Modeling Epidemiology Data with Machine Learning Technique to Detect Risk Factors for Gastric Cancer

PURPOSE

Gastric cancer (GC) ranks as the 7th most common cancer worldwide and a leading cause of cancer mortality. In Iran, stomach malignancies are the most common fatal cancers with higher than world average incidence. In recent years, methods like machine learning that provide the opportunity of merging health issues with computational power and learning capacity have caught considerable attention for prediction and diagnosis of diseases. In this study, we aimed to model GC data to find risk factors and identify GC cases in Golestan Cohort Study (GCS), using gradient boosting as a machine learning technique.

METHODS

Since the GC class (280) was smaller than not-GC (49,467), "Synthetic Minority Oversampling Technique" was used to balance the dataset. Seventy percent of the data was used to train the gradient boosting algorithm and find effective factors on gastric cancer, and the remaining 30% was used for accuracy assessment.

RESULTS

Our results indicated that out of 19 factors, age, social economical status, tea temperature, body mass index, gender, and education were the top six effective factors with impact rates of 0.24, 0.16, 0.13, 0.13, and 0.07, respectively. The trained model classified 70 out of 72 GC patients in the test set, correctly.

CONCLUSION

The results indicate that this model can effectively detect gastric cancer (GC) by utilizing important risk factors, thus avoiding the need for invasive procedures. The model's performance is reliable when provided with an adequate amount of input data, and as the dataset expands, its accuracy and generalization improve significantly. Overall, the trained system's success stems from its ability to identify risk factors and identify cancer patients.

Sample size and power calculations for causal mediation analysis: A Tutorial and Shiny App

When designing a study for causal mediation analysis, it is crucial to conduct a power analysis to determine the sample size required to detect the causal mediation effects with sufficient power. However, the development of power analysis methods for causal mediation analysis has lagged far behind. To fill the knowledge gap, I proposed a simulation-based method and an easy-to-use web application ( https://xuqin.shinyapps.io/CausalMediationPowerAnalysis/ ) for power and sample size calculations for regression-based causal mediation analysis. By repeatedly drawing samples of a specific size from a population predefined with hypothesized models and parameter values, the method calculates the power to detect a causal mediation effect based on the proportion of the replications with a significant test result. The Monte Carlo confidence interval method is used for testing so that the sampling distributions of causal effect estimates are allowed to be asymmetric, and the power analysis runs faster than if the bootstrapping method is adopted. This also guarantees that the proposed power analysis tool is compatible with the widely used R package for causal mediation analysis, mediation, which is built upon the same estimation and inference method. In addition, users can determine the sample size required for achieving sufficient power based on power values calculated from a range of sample sizes. The method is applicable to a randomized or nonrandomized treatment, a mediator, and an outcome that can be either binary or continuous. I also provided sample size suggestions under various scenarios and a detailed guideline of app implementation to facilitate study designs.

Evidence for Shifting Cognitive Strategies when Icons Appear in Unexpected Locations

OBJECTIVE

The present study examines the cognitive effects of placing icons in unexpected spatial locations within websites.

BACKGROUND

Prior research has revealed evidence for cognitive conflict when web icons occur in unexpected locations (e.g., cart, top left), generally consistent with a dynamical systems models. Here, we compare the relative strength of evidence for both dual and dynamical systems models.

METHODS

Participants clicked on icons located in either expected (e.g., cart, top right) or unexpected (e.g., cart, top left) locations while mouse trajectories were continuously recorded. Trajectories were classified according to prototypes associated with each cognitive model. The dynamical systems model predicts curved trajectories, while the dual-systems model predicts straight and change of mind trajectories.

RESULTS

Trajectory classification revealed that curved trajectories increased (+11%), while straight and change of mind trajectories decreased (-12%) when target icons occurred in unexpected locations (p < .001).

CONCLUSION

Rather than employing a single cognitive strategy, users shift from a primarily dual-systems to dynamical systems strategy when icons occur in unexpected locations.

APPLICATION

Potential applications of this work include the assessment of cognitive impacts such as mental workload and cognitive conflict during real-time interaction with websites and other screen-based interfaces, personalization and adaptive interfaces based on an individual's cognitive strategy, and data-driven A/B testing of alternative interface designs.

