic, protestant & Muslim. The heterogeneity of age at first marriage for women aged 15-49 among regions in Ethiopia was observed. The government of Ethiopia and the concerned bodies should revise the women's health policy and practice to reduce early marriage and give attention to women; illiterate, live in rural areas, and have illiterate and poor heads.

Challenges of modelling approaches for network meta-analysis of time-to-event outcomes in the presence of non-proportional hazards to aid decision making: Application to a melanoma network

BACKGROUND

Synthesis of clinical effectiveness from multiple trials is a well-established component of decision-making. Time-to-event outcomes are often synthesised using the Cox proportional hazards model assuming a constant hazard ratio over time. However, with an increasing proportion of trials reporting treatment effects where hazard ratios vary over time and with differing lengths of follow-up across trials, alternative synthesis methods are needed.

OBJECTIVES

To compare and contrast five modelling approaches for synthesis of time-to-event outcomes and provide guidance on key considerations for choosing between the modelling approaches.

METHODS

The Cox proportional hazards model and five other methods of estimating treatment effects from time-to-event outcomes, which relax the proportional hazards assumption, were applied to a network of melanoma trials reporting overall survival: restricted mean survival time, generalised gamma, piecewise exponential, fractional polynomial and Royston-Parmar models.

RESULTS

All models fitted the melanoma network acceptably well. However, there were important differences in extrapolations of the survival curve and interpretability of the modelling constraints demonstrating the potential for different conclusions from different modelling approaches.

CONCLUSION

The restricted mean survival time, generalised gamma, piecewise exponential, fractional polynomial and Royston-Parmar models can accommodate non-proportional hazards and differing lengths of trial follow-up within a network meta-analysis of time-to-event outcomes. We recommend that model choice is informed using available and relevant prior knowledge, model transparency, graphically comparing survival curves alongside observed data to aid consideration of the reliability of the survival estimates, and consideration of how the treatment effect estimates can be incorporated within a decision model.

Modeling the determinant of time to age at first marriage among women in Ethiopia using Cox models with mixed effects

Background

Time to age at first marriage of women is the duration of time until the age at which they started living with their first partner. Time to age at first marriage is widely considered a proxy indicator for the age at which women begin to be exposed to the risks inherent in sexual activity. The purpose is to model the determinant of time to age at first marriage among women in Ethiopia using Cox models with mixed effects.

Methods

The 2016 Ethiopian Demography and Health survey sample was selected using a two-stage cluster design. The data set in this study were obtained from the Demography and Health survey conducted in Ethiopia in 2016. In this study, we used Cox models with mixed effects.

Results

Of all 15,683 women aged 15–49 years, 11,405 (72.72%) were married with the median and mean age at first marriage 17 years and 18 years, respectively. Cox frailty survival model showed that residence, educational level, occupation, work status of women& head education level of households were the most significant factors whereas religion, access to media and wealth index of a household of women were not significant factors at 5% level of significance. The significant clustering effect showed that heterogeneity among the regions on age at first marriage was present.

Conclusions

The present study determined the duration of time until the age at first marriage and indicated relevant solutions for marriage-related problems of women aged 15–49 years in Ethiopia. Women residing in rural area of Ethiopia and had lower education level were married earlier. Therefore, programs to reduce the high rate of early marriage in Ethiopia should give attention to women education and women residing in rural area.

Time to age at first marriage of women is the duration of time until the age at which they started living with their first partner. African women are more likely to marry earlier than other continent women, which causes high fertility due to their long period of exposure to the risk of pregnancy. Even though Sub-Sahara Africa accounts for the highest rate of age at first marriage among countries in the Africa continent, comparably the case is very worse in Ethiopia. Furthermore, there is no study about the determinant of time to age at first marriage in Ethiopia using advanced models like Cox Model with Mixed effects. This model allows for the analysis of data with complex patterns of variability, with a focus on nested sources of variability. Very often it makes sense to use such a model to represent the variability within and between groups. For instance, in this study, our interest is not only knowing the significant effect of determinant factors on time to age at first marriage among women in Ethiopia but also the variability of time to age at first marriage within the region of Ethiopia and between regions of Ethiopia. We have used a dataset for this study from Ethiopia Demography and Health Survey which was conducted in 2016. The study helps to indicate relevant solutions for women's marriage-related problems (predominantly, women's reproductive health problems) in Ethiopia and it provides input for further studies in Ethiopia.

Comparison of statistical methods for the analysis of recurrent adverse events in the presence of non-proportional hazards and unobserved heterogeneity: a simulation study

Background

In preventive drug trials such as intermittent preventive treatment for malaria prevention during pregnancy (IPTp), where there is repeated treatment administration, recurrence of adverse events (AEs) is expected. Challenges in modelling the risk of the AEs include accounting for time-to-AE and within-patient-correlation, beyond the conventional methods. The correlation comes from two sources; (a) individual patient unobserved heterogeneity (i.e. frailty) and (b) the dependence between AEs characterised by time-dependent treatment effects. Potential AE-dependence can be modelled via time-dependent treatment effects, event-specific baseline and event-specific random effect, while heterogeneity can be modelled via subject-specific random effect. Methods that can improve the estimation of both the unobserved heterogeneity and treatment effects can be useful in understanding the evolution of risk of AEs, especially in preventive trials where time-dependent treatment effect is expected.

Methods

Using both a simulation study and the Chloroquine for Malaria in Pregnancy (NCT01443130) trial data to demonstrate the application of the models, we investigated whether the lognormal shared frailty models with restricted cubic splines and non-proportional hazards (LSF-NPH) assumption can improve estimates for both frailty variance and treatment effect compared to the conventional inverse Gaussian shared frailty model with proportional hazard (ISF-PH), in the presence of time-dependent treatment effects and unobserved patient heterogeneity. We assessed the bias, precision gain and coverage probability of 95% confidence interval of the frailty variance estimates for the models under varying known unobserved heterogeneity, sample sizes and time-dependent effects.

Results

The ISF-PH model provided a better coverage probability of 95% confidence interval, less bias and less precise frailty variance estimates compared to the LSF-NPH models. The LSF-NPH models yielded unbiased hazard ratio estimates at the expense of imprecision and high mean square error compared to the ISF-PH model.

Conclusion

The choice of the shared frailty model for the recurrent AEs analysis should be driven by the study objective. Using the LSF-NPH models is appropriate if unbiased hazard ratio estimation is of primary interest in the presence of time-dependent treatment effects. However, ISF-PH model is appropriate if unbiased frailty variance estimation is of primary interest.

