Volume 20, Issue 1

There is widespread interest in using deep learning to build prediction models for medical imaging data. These deep learning methods capture the local structure of the image and require no manual feature extraction. Despite the importance of modeling survival in the context of medical data analysis, research on deep learning methods for modeling the relationship of imaging and time-to-event data is still under-developed. We provide an overview of deep learning methods for time-to-event outcomes and compare several deep learning methods to Cox model based methods through the analysis of a histology dataset of gliomas.

In response to the COVID-19 pandemic caused by SARS-CoV-2, governments have adopted a wide range of non-pharmaceutical interventions (NPI). These include stringent measures such as strict lockdowns, closing schools, bars and restaurants, curfews, and barrier gestures such as mask-wearing and social distancing. Deciphering the effectiveness of each NPI is critical to responding to future waves and outbreaks. To this end, we first develop a dynamic model of the French COVID-19 epidemics over a one-year period. We rely on a global extended Susceptible-Infectious-Recovered (SIR) mechanistic model of infection that includes a dynamic transmission rate over time. Multilevel data across French regions are integrated using random effects on the parameters of the mechanistic model, boosting statistical power by multiplying integrated observation series. We estimate the parameters using a new population-based statistical approach based on a Kalman filter, used for the first time in analysing real-world data. We then fit the estimated time-varying transmission rate using a regression model that depends on the NPIs while accounting for vaccination coverage, the occurrence of variants of concern (VoC), and seasonal weather conditions. We show that all NPIs considered have an independent significant association with transmission rates. In addition, we show a strong association between weather conditions that reduces transmission in summer, and we also estimate increased transmissibility of VoC.

COVID-19 survival data presents a special situation where not only the time-to-event period is short, but also the two events or outcome types, death and release from hospital, are mutually exclusive, leading to two cause-specific hazard ratios (csHR d and csHR r ). The eventual mortality/release outcome is also analyzed by logistic regression to obtain odds-ratio (OR). We have the following three empirical observations: (1) The magnitude of OR is an upper limit of the csHR d : |log(OR)| ≥ |log(csHR d )|. This relationship between OR and HR might be understood from the definition of the two quantities; (2) csHR d and csHR r point in opposite directions: log(csHR d ) ⋅ log(csHR r ) < 0; This relation is a direct consequence of the nature of the two events; and (3) there is a tendency for a reciprocal relation between csHR d and csHR r : csHR d ∼ 1/csHR r . Though an approximate reciprocal trend between the two hazard ratios is in indication that the same factor causing faster death also lead to slow recovery by a similar mechanism, and vice versa, a quantitative relation between csHR d and csHR r in this context is not obvious. These results may help future analyses of data from COVID-19 or other similar diseases, in particular if the deceased patients are lacking, whereas surviving patients are abundant.

In clinical trials that are subject to noncompliance, the commonly used intention-to-treat estimand is valid as a causal effect of treatment assignment but is sensitive to the level of compliance. An alternative estimand, the complier average causal effect ( CACE ), measures the average effect of treatment received in the latent subset of subjects who would comply with either assigned treatment. Because the principal stratum of compliers can vary with the circumstances of the trial, CACE too depends on the compliance fraction. We propose a model in which an underlying latent proto-compliance interacts with trial characteristics to determine a subject’s compliance behavior. When the latent compliance is independent of the individual treatment effect, the average causal effect is constant across compliance classes, and CACE is robust across trials and equal to the population average causal effect . We demonstrate the potential degree of sensitivity of CACE in a simulation study, an analysis of data from a trial of vitamin A supplementation in children, and a meta-analysis of trials of epidural analgesia in labor.

With our increased ability to capture large data, causal inference has received renewed attention and is playing an ever-important role in biomedicine and economics. However, one major methodological hurdle is that existing methods rely on many unverifiable model assumptions. Thus robust modeling is a critically important approach complementary to sensitivity analysis, where it compares results under various model assumptions. The more robust a method is with respect to model assumptions, the more worthy it is. The doubly robust estimator (DRE) is a significant advance in this direction. However, in practice, many outcome measures are functionals of multiple distributions, and so are the associated estimands, which can only be estimated via U-statistics. Thus most existing DREs do not apply. This article proposes a broad class of highly robust U-statistic estimators (HREs), which use semiparametric specifications for both the propensity score and outcome models in constructing the U-statistic. Thus, the HRE is more robust than the existing DREs. We derive comprehensive asymptotic properties of the proposed estimators and perform extensive simulation studies to evaluate their finite sample performance and compare them with the corresponding parametric U-statistics and the naive estimators, which show significant advantages. Then we apply the method to analyze a clinical trial from the AIDS Clinical Trials Group.