Regularized Variational Estimation for Exploratory Item Factor Analysis

Item factor analysis (IFA), also known as Multidimensional Item Response Theory (MIRT), is a general framework for specifying the functional relationship between respondents' multiple latent traits and their responses to assessment items. The key element in MIRT is the relationship between the items and the latent traits, so-called item factor loading structure. The correct specification of this loading structure is crucial for accurate calibration of item parameters and recovery of individual latent traits. This paper proposes a regularized Gaussian Variational Expectation Maximization (GVEM) algorithm to efficiently infer item factor loading structure directly from data. The main idea is to impose an adaptive L 1 -type penalty to the variational lower bound of the likelihood to shrink certain loadings to 0. This new algorithm takes advantage of the computational efficiency of GVEM algorithm and is suitable for high-dimensional MIRT applications. Simulation studies show that the proposed method accurately recovers the loading structure and is computationally efficient. The new method is also illustrated using the National Education Longitudinal Study of 1988 (NELS:88) mathematics and science assessment data.

A critical assessment of sparse PCA (research): why (one should acknowledge that) weights are not loadings

Principal component analysis (PCA) is an important tool for analyzing large collections of variables. It functions both as a pre-processing tool to summarize many variables into components and as a method to reveal structure in data. Different coefficients play a central role in these two uses. One focuses on the weights when the goal is summarization, while one inspects the loadings if the goal is to reveal structure. It is well known that the solutions to the two approaches can be found by singular value decomposition; weights, loadings, and right singular vectors are mathematically equivalent. What is often overlooked, is that they are no longer equivalent in the setting of sparse PCA methods which induce zeros either in the weights or the loadings. The lack of awareness for this difference has led to questionable research practices in sparse PCA. First, in simulation studies data is generated mostly based only on structures with sparse singular vectors or sparse loadings, neglecting the structure with sparse weights. Second, reported results represent local optima as the iterative routines are often initiated with the right singular vectors. In this paper we critically re-assess sparse PCA methods by also including data generating schemes characterized by sparse weights and different initialization strategies. The results show that relying on commonly used data generating models can lead to over-optimistic conclusions. They also highlight the impact of choice between sparse weights versus sparse loadings methods and the initialization strategies. The practical consequences of this choice are illustrated with empirical datasets.

Testing the missing at random assumption in generalized linear models in the presence of instrumental variables

Practical problems with missing data are common, and many methods have been developed concerning the validity and/or efficiency of statistical procedures. On a central focus, there have been longstanding interests on the mechanism governing data missingness, and correctly deciding the appropriate mechanism is crucially relevant for conducting proper practical investigations. In this paper, we present a new hypothesis testing approach for deciding between the conventional notions of missing at random and missing not at random in generalized linear models in the presence of instrumental variables. The foundational idea is to develop appropriate discrepancy measures between estimators whose properties significantly differ only when missing at random does not hold. We show that our testing approach achieves an objective data-oriented choice between missing at random or not. We demonstrate the feasibility, validity, and efficacy of the new test by theoretical analysis, simulation studies, and a real data analysis.

Using a centered general linear model for detection of interactions among biomarkers

The dummy variable based general linear model (gLM) is commonly used to model categorical factors and their interactions. However, the main factors and their interactions in a general linear model are often correlated even when the factors are independently distributed. Alternatively, the classical two-way factorial analysis of variance (ANOVA) model can avoid the correlation between the main factors and their interactions when the main factors are independent. But the ANOVA model is hardly applicable to a regular linear regression model especially in the presence of other covariates due to constraints on its model parameters. In this study, a centered general linear model (cgLM) is proposed for modeling interactions between categorical factors based on their centered dummy variables. We show that the cgLM can avoid the correlation between the main factors and their interactions as the ANOVA model when the main factors are independent. Meanwhile, similar to gLM, it can be used in regular regression and fitted conveniently using the standard least square approach by choosing appropriate baselines to avoid constraints on its model parameters. The potential advantage of cgLM over gLM for detection of interactions in model building procedures is also illustrated and compared via a simulation study. Finally, the cgLM is applied to a postmortem brain gene expression data set.

Propensity score stratified MAP prior and posterior inference for incorporating information across multiple potentially heterogeneous data sources

Incorporation of external information is becoming increasingly common when designing clinical trials. Availability of multiple sources of information has inspired the development of methodologies that account for potential heterogeneity not only between the prospective trial and the pooled external data sources but also between the different external data sources themselves. Our approach proposes an intuitive way of handling such a scenario for the continuous outcomes setting by using propensity score-based stratification and then utilizing robust meta-analytic predictive priors for each stratum to incorporate the prior data to distinguish among different external data sources in each stratum. Through extensive simulations, our approach proves to be more efficient and less biased than the currently available methods. A real case study using clinical trials that study schizophrenia from multiple different sources is also included.