Trial registration

ClinicalTrials.gov; NCT01443130

Individual participant data meta-analysis with mixed-effects transformation models

One-stage meta-analysis of individual participant data (IPD) poses several statistical and computational challenges. For time-to-event outcomes, the approach requires the estimation of complicated nonlinear mixed-effects models that are flexible enough to realistically capture the most important characteristics of the IPD. We present a model class that incorporates general normally distributed random effects into linear transformation models. We discuss extensions to model between-study heterogeneity in baseline risks and covariate effects and also relax the assumption of proportional hazards. Within the proposed framework, data with arbitrary random censoring patterns can be handled. The accompanying \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$\textsf{R}$\end{document} package tramME utilizes the Laplace approximation and automatic differentiation to perform efficient maximum likelihood estimation and inference in mixed-effects transformation models. We compare several variants of our model to predict the survival of patients with chronic obstructive pulmonary disease using a large data set of prognostic studies. Finally, a simulation study is presented that verifies the correctness of the implementation and highlights its efficiency compared to an alternative approach.

A cross-sectional and longitudinal study of neighbourhood disadvantage and cardiovascular disease and the mediating role of physical activity

We investigate the prospective association between neighbourhood-level disadvantage and cardiovascular disease (CVD) among mid-to-older aged adults and whether physical activity (PA) mediates this association. The data come from the HABITAT project, a multilevel longitudinal investigation of health and wellbeing in Brisbane. The participants were 11,035 residents of 200 neighbourhoods in 2007, with follow-up data collected in 2009, 2011, 2013 and 2016. Multilevel binomial regression was used for the cross-sectional analysis and mixed-effect parametric survival models were used for the longitudinal analysis. Models were adjusted for age, sex, education, occupation, and household income. Those with pre-existing CVD at baseline were excluded from the longitudinal analyses. The mediated effect of PA on CVD was examined using multilevel generalized structural equation modelling. There was a total of 20,064 person-year observations across the five time-points clustered at three levels. Results indicated that the incidence of CVD was significantly higher in the most disadvantaged neighbourhoods (OR 1.50; HR 1.29) compared with the least disadvantaged. Mediation analysis results revealed that 11.5% of the effect of neighbourhood disadvantage on CVD occurs indirectly through PA in the most disadvantaged neighbourhoods while the corresponding figure is 5.2% in the more advantaged areas. Key findings showed that neighbourhood disadvantage is associated with the incidence of CVD, and PA is a significant mediator of this relationship. Future research should investigate which specific social and built environment features promote or inhibit PA in disadvantaged areas as the basis for policy initiatives to address inequities in CVD.

Unmarried or less-educated patients with mantle cell lymphoma are less likely to undergo a transplant, leading to lower survival

It is unknown how many mantle cell lymphoma (MCL) patients undergo consolidation with autologous hematopoietic cell transplantation (AHCT), and the reasons governing the decision, are also unknown. The prognostic impact of omitting AHCT is also understudied. We identified all MCL patients diagnosed from 2000 to 2014, aged 18 to 65 years, in the Swedish Lymphoma Register. Odds ratios (ORs) and 95% confidence intervals (CIs) from logistic regression models were used to compare the likelihood of AHCT within 18 months of diagnosis. All-cause mortality was compared between patients treated with/without AHCT using hazard ratios (HRs) and 95% CIs estimated from Cox regression models. Probabilities of being in each of the following states: alive without AHCT, alive with AHCT, dead before AHCT, and dead after AHCT, were estimated over time from an illness-death model. Among 369 patients, 148 (40%) were not treated with AHCT within 18 months. Compared with married patients, never married and divorced patients had lower likelihood of undergoing AHCT, as had patients with lower educational level, and comorbid patients. Receiving AHCT was associated with reduced all-cause mortality (HR = 0.58, 95% CI: 0.40-0.85). Transplantation-related mortality was low (2%). MCL patients not receiving an AHCT had an increased mortality rate, and furthermore, an undue concern about performing an AHCT in certain societal groups was seen. Improvements in supportive functions potentially increasing the likelihood of tolerating an AHCT and introduction of more tolerable treatments for these groups are needed.

Biodegradable polymer sirolimus-eluting stents vs durable polymer everolimus-eluting stents in patients undergoing percutaneous coronary intervention: A meta-analysis of individual patient data from 5 randomized trials

BACKGROUND

Newest generation drug-eluting stents combine biodegradable polymers with ultrathin stent platforms in order to minimize vessel injury and inflammatory response. Evidence from randomized controlled trials suggested that differences in stent design translate into differences in clinical outcome. The aim of the present study was to evaluate the safety and efficacy of ultrathin strut, biodegradable polymer sirolimus eluting stents (BP SES) compared with thin strut, durable polymer everolimus-eluting stents (DP EES) among patients undergoing percutaneous coronary intervention (PCI).

METHODS

We pooled individual participant data from 5 randomized trials (NCT01356888, NCT01939249, NCT02389946, NCT01443104, NCT02579031) including a total of 5,780 patients, and performed a one-stage meta-analysis using a mixed effects Cox regression model.

RESULTS

At a median duration of follow-up of 739 days (interquartile range 365-1,806 days), target-lesion failure occurred in 337 (10.3%) and 304 (12.2%) patients treated with BP SES and DP EES (HR 0.86, 95%CI 0.71-1.06, P = .16). There were no significant differences between BP SES and DP EES with regards to cardiac death (111 (3.4%) vs 102 (4.1%); HR 1.05, 95%CI 0.80-1.37, P = .73), target-vessel myocardial infarction (136 (4.1%) vs 126 (5.0%), HR 0.79, 95%CI 0.62-1.01, P = .061), and clinically-driven target-lesion revascularization (163 (5.0%) vs 147 (5.9%); HR 0.94, 95%CI 0.75-1.18, P = .61). The effect was consistent across major subgroups. In a landmark analysis, there was no significant interaction between treatment effect and timing of events.

CONCLUSIONS

In this patient-level meta-analysis of 5 randomized controlled trials, BP SES were associated with a similar risk of target-lesion failure compared with DP EES among patients undergoing PCI.

STUDY REGISTRATION

PROSPERO registry (CRD42018109098).

Bayesian Stacked Parametric Survival with Frailty Components and Interval-Censored Failure Times: An Application to Food Allergy Risk

To better understand the risk of exposure to food allergens, food challenge studies are designed to slowly increase the dose of an allergen delivered to allergic individuals until an objective reaction occurs. These dose-to-failure studies are used to determine acceptable intake levels and are analyzed using parametric failure time models. Though these models can provide estimates of the survival curve and risk, their parametric form may misrepresent the survival function for doses of interest. Different models that describe the data similarly may produce different dose-to-failure estimates. Motivated by predictive inference, we developed a Bayesian approach to combine survival estimates based on posterior predictive stacking, where the weights are formed to maximize posterior predictive accuracy. The approach defines a model space that is much larger than traditional parametric failure time modeling approaches. In our case, we use the approach to include random effects accounting for frailty components. The methodology is investigated in simulation, and is used to estimate allergic population eliciting doses for multiple food allergens.

Risk factors for indicators of opioid-related harms amongst people living with chronic non-cancer pain: Findings from a 5-year prospective cohort study

BACKGROUND

The literature suggests patient characteristics and higher opioid doses and long-term duration are associated with problematic opioid behaviours but no one study has examined the role of all these factors simultaneously in a long-term prospective cohort study.