A major focus of causal inference is the estimation of heterogeneous average treatment effects (HTE) – average treatment effects within strata of another variable of interest such as levels of a biomarker, education, or age strata. Inference involves estimating a stratum-specific regression and integrating it over the distribution of confounders in that stratum – which itself must be estimated. Standard practice involves estimating these stratum-specific confounder distributions independently (e.g. via the empirical distribution or Rubin’s Bayesian bootstrap), which becomes problematic for sparsely populated strata with few observed confounder vectors. In this paper, we develop a nonparametric hierarchical Bayesian bootstrap (HBB) prior over the stratum-specific confounder distributions for HTE estimation. The HBB partially pools the stratum-specific distributions, thereby allowing principled borrowing of confounder information across strata when sparsity is a concern. We show that posterior inference under the HBB can yield efficiency gains over standard marginalization approaches while avoiding strong parametric assumptions about the confounder distribution. We use our approach to estimate the adverse event risk of proton versus photon chemoradiotherapy across various cancer types.

Cancer classification and gene selection are important applications in DNA microarray gene expression data analysis. Since DNA microarray data suffers from the high-dimensionality problem, automatic gene selection methods are used to enhance the classification performance of expert classifier systems. In this paper, a new penalized logistic regression method that performs simultaneous gene coefficient estimation and variable selection in DNA microarray data is discussed. The method employs prior information about the gene coefficients to improve the classification accuracy of the underlying model. The coordinate descent algorithm with screening rules is given to obtain the gene coefficient estimates of the proposed method efficiently. The performance of the method is examined on five high-dimensional cancer classification datasets using the area under the curve, the number of selected genes, misclassification rate and F-score measures. The real data analysis results indicate that the proposed method achieves a good cancer classification performance with a small misclassification rate, large area under the curve and F-score by trading off some sparsity level of the underlying model. Hence, the proposed method can be seen as a reliable penalized logistic regression method in the scope of high-dimensional cancer classification.

Selection of relevant fixed and random effects without prior choices made from possibly insufficient theory is important in mixed models. Inference with current boosting techniques suffers from biased estimates of random effects and the inflexibility of random effects selection. This paper proposes a new inference method “BayesBoost” that integrates a Bayesian learner into gradient boosting with simultaneous estimation and selection of fixed and random effects in linear mixed models. The method introduces a novel selection strategy for random effects, which allows for computationally fast selection of random slopes even in high-dimensional data structures. Additionally, the new method not only overcomes the shortcomings of Bayesian inference in giving precise and unambiguous guidelines for the selection of covariates by benefiting from boosting techniques, but also provides Bayesian ways to construct estimators for the precision of parameters such as variance components or credible intervals, which are not available in conventional boosting frameworks. The effectiveness of the new approach can be observed via simulation and in a real-world application.

For non-inferiority/superiority and equivalence tests of two Poisson rates, the determination of the required number of sample sizes has been studied but the studies for the number of events to be observed are very limited. To fill the gap, the present study first is aimed toward determining the number of events to be observed for testing non-inferiority/superiority and equivalence of two Poisson rates, respectively. Also, considering the cost for each event, the second purpose is to apply an exhaustive search to find the unequal but optimal allocation of events for each group such that the budget is minimal for a user-specified power level, or the statistical power is maximal for a user-specified budget. Four R Shiny apps were developed to obtain the number of events needed for each group. A simulation study showed the proposed approach to be valid in terms of Type I error and statistical power. A comparison of the proposed approach with extant methods from various disciplines was performed, and an illustrative example of comparing the adverse reactions to the COVID-19 vaccines was demonstrated. By applying the proposed approach, researchers also can estimate the most economical number of subjects or time intervals after determining the number of events.

Common statistical approaches are not designed to deal with so-called “short fat data” in biomarker pilot studies, where the number of biomarker candidates exceeds the sample size by magnitudes. High-throughput technologies for omics data enable the measurement of ten thousands and more biomarker candidates for specific diseases or states of a disease. Due to the limited availability of study participants, ethical reasons and high costs for sample processing and analysis researchers often prefer to start with a small sample size pilot study in order to judge the potential of finding biomarkers that enable – usually in combination – a sufficiently reliable classification of the disease state under consideration. We developed a user-friendly tool, called HiPerMAb that allows to evaluate pilot studies based on performance measures like multiclass AUC, entropy, area above the cost curve, hypervolume under manifold, and misclassification rate using Monte-Carlo simulations to compute the p -values and confidence intervals. The number of “good” biomarker candidates is compared to the expected number of “good” biomarker candidates in a data set with no association to the considered disease states. This allows judging the potential in the pilot study even if statistical tests with correction for multiple testing fail to provide any hint of significance.

In recent years, meta-analysis has evolved to a critically important field of Statistics, and has significant applications in Medicine and Health Sciences. In this work we briefly present existing methodologies to conduct meta-analysis along with any discussion and recent developments accompanying them. Undoubtedly, studies brought together in a systematic review will differ in one way or another. This yields a considerable amount of variability, any kind of which may be termed heterogeneity. To this end, reports of meta-analyses commonly present a statistical test of heterogeneity when attempting to establish whether the included studies are indeed similar in terms of the reported output or not. We intend to provide an overview of the topic, discuss the potential sources of heterogeneity commonly met in the literature and provide useful guidelines on how to address this issue and to detect heterogeneity. Moreover, we review the recent developments in the Bayesian approach along with the various graphical tools and statistical software that are currently available to the analyst. In addition, we discuss sensitivity analysis issues and other approaches of understanding the causes of heterogeneity. Finally, we explore heterogeneity in meta-analysis for time to event data in a nutshell, pointing out its unique characteristics.