Non-inferiority trials with time-to-event data: clarifying the impact of censoring

In non-inferiority (NI) trials with time-to-event data, different types and patterns of censoring may occur, but their impact on trial results is not entirely clear. We investigated the influence of informative and non-informative censoring by conducting extensive simulation studies under the assumption that the NI margin is defined as a maximum acceptable hazard ratio and scenarios typically observed in recent NI trials. We found that while non-informative censoring tends to only affect the power, informative censoring can impact the treatment effect estimates, type I error rate, and power. The magnitude of these effects depends on the between-group differences in the failure and informative censoring risks, as well as the correlation between censoring and failure times, among other factors. The adverse impact of informative censoring was generally decreased with larger NI margins.

A mark-specific quantile regression model

Quantile regression has become a widely used tool for analysing competing risk data. However, quantile regression for competing risk data with a continuous mark is still scarce. The mark variable is an extension of cause of failure in a classical competing risk model where cause of failure is replaced by a continuous mark only observed at uncensored failure times. An example of the continuous mark variable is the genetic distance that measures dissimilarity between the infecting virus and the virus contained in the vaccine construct. In this article, we propose a novel mark-specific quantile regression model. The proposed estimation method borrows strength from data in a neighbourhood of a mark and is based on an induced smoothed estimation equation, which is very different from the existing methods for competing risk data with discrete causes. The asymptotic properties of the resulting estimators are established across mark and quantile continuums. In addition, a mark-specific quantile-type vaccine efficacy is proposed and its statistical inference procedures are developed. Simulation studies are conducted to evaluate the finite sample performances of the proposed estimation and hypothesis testing procedures. An application to the first HIV vaccine efficacy trial is provided.

Estimation of the selected treatment mean in two stage drop-the-losers design

A common problem faced in clinical studies is that of estimating the effect of the most effective (e.g. the one having the largest mean) treatment among k ( ≥ 2 ) available treatments. The most effective treatment is adjudged based on numerical values of some statistic corresponding to the k treatments. A proper design for such problems is the so-called "Drop-the-Losers Design (DLD)". We consider two treatments whose effects are described by independent Gaussian distributions having different unknown means and a common known variance. To select the more effective treatment, the two treatments are independently administered to n 1 subjects each and the treatment corresponding to the larger sample mean is selected. To study the effect of the adjudged more effective treatment (i.e. estimating its mean), we consider the two-stage DLD in which n 2 subjects are further administered the adjudged more effective treatment in the second stage of the design. We obtain some admissibility and minimaxity results for estimating the mean effect of the adjudged more effective treatment. The maximum likelihood estimator is shown to be minimax and admissible. We show that the uniformly minimum variance conditionally unbiased estimator (UMVCUE) of the selected treatment mean is inadmissible and obtain an improved estimator. In this process, we also derive a sufficient condition for inadmissibility of an arbitrary location and permutation equivariant estimator and provide dominating estimators in cases, where this sufficient condition is satisfied. The mean squared error and the bias performances of various competing estimators are compared via a simulation study. A real data example is also provided for illustration purpose.

Assumption-lean falsification tests of rate double-robustness of double-machine-learning estimators

The class of doubly robust (DR) functionals studied by Rotnitzky et al. (2021) is of central importance in economics and biostatistics. It strictly includes both (i) the class of mean-square continuous functionals that can be written as an expectation of an affine functional of a conditional expectation studied by Chernozhukov et al. (2022b) and the class of functionals studied by Robins et al. (2008). The present state-of-the-art estimators for DR functionals ψ are double-machine-learning (DML) estimators (Chernozhukov et al., 2018). A DML estimator ψ ^ 1 of ψ depends on estimates p ^ ( x ) and b ^ x of a pair of nuisance functions p ( x ) and b x , and is said to satisfy "rate double-robustness" if the Cauchy-Schwarz upper bound of its bias is o ( n - 1 / 2 ) . Were it achievable, our scientific goal would have been to construct valid, assumption-lean (i.e. no complexity-reducing assumptions on b or p ) tests of the validity of a nominal (1 - α ) Wald confidence interval (CI) centered at ψ ^ 1 . But this would require a test of the bias to be o ( n - 1 / 2 ) , which can be shown not to exist. We therefore adopt the less ambitious goal of falsifying, when possible, an analyst's justification for her claim that the reported (1 - α ) Wald CI is valid. In many instances, an analyst justifies her claim by imposing complexity-reducing assumptions on b and p to ensure "rate double-robustness". Here we exhibit valid, assumption-lean tests of H 0 : "rate double-robustness holds", with non-trivial power against certain alternatives. If H 0 is rejected, we will have falsified her justification. However, no assumption-lean test of H 0 , including ours, can be a consistent test. Thus, the failure of our test to reject is not meaningful evidence in favor of H 0 .