METHODS

Five-year, community-based, prospective cohort of people prescribed opioids for chronic non-cancer pain (CNCP). Logistic mixed effect models with multiple imputation were used to address missing data. Oral morphine equivalent (OME) mg per day was categorised as: 0 mg OME/day, 1-49 mg OME/day (reference), 50-89 mg OME/day, 90-199 mg OME/day and 200mg+ OME/day. Patient risk factors included: age, gender, substance use, mental health history and pain-related factors. Main outcomes included: Prescribed Opioids Difficulties Scale (PODS), Opioid-Related Behaviours In Treatment (ORBIT) scale, and ICD-10 opioid dependence. Multiple confounders for problematic opioid behaviours were assessed.

FINDINGS

Of 1,514 participants 44.4% were male (95%CI 41.9-46.9) and their mean age was 58 years (IQR 48-67). Participants had a mean duration of pain of 10 years (IQR 4.5-20.0) and had been taking strong opioids for a median of four years (IQR 1.0-10.0). At baseline, median OME/day was 73 (IQR 35-148). At 5-years, 85% were still taking strong opioids. PODS moderate-high scores reduced from 59.9% (95%CI 58.8-61.0) at baseline to 51.5% (95%CI 50.0-53.0) at 5-years. Around 9% met criteria for ICD-10 opioid dependence at each wave. In adjusted mixed effect models, the risk factors most consistently associated with problematic opioid use were: younger age, substance dependence, mental health histories and higher opioid doses.

INTERPRETATION

Both patient risk factors and opioid dose are associated with problematic opioid use behaviours.

Socioeconomic Inequalities in Colorectal Cancer Survival in Southern Spain: A Multilevel Population-Based Cohort Study

BACKGROUND

Colorectal cancer (CRC) is the most frequently diagnosed cancer in Spain. Socioeconomic inequalities in cancer survival are not documented in Spain. We aim to study the association of socioeconomic inequalities with overall mortality and survival among CRC patients in southern Spain.

METHODS

We conducted a multilevel population-based cohort study, including CRC cases for the period 2011-2013. The study time-to-event outcome was death, and the primary exposure was CRC patients' socioeconomic status assessed by the Spanish deprivation index at the census tract level. We used a mixed-effects flexible hazard model, including census tract as a random intercept, to derive overall survival estimates by deprivation.

RESULTS

Among 3589 CRC patients and 12,148 person-years at risk (pyr), 964 patients died before the end of the follow-up. Mortality by deprivation showed the highest mortality rate for the most deprived group (96.2 per 1000 pyr, 95% CI: 84.0-110.2). After adjusting for sex, age, cancer stage, and the area of residence, the most deprived had a 60% higher excess mortality risk than the less deprived group (excess mortality risk ratio: 1.6, 95% CI: 1.1-2.3).

CONCLUSIONS

We found a consistent association between deprivation and CRC excess mortality and survival. The reasons behind these inequalities need further investigation in order to improve equality cancer outcomes in all social groups.

Continuation of emtricitabine/lamivudine within combination antiretroviral therapy following detection of the M184V/I HIV-1 resistance mutation

OBJECTIVES

The aim of the study was to investigate whether lamivudine (3TC) or emtricitabine (FTC) use following detection of M184V/I is associated with better virological outcomes.

METHODS

We identified people with viruses harbouring the M184V/I mutation in UK multicentre data sets who had treatment change/initiation within 1 year. We analysed outcomes of viral suppression (< 200 HIV-1 RNA copies/mL) and appearance of new major drug resistance mutations (DRMs) using Cox and Poisson models, with stratification by new drug regimen (excluding 3TC/FTC) and Bayesian implementation, and estimated the effect of 3TC/FTC adjusted for individual and viral characteristics.

RESULTS

We included 2597 people with the M184V/I resistance mutation, of whom 665 (25.6%) were on 3TC and 458 (17.6%) on FTC. We found a negative adjusted association between 3TC/FTC use and viral suppression [hazard ratio (HR) 0.84; 95% credibility interval (CrI) 0.71-0.98]. On subgroup analysis of individual drugs, there was no evidence of an association with viral suppression for 3TC (n = 184; HR 0.94; 95% CrI 0.73-1.15) or FTC (n = 454; HR 0.99; 95% CrI 0.80-1.19) amongst those on tenofovir-containing regimens, but we estimated a reduced rate of viral suppression for people on 3TC amongst those without tenofovir use (n = 481; HR 0.71; 95% CrI 0.54-0.90). We found no association between 3TC/FTC and detection of any new DRM (overall HR 0.92; 95% CrI 0.64-1.18), but found inconclusive evidence of a lower incidence rate of new DRMs (overall incidence rate ratio 0.69; 95% CrI 0.34-1.11).

CONCLUSIONS

We did not find evidence that 3TC or FTC use is associated with an increase in viral suppression, but it may reduce the appearance of additional DRMs in people with M184V/I. 3TC was associated with reduced viral suppression amongst people on regimens without tenofovir.

Individual participant data meta-analysis to examine interactions between treatment effect and participant-level covariates: Statistical recommendations for conduct and planning

Precision medicine research often searches for treatment‐covariate interactions, which refers to when a treatment effect (eg, measured as a mean difference, odds ratio, hazard ratio) changes across values of a participant‐level covariate (eg, age, gender, biomarker). Single trials do not usually have sufficient power to detect genuine treatment‐covariate interactions, which motivate the sharing of individual participant data (IPD) from multiple trials for meta‐analysis. Here, we provide statistical recommendations for conducting and planning an IPD meta‐analysis of randomized trials to examine treatment‐covariate interactions. For conduct, two‐stage and one‐stage statistical models are described, and we recommend: (i) interactions should be estimated directly, and not by calculating differences in meta‐analysis results for subgroups; (ii) interaction estimates should be based solely on within‐study information; (iii) continuous covariates and outcomes should be analyzed on their continuous scale; (iv) nonlinear relationships should be examined for continuous covariates, using a multivariate meta‐analysis of the trend (eg, using restricted cubic spline functions); and (v) translation of interactions into clinical practice is nontrivial, requiring individualized treatment effect prediction. For planning, we describe first why the decision to initiate an IPD meta‐analysis project should not be based on between‐study heterogeneity in the overall treatment effect; and second, how to calculate the power of a potential IPD meta‐analysis project in advance of IPD collection, conditional on characteristics (eg, number of participants, standard deviation of covariates) of the trials (potentially) promising their IPD. Real IPD meta‐analysis projects are used for illustration throughout.