In this paper, a Markov Regime Switching Model of Conditional Mean with covariates, is proposed and investigated for the analysis of incidence rate data. The components of the model are selected by both penalized likelihood techniques in conjunction with the Expectation Maximization algorithm, with the goal of achieving a high level of robustness regarding the modeling of dynamic behaviors of epidemiological data. In addition to statistical inference, Changepoint Detection Analysis is performed for the selection of the number of regimes, which reduces the complexity associated with Likelihood Ratio Tests. Within this framework, a three-phase procedure for modeling incidence data is proposed and tested via real and simulated data.

Mediation analysis studies situations where an exposure may affect an outcome both directly and indirectly through intervening variables called mediators. It is frequently of interest to test for the effect of the exposure on the outcome, and the standard approach is simply to regress the latter on the former. However, it seems plausible that a more powerful test statistic could be achieved by also incorporating the mediators. This would be useful in cases where the exposure effect size might be small, which for example is common in genomics applications. Previous work has shown that this is indeed possible under complete mediation, where there is no direct effect. In most applications, however, the direct effect is likely nonzero. In this paper we study linear mediation models and find that under certain conditions, power gain is still possible under this incomplete mediation setting for testing the null hypothesis that there is neither a direct nor an indirect effect. We study a class of procedures that can achieve this performance and develop their application to both low- and high-dimensional mediators. We then illustrate their performances in simulations as well as in an analysis using DNA methylation mediators to study the effect of cigarette smoking on gene expression.

In case-control studies, odds ratios (OR) are calculated from 2 × 2 tables and in some instances, we observe small cell counts or zero counts in one of the cells. The corrections to calculate the ORs in the presence of empty cells are available in literature. Some of these include Yates continuity correction and Agresti and Coull correction. However, the available methods provided different corrections and the situations where each could be applied are not very apparent. Therefore, the current research proposes an iterative algorithm of estimating an exact (optimum) correction factor for the respective sample size. This was evaluated by simulating data with varying proportions and sample sizes. The estimated correction factor was considered after obtaining the bias, standard error of odds ratio, root mean square error and the coverage probability. Also, we have presented a linear function to identify the exact correction factor using sample size and proportion.

This paper considers a partially linear regression model relating a right-censored response variable to predictors and an extra covariate with measured error. The main problem here is that censorship and measurement error problems need to be solved to estimate the model correctly. In this sense, we propose three modified semiparametric estimators obtained from local polynomial regression, kernel smoothing, and B-spline smoothing methods based on kernel deconvolution approach and synthetic data transformation. Here, kernel deconvolution technique is used to solve the measurement error problem in the model and synthetic data transformation is considered to add the effect of censorship to the estimation procedure, which is a very common method in the literature. The performances of the introduced estimators are compared in the detailed Monte-Carlo simulation study. In addition, Carotid endarterectomy data is used as real-world data example and results are presented. According to the results, it is seen that the deconvoluted local polynomial method gives more qualified estimates than other two methods.

Differentiated care delivery aims to simplify care of people living with HIV, reflect their preferences, reduce burdens on the healthcare system, maintain care quality and preserve resources. However, assessing program effectiveness using observational data is difficult due to confounding by indication and randomized trials may be infeasible. Also, benefits can reach patients directly, through enrollment in the program, and indirectly, by increasing quality of and accessibility to care. Low-risk express care (LREC), the program under evaluation, is a nurse-centered model which assigns patients stable on ART to a nurse every two months and a clinician every third visit, reducing annual clinician visits by two thirds. Study population is comprised of 16,832 subjects from 15 clinics in Kenya. We focus on patient retention in care based on whether the LREC program is available at a clinic and whether the patient is enrolled in LREC. We use G-estimation to assess the effect on retention of two “strategies”: (i) program availability but no enrollment; (ii) enrollment at an available program; versus no program availability. Compared to no availability, LREC results in a non-significant increase in patient retention, among patients not enrolled in the program (indirect effect), while enrollment in LREC is associated with a significant extension of the time retained in care (direct effect). G-estimation provides an analytical framework useful to the assessment of similar programs using observational data.

Model-based component-wise gradient boosting is a popular tool for data-driven variable selection. In order to improve its prediction and selection qualities even further, several modifications of the original algorithm have been developed, that mainly focus on different stopping criteria, leaving the actual variable selection mechanism untouched. We investigate different prediction-based mechanisms for the variable selection step in model-based component-wise gradient boosting. These approaches include Akaikes Information Criterion (AIC) as well as a selection rule relying on the component-wise test error computed via cross-validation. We implemented the AIC and cross-validation routines for Generalized Linear Models and evaluated them regarding their variable selection properties and predictive performance. An extensive simulation study revealed improved selection properties whereas the prediction error could be lowered in a real world application with age-standardized COVID-19 incidence rates.