Variable selection for individualised treatment rules with discrete outcomes

An individualised treatment rule (ITR) is a decision rule that aims to improve individuals' health outcomes by recommending treatments according to subject-specific information. In observational studies, collected data may contain many variables that are irrelevant to treatment decisions. Including all variables in an ITR could yield low efficiency and a complicated treatment rule that is difficult to implement. Thus, selecting variables to improve the treatment rule is crucial. We propose a doubly robust variable selection method for ITRs, and show that it compares favourably with competing approaches. We illustrate the proposed method on data from an adaptive, web-based stress management tool.

Improving the reliability and validity of the IAT with a dynamic model driven by similarity

The Implicit Association Test (IAT), like many behavioral measures, seeks to quantify meaningful individual differences in cognitive processes that are difficult to assess with approaches like self-reports. However, much like other behavioral measures, many IATs appear to show low test-retest reliability and typical scoring methods fail to quantify all of the decision-making processes that generate the overt task performance. Here, we develop a new modeling approach for IATs based on the geometric similarity representation (GSR) model. This model leverages both response times and accuracy on IATs to make inferences about representational similarity between the stimuli and categories. The model disentangles processes related to response caution, stimulus encoding, similarities between concepts and categories, and response processes unrelated to the choice itself. This approach to analyzing IAT data illustrates that the unreliability in IATs is almost entirely attributable to the methods used to analyze data from the task: GSR model parameters show test-retest reliability around .80-.90, on par with reliable self-report measures. Furthermore, we demonstrate how model parameters result in greater validity compared to the IAT D-score, Quad model, and simple diffusion model contrasts, predicting outcomes related to intergroup contact and motivation. Finally, we present a simple point-and-click software tool for fitting the model, which uses a pre-trained neural network to estimate best-fit parameters of the GSR model. This approach allows easy and instantaneous fitting of IAT data with minimal demands on coding or technical expertise on the part of the user, making the new model accessible and effective.

Prospective comparison of restriction spectrum imaging and non-invasive biomarkers to predict upgrading on active surveillance prostate biopsy

BACKGROUND

Protocol-based active surveillance (AS) biopsies have led to poor compliance. To move to risk-based protocols, more accurate imaging biomarkers are needed to predict upgrading on AS prostate biopsy. We compared restriction spectrum imaging (RSI-MRI) generated signal maps as a biomarker to other available non-invasive biomarkers to predict upgrading or reclassification on an AS biopsy.

METHODS

We prospectively enrolled men on prostate cancer AS undergoing repeat biopsy from January 2016 to June 2019 to obtain an MRI and biomarkers to predict upgrading. Subjects underwent a prostate multiparametric MRI and a short duration, diffusion-weighted enhanced MRI called RSI to generate a restricted signal map along with evaluation of 30 biomarkers (14 clinico-epidemiologic features, 9 molecular biomarkers, and 7 radiologic-associated features). Our primary outcome was upgrading or reclassification on subsequent AS prostate biopsy. Statistical analysis included operating characteristic improvement using AUROC and AUPRC.

RESULTS

The individual biomarker with the highest area under the receiver operator characteristic curve (AUC) was RSI-MRI (AUC = 0.84; 95% CI: 0.71-0.96). The best non-imaging biomarker was prostate volume-corrected Prostate Health Index density (PHI, AUC = 0.68; 95% CI: 0.53-0.82). Non-imaging biomarkers had a negligible effect on predicting upgrading at the next biopsy but did improve predictions of overall time to progression in AS.

CONCLUSIONS

RSI-MRI, PIRADS, and PHI could improve the predictive ability to detect upgrading in AS. The strongest predictor of clinically significant prostate cancer on AS biopsy was RSI-MRI signal output.

Using machine learning to improve Q-matrix validation

The Q-matrix, which specifies the relationship between items and attributes, is a crucial component of cognitive diagnostic models (CDMs). A precisely specified Q-matrix allows for valid cognitive diagnostic assessments. In practice, a Q-matrix is usually developed by domain experts, and noted as being subjective and potentially containing misspecifications which can decrease the classification accuracy of examinees. To overcome this, some promising validation methods have been proposed, such as the general discrimination index (GDI) method and the Hull method. In this article, we propose four new methods for Q-matrix validation based on random forest and feed-forward neural network techniques. Proportion of variance accounted for (PVAF) and coefficient of determination (i.e., the McFadden pseudo-R2) are used as input features for developing the machine learning models. Two simulation studies are carried out to examine the feasibility of the proposed methods. Finally, a sub-dataset of the PISA 2000 reading assessment is analyzed as illustration.