Individual participant data meta-analysis of intervention studies with time-to-event outcomes: A review of the methodology and an applied example

Many randomized trials evaluate an intervention effect on time-to-event outcomes. Individual participant data (IPD) from such trials can be obtained and combined in a so-called IPD meta-analysis (IPD-MA), to summarize the overall intervention effect. We performed a narrative literature review to provide an overview of methods for conducting an IPD-MA of randomized intervention studies with a time-to-event outcome. We focused on identifying good methodological practice for modeling frailty of trial participants across trials, modeling heterogeneity of intervention effects, choosing appropriate association measures, dealing with (trial differences in) censoring and follow-up times, and addressing time-varying intervention effects and effect modification (interactions).We discuss how to achieve this using parametric and semi-parametric methods, and describe how to implement these in a one-stage or two-stage IPD-MA framework. We recommend exploring heterogeneity of the effect(s) through interaction and non-linear effects. Random effects should be applied to account for residual heterogeneity of the intervention effect. We provide further recommendations, many of which specific to IPD-MA of time-to-event data from randomized trials examining an intervention effect.We illustrate several key methods in a real IPD-MA, where IPD of 1225 participants from 5 randomized clinical trials were combined to compare the effects of Carbamazepine and Valproate on the incidence of epileptic seizures.

Individual participant data (IPD) meta-analysis of psychological relapse prevention interventions versus control for patients in remission from depression: a protocol

INTRODUCTION

Psychological interventions and antidepressant medication can be effective interventions to prevent depressive relapse for patients currently in remission of depression. Less is known about overall factors that predict or moderate treatment response for patients receiving a psychological intervention for recurrent depression. This is a protocol for an individual participant data (IPD) meta-analysis which aims to assess predictors and moderators of relapse or recurrence for patients currently in remission from depression.

METHODS AND ANALYSIS

Searches of PubMed, PsycINFO, Embase and Cochrane Central Register of Controlled Trials were completed on 13 October 2019. Study extractions and risk of bias assessments have been completed. Study authors will be asked to contribute IPD. Standard aggregate meta-analysis and IPD analysis will be conducted, and the outcomes will be compared with assess whether results differ between studies supplying data and those that did not. IPD files of individual data will be merged and variables homogenised where possible for consistency. IPD will be analysed via Cox regression and one and two-stage analyses will be conducted.

ETHICS AND DISSEMINATION

The results will be published in peer review journals and shared in a policy briefing as well as accessible formats and shared with a range of stakeholders. The results will inform patients and clinicians and researchers about our current understanding of more personalised ways to prevent a depressive relapse. No local ethics approval was necessary following consultation with the legal department. Guidance on patient data storage and management will be adhered to.

PROSPERO REGISTRATION NUMBER

CRD42019127844.

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.

Assessing the impacts of signal coordination on the crash risks of various driving cohorts

INTRODUCTION

Signal coordination has been wildly implemented on urban arterials to improve traffic efficiency. The impacts of signal coordination on traffic safety, however, are largely overlooked, particularly on crash propensities of driver-vehicle cohorts, which will vary due to changing traffic flow patterns.

METHOD

The paper aims to compare crash risks of various driving cohorts (measured by relative crash involvement ratio) on arterials with and without signal coordination with quasi-induced exposure technique, which has been well developed in estimating crash risks for driver-vehicle characteristics (i.e., driver age, gender, and vehicle type). Michigan traffic crash data (2000-2014) are retrieved for the case study.

RESULTS

The results indicate that: (a) when signal coordination is implemented, young, male drivers, and pickups are associated with more crash responsibilities; (b) crash propensities vary for different disaggregated situations, e.g., young drivers may experience the rapid increase in crash risks during the peak hours; and (c) more hazardous actions (e.g., failing to stop in assured clear distance) are witnessed for the high-risk driving cohorts on the coordinated arterials than non-coordinated ones. Conclusions and practical applications: The findings highlight the importance of safety impact analysis of signal coordination, and serve to guide the potential improvements of safety operation and management of signal coordinated arterials.

Visual memory tests enhance the identification of amnestic MCI cases at greater risk of Alzheimer's disease

OBJECTIVES

To investigate whether amnestic mild cognitive impairment (aMCI) identified with visual memory tests conveys an increased risk of Alzheimer's disease (risk-AD) and if the risk-AD differs from that associated with aMCI based on verbal memory tests.

PARTICIPANTS

4,771 participants aged 70.76 (SD = 6.74, 45.4% females) from five community-based studies, each a member of the international COSMIC consortium and from a different country, were classified as having normal cognition (NC) or one of visual, verbal, or combined (visual and verbal) aMCI using international criteria and followed for an average of 2.48 years. Hazard ratios (HR) and individual patient data (IPD) meta-analysis analyzed the risk-AD with age, sex, education, single/multiple domain aMCI, and Mini-Mental State Examination (MMSE) scores as covariates.

RESULTS

All aMCI groups (n = 760) had a greater risk-AD than NC (n = 4,011; HR range = 3.66 - 9.25). The risk-AD was not different between visual (n = 208, 17 converters) and verbal aMCI (n = 449, 29 converters, HR = 1.70, 95%CI: 0.88, 3.27, p = 0.111). Combined aMCI (n = 103, 12 converters, HR = 2.34, 95%CI: 1.13, 4.84, p = 0.023) had a higher risk-AD than verbal aMCI. Age and MMSE scores were related to the risk-AD. The IPD meta-analyses replicated these results, though with slightly lower HR estimates (HR range = 3.68, 7.43) for aMCI vs. NC.

CONCLUSIONS

Although verbal aMCI was most common, a significant proportion of participants had visual-only or combined visual and verbal aMCI. Compared with verbal aMCI, the risk-AD was the same for visual aMCI and higher for combined aMCI. Our results highlight the importance of including both verbal and visual memory tests in neuropsychological assessments to more reliably identify aMCI.

Efficacy of Mindfulness-Based Cognitive Therapy in Prevention of Depressive Relapse: An Individual Patient Data Meta-analysis From Randomized Trials

IMPORTANCE

Relapse prevention in recurrent depression is a significant public health problem, and antidepressants are the current first-line treatment approach. Identifying an equally efficacious nonpharmacological intervention would be an important development.

OBJECTIVE

To conduct a meta-analysis on individual patient data to examine the efficacy of mindfulness-based cognitive therapy (MBCT) compared with usual care and other active treatments, including antidepressants, in treating those with recurrent depression.

DATA SOURCES

English-language studies published or accepted for publication in peer-reviewed journals identified from EMBASE, PubMed/Medline, PsycINFO, Web of Science, Scopus, and the Cochrane Controlled Trials Register from the first available year to November 22, 2014. Searches were conducted from November 2010 to November 2014.

STUDY SELECTION

Randomized trials of manualized MBCT for relapse prevention in recurrent depression in full or partial remission that compared MBCT with at least 1 non-MBCT treatment, including usual care.

DATA EXTRACTION AND SYNTHESIS

This was an update to a previous meta-analysis. We screened 2555 new records after removing duplicates. Abstracts were screened for full-text extraction (S.S.) and checked by another researcher (T.D.). There were no disagreements. Of the original 2555 studies, 766 were evaluated against full study inclusion criteria, and we acquired full text for 8. Of these, 4 studies were excluded, and the remaining 4 were combined with the 6 studies identified from the previous meta-analysis, yielding 10 studies for qualitative synthesis. Full patient data were not available for 1 of these studies, resulting in 9 studies with individual patient data, which were included in the quantitative synthesis.

RESULTS

Of the 1258 patients included, the mean (SD) age was 47.1 (11.9) years, and 944 (75.0%) were female. A 2-stage random effects approach showed that patients receiving MBCT had a reduced risk of depressive relapse within a 60-week follow-up period compared with those who did not receive MBCT (hazard ratio, 0.69; 95% CI, 0.58-0.82). Furthermore, comparisons with active treatments suggest a reduced risk of depressive relapse within a 60-week follow-up period (hazard ratio, 0.79; 95% CI, 0.64-0.97). Using a 1-stage approach, sociodemographic (ie, age, sex, education, and relationship status) and psychiatric (ie, age at onset and number of previous episodes of depression) variables showed no statistically significant interaction with MBCT treatment. However, there was some evidence to suggest that a greater severity of depressive symptoms prior to treatment was associated with a larger effect of MBCT compared with other treatments.

CONCLUSIONS AND RELEVANCE

Mindfulness-based cognitive therapy appears efficacious as a treatment for relapse prevention for those with recurrent depression, particularly those with more pronounced residual symptoms. Recommendations are made concerning how future trials can address remaining uncertainties and improve the rigor of the field.

Using simulation studies to evaluate statistical methods

Simulation studies are computer experiments that involve creating data by pseudo-random sampling. A key strength of simulation studies is the ability to understand the behavior of statistical methods because some "truth" (usually some parameter/s of interest) is known from the process of generating the data. This allows us to consider properties of methods, such as bias. While widely used, simulation studies are often poorly designed, analyzed, and reported. This tutorial outlines the rationale for using simulation studies and offers guidance for design, execution, analysis, reporting, and presentation. In particular, this tutorial provides a structured approach for planning and reporting simulation studies, which involves defining aims, data-generating mechanisms, estimands, methods, and performance measures ("ADEMP"); coherent terminology for simulation studies; guidance on coding simulation studies; a critical discussion of key performance measures and their estimation; guidance on structuring tabular and graphical presentation of results; and new graphical presentations. With a view to describing recent practice, we review 100 articles taken from Volume 34 of Statistics in Medicine, which included at least one simulation study and identify areas for improvement.

Fractional flow reserve-guided percutaneous coronary intervention vs. medical therapy for patients with stable coronary lesions: meta-analysis of individual patient data

Aims

To assess the effect of fractional flow reserve (FFR)-guided percutaneous coronary intervention (PCI) with contemporary drug-eluting stents on the composite of cardiac death or myocardial infarction (MI) vs. medical therapy in patients with stable coronary lesions.

Methods and results

We performed a systematic review and meta-analysis of individual patient data (IPD) of the three available randomized trials of contemporary FFR-guided PCI vs. medical therapy for patients with stable coronary lesions: FAME 2 (NCT01132495), DANAMI-3-PRIMULTI (NCT01960933), and Compare-Acute (NCT01399736). FAME 2 enrolled patients with stable coronary artery disease (CAD), while the other two focused on non-culprit lesions in stabilized patients after acute coronary syndrome. A total of 2400 subjects were recruited from 54 sites world-wide with 1056 randomly assigned to FFR-guided PCI and 1344 to medical therapy. The pre-specified primary outcome was a composite of cardiac death or MI. We included data from extended follow-ups for FAME 2 (up to 5.5 years follow-up) and DANAMI-3-PRIMULTI (up to 4.7 years follow-up). After a median follow-up of 35 months (interquartile range 12-60 months), a reduction in the composite of cardiac death or MI was observed with FFR-guided PCI as compared with medical therapy (hazard ratio 0.72, 95% confidence interval 0.54-0.96; P = 0.02). The difference between groups was driven by MI.

Conclusion

In this IPD meta-analysis of the three available randomized controlled trials to date, FFR-guided PCI resulted in a reduction of the composite of cardiac death or MI compared with medical therapy, which was driven by a decreased risk of MI.

Prognostic role of preoperative platelet, fibrinogen, and D-dimer levels in patients with non-small cell lung cancer: A multicenter prospective study

Background

The relationships between coagulation factors and non‐small cell lung cancer (NSCLC) prognosis have been intensively studied. However, no previous study has investigated the combined effects of preoperative platelet (PLT), fibrinogen (FIB), and D‐dimer (D‐D) levels on the prognosis of NSCLC.

Methods

A multicenter prospective study was conducted over seven hospitals. A total of 395 patients diagnosed with operable NSCLC for the first time were included and followed‐up until disease progression or the end of the study. Baseline demographic and clinicopathological information, and preoperative coagulation test results were collected for each patient. Univariate and multilevel survival analyses were conducted using Cox regression and shared frailty models.

Results

Multilevel analyses revealed that there was a marginally significant association between elevated PLT level (> 215 × 10⁹/L) and unfavorable progression‐free survival (PFS) (hazard ratio 2.42, P = 0.05), whereas preoperative FIB and D‐D were not significant prognostic factors for PFS (P = 0.31 and 0.30, respectively). Compared to patients with one elevation of the three coagulation factors, patients with at least two elevations of the three factors had a significantly higher risk of cancer progression (hazard ratio 4.62, P = 0.02).

Conclusion

The number of elevated preoperative coagulation factors may have a significant effect on PFS and could be used to predict the prognosis of NSCLC patients after surgery. Future studies are warranted to further investigate the interactions between these three coagulation factors.

PROBAST: A Tool to Assess Risk of Bias and Applicability of Prediction Model Studies: Explanation and Elaboration

Prediction models in health care use predictors to estimate for an individual the probability that a condition or disease is already present (diagnostic model) or will occur in the future (prognostic model). Publications on prediction models have become more common in recent years, and competing prediction models frequently exist for the same outcome or target population. Health care providers, guideline developers, and policymakers are often unsure which model to use or recommend, and in which persons or settings. Hence, systematic reviews of these studies are increasingly demanded, required, and performed. A key part of a systematic review of prediction models is examination of risk of bias and applicability to the intended population and setting. To help reviewers with this process, the authors developed PROBAST (Prediction model Risk Of Bias ASsessment Tool) for studies developing, validating, or updating (for example, extending) prediction models, both diagnostic and prognostic. PROBAST was developed through a consensus process involving a group of experts in the field. It includes 20 signaling questions across 4 domains (participants, predictors, outcome, and analysis). This explanation and elaboration document describes the rationale for including each domain and signaling question and guides researchers, reviewers, readers, and guideline developers in how to use them to assess risk of bias and applicability concerns. All concepts are illustrated with published examples across different topics. The latest version of the PROBAST checklist, accompanying documents, and filled-in examples can be downloaded from www.probast.org.

Ten-year survival of immediate-loading implants in fully edentulous mandibles in the Japanese population: a multilevel analysis

PURPOSE

To evaluate the long-term clinical results of and risk factors for immediate-loading implant treatment of completely edentulous mandibles.

METHODS

We retrospectively studied 220 implants in 52 patients who received immediate-loading implants in fully edentulous mandibles. Kaplan-Meier survival analyses, log-rank tests, and multilevel mixed-effects parametric survival analysis was used for statistical analyses.

RESULTS

Thirteen of implants in seven patients failed, and the 10-year cumulative implant survival rate was 93.9 % and significantly (p=0.049) higher in women than in men. None of the predictor variables were significantly associated with implant survival, although sex tended to be associated with implant survival.

CONCLUSIONS

Immediate-loading implant treatment for completely edentulous mandibles had acceptable clinical results in the long term. Although we could not identify significant risk factors, we established a multilevel mixed-effects parametric survival analysis with the immediate-loading implant survival data.

Spatially varying effects of predictors for the survival prediction of nonmetastatic colorectal Cancer

Background

An increasing number of studies have identified spatial differences in colorectal cancer survival. However, little is known about the spatially varying effects of predictors in survival prediction modeling studies of colorectal cancer that have focused on estimating the absolute survival risk for patients from a wide range of populations. This study aimed to demonstrate the spatially varying effects of predictors of survival for nonmetastatic colorectal cancer patients.

Methods

Patients diagnosed with nonmetastatic colorectal cancer from 2004 to 2013 who were followed up through the end of 2013 were extracted from the Surveillance Epidemiology End Results registry (Patients: 128061). The log-rank test and the restricted mean survival time were used to evaluate survival outcome differences among spatial clusters corresponding to a widely used clinical predictor: stage determined by AJCC 7th edition staging system. The heterogeneity test, which is used in meta-analyses, revealed the spatially varying effects of single predictors. Then, considering the above predictors in a standard survival prediction model based on spatially clustered data, the spatially varying coefficients of these models revealed that some covariate effects may not be constant across the geographic regions of the study. Then, two types of survival prediction models (a statistical model and a machine learning model) were built; these models considered the predictors and enabled survival prediction for patients from a wide range of geographic regions.

Results

Based on univariate and multivariate analysis, some prognostic factors, such as “TNM stage”, “tumor size” and “age at diagnosis,” have significant spatially varying effects among different regions. When considering these spatially varying effects, machine learning models have fewer assumption constraints (such as proportional hazard assumptions) and better predictive performance compared with statistical models. Upon comparing the concordance indexes of these two models, the machine learning model was found to be more accurate (0.898[0.895,0.902]) than the statistical model (0.732 [0.726, 0.738]).

Conclusions

Based on this study, it’s recommended that the spatially varying effect of predictors should be considered when building survival prediction models involving large-scale and multicenter research data. Machine learning models that are not limited by the requirement of a statistical hypothesis are promising alternative models.

Electronic supplementary material

The online version of this article (10.1186/s12885-018-4985-2) contains supplementary material, which is available to authorized users.

Bayesian one-step IPD network meta-analysis of time-to-event data using Royston-Parmar models

Network meta‐analysis (NMA) combines direct and indirect evidence from trials to calculate and rank treatment estimates. While modelling approaches for continuous and binary outcomes are relatively well developed, less work has been done with time‐to‐event outcomes. Such outcomes are usually analysed using Cox proportional hazard (PH) models. However, in oncology with longer follow‐up time, and time‐dependent effects of targeted treatments, this may no longer be appropriate. Network meta‐analysis conducted in the Bayesian setting has been increasing in popularity. However, fitting the Cox model is computationally intensive, making it unsuitable for many datasets. Royston‐Parmar models are a flexible alternative that can accommodate time‐dependent effects. Motivated by individual participant data (IPD) from 37 cervical cancer trials (5922 women) comparing surgery, radiotherapy, and chemotherapy, this paper develops an IPD Royston‐Parmar Bayesian NMA model for overall survival. We give WinBUGS code for the model. We show how including a treatment‐ln(time) interaction can be used to conduct a global test for PH, illustrate how to test for consistency of direct and indirect evidence, and assess within‐design heterogeneity. Our approach provides a computationally practical, flexible Bayesian approach to NMA of IPD survival data, which readily extends to include additional complexities, such as non‐PH, increasingly found in oncology trials.

Associations between interarm differences in blood pressure and cardiovascular disease outcomes: protocol for an individual patient data meta-analysis and development of a prognostic algorithm

INTRODUCTION

Individual cohort studies in various populations and study-level meta-analyses have shown interarm differences (IAD) in blood pressure to be associated with increased cardiovascular and all-cause mortality. However, key questions remain, such as follows: (1) What is the additional contribution of IAD to prognostic risk estimation for cardiovascular and all-cause mortality? (2) What is the minimum cut-off value for IAD that defines elevated risk? (3) Is there a prognostic value of IAD and do different methods of IAD measurement impact on the prognostic value of IAD? We aim to address these questions by conducting an individual patient data (IPD) meta-analysis.

METHODS AND ANALYSIS

This study will identify prospective cohort studies that measured blood pressure in both arms during recruitment, and invite authors to contribute IPD datasets to this collaboration. All patient data received will be combined into a single dataset. Using one-stage meta-analysis, we will undertake multivariable time-to-event regression modelling, with the aim of developing a new prognostic model for cardiovascular risk estimation that includes IAD. We will explore variations in risk contribution of IAD across predefined population subgroups (eg, hypertensives, diabetics), establish the lower limit of IAD that is associated with additional cardiovascular risk and assess the impact of different methods of IAD measurement on risk prediction.

ETHICS AND DISSEMINATION

This study will not include any patient identifiable data. Included datasets will already have ethical approval and consent from their sponsors. Findings will be presented to international conferences and published in peer reviewed journals, and we have a comprehensive dissemination strategy in place with integrated patient and public involvement.

PROSPERO REGISTRATION NUMBER

CRD42015031227.

A Tutorial on Multilevel Survival Analysis: Methods, Models and Applications

Data that have a multilevel structure occur frequently across a range of disciplines, including epidemiology, health services research, public health, education and sociology. We describe three families of regression models for the analysis of multilevel survival data. First, Cox proportional hazards models with mixed effects incorporate cluster-specific random effects that modify the baseline hazard function. Second, piecewise exponential survival models partition the duration of follow-up into mutually exclusive intervals and fit a model that assumes that the hazard function is constant within each interval. This is equivalent to a Poisson regression model that incorporates the duration of exposure within each interval. By incorporating cluster-specific random effects, generalised linear mixed models can be used to analyse these data. Third, after partitioning the duration of follow-up into mutually exclusive intervals, one can use discrete time survival models that use a complementary log–log generalised linear model to model the occurrence of the outcome of interest within each interval. Random effects can be incorporated to account for within-cluster homogeneity in outcomes. We illustrate the application of these methods using data consisting of patients hospitalised with a heart attack. We illustrate the application of these methods using three statistical programming languages (R, SAS and Stata).

Is the intake of selective serotonin reuptake inhibitors associated with an increased risk of dental implant failure?

The aim of this retrospective study was to investigate the association between the intake of selective serotonin reuptake inhibitors (SSRIs) and the risk of dental implant failure. Patients were included if they were taking SSRIs only and no other medication, did not present any other systemic condition or compromising habits (bruxism, smoking, snuff), and complied with the use of prophylactic antibiotics for implant surgery. The multivariate generalized estimating equation (GEE) method and multilevel mixed-effects parametric survival analysis were used to test the association between SSRI exposure (predictor variable) and the risk of implant failure (outcome variable), adjusting for several potential confounders (other variables). The total number of implants with information available and meeting the necessary eligibility criteria was 931 (35 failures). These were placed in 300 patients. The implant failure rate was 12.5% for SSRI users and 3.3% for non-users (P=0.007). Kaplan-Meier analysis showed a statistically significant difference in the cumulative survival rate (P<0.001). The multivariate GEE model did not show a statistically significant association between SSRI intake and implant failure (P=0.530), nor did the multilevel model (P=0.125). It is suggested that the intake of SSRIs may not be associated with an increased risk of dental implant failure.

One-stage individual participant data meta-analysis models: estimation of treatment-covariate interactions must avoid ecological bias by separating out within-trial and across-trial information

Stratified medicine utilizes individual‐level covariates that are associated with a differential treatment effect, also known as treatment‐covariate interactions. When multiple trials are available, meta‐analysis is used to help detect true treatment‐covariate interactions by combining their data. Meta‐regression of trial‐level information is prone to low power and ecological bias, and therefore, individual participant data (IPD) meta‐analyses are preferable to examine interactions utilizing individual‐level information. However, one‐stage IPD models are often wrongly specified, such that interactions are based on amalgamating within‐ and across‐trial information. We compare, through simulations and an applied example, fixed‐effect and random‐effects models for a one‐stage IPD meta‐analysis of time‐to‐event data where the goal is to estimate a treatment‐covariate interaction. We show that it is crucial to centre patient‐level covariates by their mean value in each trial, in order to separate out within‐trial and across‐trial information. Otherwise, bias and coverage of interaction estimates may be adversely affected, leading to potentially erroneous conclusions driven by ecological bias. We revisit an IPD meta‐analysis of five epilepsy trials and examine age as a treatment effect modifier. The interaction is −0.011 (95% CI: −0.019 to −0.003; p = 0.004), and thus highly significant, when amalgamating within‐trial and across‐trial information. However, when separating within‐trial from across‐trial information, the interaction is −0.007 (95% CI: −0.019 to 0.005; p = 0.22), and thus its magnitude and statistical significance are greatly reduced. We recommend that meta‐analysts should only use within‐trial information to examine individual predictors of treatment effect and that one‐stage IPD models should separate within‐trial from across‐trial information to avoid ecological bias. © 2016 The Authors. Statistics in Medicine published by John Wiley & Sons Ltd.

Analyzing recurrent events when the history of previous episodes is unknown or not taken into account: proceed with caution

OBJECTIVE

Researchers in public health are often interested in examining the effect of several exposures on the incidence of a recurrent event. The aim of the present study is to assess how well the common-baseline hazard models perform to estimate the effect of multiple exposures on the hazard of presenting an episode of a recurrent event, in presence of event dependence and when the history of prior-episodes is unknown or is not taken into account.

METHODS

Through a comprehensive simulation study, using specific-baseline hazard models as the reference, we evaluate the performance of common-baseline hazard models by means of several criteria: bias, mean squared error, coverage, confidence intervals mean length and compliance with the assumption of proportional hazards.

RESULTS

Results indicate that the bias worsen as event dependence increases, leading to a considerable overestimation of the exposure effect; coverage levels and compliance with the proportional hazards assumption are low or extremely low, worsening with increasing event dependence, effects to be estimated, and sample sizes.

CONCLUSIONS

Common-baseline hazard models cannot be recommended when we analyse recurrent events in the presence of event dependence. It is important to have access to the history of prior-episodes per subject, it can permit to obtain better estimations of the effects of the exposures.

Short term versus long term dual antiplatelet therapy after implantation of drug eluting stent in patients with or without diabetes: systematic review and meta-analysis of individual participant data from randomised trials

OBJECTIVE

 To compare clinical outcomes between short term (up to 6 months) and long term (12 months) dual antiplatelet therapy (DAPT) after placement of a drug eluting stent in patients with and without diabetes.

DESIGN

 Individual participant data meta-analysis. Cox proportional regression models stratified by trial were used to assess the impact of diabetes on outcomes.

DATA SOURCE

 Medline, Embase, and Cochrane databases and proceedings of international meetings searched for randomised controlled trials comparing durations of DAPT after placement of a drug eluting stent. Individual patient data pooled from six DAPT trials.

PRIMARY OUTCOME

 Primary study outcome was one year risk of major adverse cardiac events (MACE), defined as cardiac death, myocardial infarction, or definite/probable stent thrombosis. All analyses were conducted by intention to treat.

RESULTS

 Six trials including 11 473 randomised patients were pooled. Of these patients, 3681 (32.1%) had diabetes and 7708 (67.2%) did not (mean age 63.7 (SD 9.9) and 62.8 (SD 10.1), respectively), and in 84 (0.7%) the information was missing. Diabetes was an independent predictor of MACE (hazard ratio 2.30, 95% confidence interval 1.01 to 5.27; P=0.048 At one year follow-up, long term DAPT was not associated with a decreased risk of MACE compared with short term DAPT in patients with (1.05, 0.62 to 1.76; P=0.86) or without (0.97, 0.67 to 1.39; P=0.85) diabetes (P=0.33 for interaction). The risk of myocardial infarction did not differ between the two DAPT regimens (0.95, 0.58 to 1.54; P=0.82; for those with diabetes and 1.15, 0.68 to 1.94; P=0.60; for those without diabetes (P=0.84 for interaction). There was a lower risk of definite/probable stent thrombosis with long term DAPT among patients with (0.26, 0.09 to 0.80; P=0.02) than without (1.42, 0.68 to 2.98; P=0.35) diabetes, with positive interaction testing (P=0.04 for interaction), although the landmark analysis showed a trend towards benefit in both groups. Long term DAPT was associated with higher rates of major or minor bleeding, irrespective of diabetes (P=0.37 for interaction).

CONCLUSIONS

 Although the presence of diabetes emerged as an independent predictor of MACE after implantation of a drug eluting stent, compared with short term DAPT, long term DAPT did not reduce the risk of MACE but increased the risk of bleeding among patients with stents with and without diabetes.

Meta-analysis using individual participant data: one-stage and two-stage approaches, and why they may differ

Meta‐analysis using individual participant data (IPD) obtains and synthesises the raw, participant‐level data from a set of relevant studies. The IPD approach is becoming an increasingly popular tool as an alternative to traditional aggregate data meta‐analysis, especially as it avoids reliance on published results and provides an opportunity to investigate individual‐level interactions, such as treatment‐effect modifiers. There are two statistical approaches for conducting an IPD meta‐analysis: one‐stage and two‐stage. The one‐stage approach analyses the IPD from all studies simultaneously, for example, in a hierarchical regression model with random effects. The two‐stage approach derives aggregate data (such as effect estimates) in each study separately and then combines these in a traditional meta‐analysis model. There have been numerous comparisons of the one‐stage and two‐stage approaches via theoretical consideration, simulation and empirical examples, yet there remains confusion regarding when each approach should be adopted, and indeed why they may differ.

In this tutorial paper, we outline the key statistical methods for one‐stage and two‐stage IPD meta‐analyses, and provide 10 key reasons why they may produce different summary results. We explain that most differences arise because of different modelling assumptions, rather than the choice of one‐stage or two‐stage itself. We illustrate the concepts with recently published IPD meta‐analyses, summarise key statistical software and provide recommendations for future IPD meta‐analyses. © 2016 The Authors. Statistics in Medicine published by John Wiley & Sons Ltd.

Bruxism and dental implant failures: a multilevel mixed effects parametric survival analysis approach

Recent studies have suggested that the insertion of dental implants in patients being diagnosed with bruxism negatively affected the implant failure rates. The aim of the present study was to investigate the association between the bruxism and the risk of dental implant failure. This retrospective study is based on 2670 patients who received 10 096 implants at one specialist clinic. Implant- and patient-related data were collected. Descriptive statistics were used to describe the patients and implants. Multilevel mixed effects parametric survival analysis was used to test the association between bruxism and risk of implant failure adjusting for several potential confounders. Criteria from a recent international consensus (Lobbezoo et al., J Oral Rehabil, 40, 2013, 2) and from the International Classification of Sleep Disorders (International classification of sleep disorders, revised: diagnostic and coding manual, American Academy of Sleep Medicine, Chicago, 2014) were used to define and diagnose the condition. The number of implants with information available for all variables totalled 3549, placed in 994 patients, with 179 implants reported as failures. The implant failure rates were 13·0% (24/185) for bruxers and 4·6% (155/3364) for non-bruxers (P < 0·001). The statistical model showed that bruxism was a statistically significantly risk factor to implant failure (HR 3·396; 95% CI 1·314, 8·777; P = 0·012), as well as implant length, implant diameter, implant surface, bone quantity D in relation to quantity A, bone quality 4 in relation to quality 1 (Lekholm and Zarb classification), smoking and the intake of proton pump inhibitors. It is suggested that the bruxism may be associated with an increased risk of dental implant failure.

A multilevel excess hazard model to estimate net survival on hierarchical data allowing for non-linear and non-proportional effects of covariates

The excess hazard regression model is an approach developed for the analysis of cancer registry data to estimate net survival, that is, the survival of cancer patients that would be observed if cancer was the only cause of death. Cancer registry data typically possess a hierarchical structure: individuals from the same geographical unit share common characteristics such as proximity to a large hospital that may influence access to and quality of health care, so that their survival times might be correlated. As a consequence, correct statistical inference regarding the estimation of net survival and the effect of covariates should take this hierarchical structure into account. It becomes particularly important as many studies in cancer epidemiology aim at studying the effect on the excess mortality hazard of variables, such as deprivation indexes, often available only at the ecological level rather than at the individual level. We developed here an approach to fit a flexible excess hazard model including a random effect to describe the unobserved heterogeneity existing between different clusters of individuals, and with the possibility to estimate non-linear and time-dependent effects of covariates. We demonstrated the overall good performance of the proposed approach in a simulation study that assessed the impact on parameter estimates of the number of clusters, their size and their level of unbalance. We then used this multilevel model to describe the effect of a deprivation index defined at the geographical level on the excess mortality hazard of patients diagnosed with cancer of the oral cavity. Copyright © 2016 John Wiley & Sons, Ltd.

flexsurv: A Platform for Parametric Survival Modeling in R

flexsurv is an R package for fully-parametric modeling of survival data. Any parametric time-to-event distribution may be fitted if the user supplies a probability density or hazard function, and ideally also their cumulative versions. Standard survival distributions are built in, including the three and four-parameter generalized gamma and F distributions. Any parameter of any distribution can be modeled as a linear or log-linear function of covariates. The package also includes the spline model of Royston and Parmar (2002), in which both baseline survival and covariate effects can be arbitrarily flexible parametric functions of time. The main model-fitting function, flexsurvreg, uses the familiar syntax of survreg from the standard survival package (Therneau 2016). Censoring or left-truncation are specified in 'Surv' objects. The models are fitted by maximizing the full log-likelihood, and estimates and confidence intervals for any function of the model parameters can be printed or plotted. flexsurv also provides functions for fitting and predicting from fully-parametric multi-state models, and connects with the mstate package (de Wreede, Fiocco, and Putter 2011). This article explains the methods and design principles of the package, giving several worked examples of its use.

flexsurv: A Platform for Parametric Survival Modeling in R

flexsurv is an R package for fully-parametric modeling of survival data. Any parametric time-to-event distribution may be fitted if the user supplies a probability density or hazard function, and ideally also their cumulative versions. Standard survival distributions are built in, including the three and four-parameter generalized gamma and F distributions. Any parameter of any distribution can be modeled as a linear or log-linear function of covariates. The package also includes the spline model of Royston and Parmar (2002), in which both baseline survival and covariate effects can be arbitrarily flexible parametric functions of time. The main model-fitting function, flexsurvreg, uses the familiar syntax of survreg from the standard survival package (Therneau 2016). Censoring or left-truncation are specified in 'Surv' objects. The models are fitted by maximizing the full log-likelihood, and estimates and confidence intervals for any function of the model parameters can be printed or plotted. flexsurv also provides functions for fitting and predicting from fully-parametric multi-state models, and connects with the mstate package (de Wreede, Fiocco, and Putter 2011). This article explains the methods and design principles of the package, giving several worked examples of its use.

Longitudinal measures of perfluoroalkyl substances (PFAS) in serum of Gullah African Americans in South Carolina: 2003-2013

BACKGROUND

Charleston Harbor has elevated concentrations of PFAS in dolphins, but local human exposure data are limited.

OBJECTIVES

We sought to describe PFAS serum concentrations' temporal trends among Gullah African American residents of coastal South Carolina.

METHODS

Longitudinal measures of PFAS in blood serum from a Gullah clinical sample, without lupus, were examined using spaghetti plots and visit-to-visit change scores (e.g., differences in concentrations between visits) among the 68 participants with repeated measures available. We also modeled population-level trends among the 71 participants with any data using proportionate percentile models, accounting for clustering through robust standard errors. In a post-hoc analysis we examined heterogeneity of temporal trends by age through mixed-effects models for the log-transformed PFAS compounds.

RESULTS

Population concentrations of PFOS dropped approximately 9 (95% CI: 8, 10) percent each year over 2003-2013. This was concordant with individual PFOS trajectories (median PFOS change score -21.7 ng/g wet weight, interquartile range of PFOS change scores: -32.8, -14.9) and reports for other populations over this time period. Several other compounds including PFOA, PFHxS, and PFuNDA also showed a population-level decrease. However, examination of individual trajectories suggested substantial heterogeneity. Post-hoc analyses indicated that PFAS trajectories were heterogeneous by age.

CONCLUSIONS

Many PFAS compounds are decreasing in a sample of Gullah African Americans from coastal South Carolina. There may be age differences in the elimination kinetics of PFASs. The possible role of age as a modifier of PFAS serum trends merits further research.