A tutorial on Bayesian model-averaged meta-analysis in JASP

Researchers conduct meta-analyses in order to synthesize information across different studies. Compared to standard meta-analytic methods, Bayesian model-averaged meta-analysis offers several practical advantages including the ability to quantify evidence in favor of the absence of an effect, the ability to monitor evidence as individual studies accumulate indefinitely, and the ability to draw inferences based on multiple models simultaneously. This tutorial introduces the concepts and logic underlying Bayesian model-averaged meta-analysis and illustrates its application using the open-source software JASP. As a running example, we perform a Bayesian meta-analysis on language development in children. We show how to conduct a Bayesian model-averaged meta-analysis and how to interpret the results.

Development of the Online Problem Gaming Behavior Index: A New Scale Based on Actual Problem Gambling Behavior Rather Than the Consequences of it

Many items in current problem gambling screens focus on negative consequences of gambling and gambling-related harms. However, few problem gambling screens comprise items that are totally based on actual gambling behavior such as gambling duration, gambling frequency, or gambling late at night. The aim of the present study was to develop and validate the 12-item Online Problem Gambling Behavior Index (OPGBI). A total of 10,000 online Croatian gamblers were administered the OPGBI alongside the nine-item Problem Gambling Severity Index (PGSI), as well as questions regarding types of gambling engaged in and socio-demographic factors. The 12 OPGBI items mainly concern actual gambling behavior. The correlation between OPGBI and PGSI was highly significant (r = 0.68). Three latent factors in the OPGBI were identified (gambling behavior, limit setting, communication with operator). The three factors all significantly correlated with the PGSI score (R2- = 51.8%). The fact that pure gambling behavior related items explained over 50% of the PGSI score strengthens the idea that player tracking could be an important approach in identifying problem gambling.

Analyzing risk factors for post-acute recovery in older adults with Alzheimer's disease and related dementia: A new semi-parametric model for large-scale medicare claims

Nearly 300,000 older adults experience a hip fracture every year, the majority of which occur following a fall. Unfortunately, recovery after fall-related trauma such as hip fracture is poor, where older adults diagnosed with Alzheimer's disease and related dementia (ADRD) spend a particularly long time in hospitals or rehabilitation facilities during the post-operative recuperation period. Because older adults value functional recovery and spending time at home versus facilities as key outcomes after hospitalization, identifying factors that influence days spent at home after hospitalization is imperative. While several individual-level factors have been identified, the characteristics of the treating hospital have recently been identified as contributors. However, few methodological rigorous approaches are available to help overcome potential sources of bias such as hospital-level unmeasured confounders, informative hospital size, and loss to follow-up due to death. This article develops a useful tool equipped with unsupervised learning to simultaneously handle statistical complexities that are often encountered in health services research, especially when using large administrative claims databases. The proposed estimator has a closed form, thus only requiring light computation load in a large-scale study. We further develop its asymptotic properties with stabilized inference assisted by unsupervised clustering. Extensive simulation studies demonstrate superiority of the proposed estimator compared to existing estimators.

Lower Rates of Hypocalcemia Following Near-Infrared Autofluorescence Use in Thyroidectomy: A Meta-Analysis of RCTs

BACKGROUND

Iatrogenic injury of the parathyroid glands is the most frequent complication after total thyroidectomy.

OBJECTIVE

To determine the effectiveness of near-infrared autofluorescence (NIRAF) in reducing postoperative hypocalcemia following total thyroidectomy.

METHODS

PubMed, Scopus, and Google Scholar databases were searched. Randomised trials reporting at least one hypocalcemia outcome following total thyroidectomy using NIRAF were included.

RESULTS

The qualitative data synthesis comprised 1363 patients from nine randomised studies, NIRAF arm = 636 cases and non-NIRAF arm = 637 cases. There was a statistically significant difference in the overall rate of hypocalcemia log(OR) = -0.7 [(-1.01, -0.40), M-H, REM, CI = 95%] and temporary hypocalcemia log(OR) = -0.8 [(-1.01, -0.59), M-H, REM, CI = 95%] favouring the NIRAF. The difference in the rate of permanent hypocalcemia log(OR) = -1.09 [(-2.34, 0.17), M-H, REM, CI = 95%] between the two arms was lower in the NIRAF arm but was not statistically significant.

CONCLUSIONS

NIRAF during total thyroidectomy helps in reducing postoperative hypocalcemia. Level of evidence-1.

ADELLE: A global testing method for Trans-eQTL mapping

Understanding the genetic regulatory mechanisms of gene expression is a challenging and ongoing problem. Genetic variants that are associated with expression levels are readily identified when they are proximal to the gene (i.e., cis-eQTLs), but SNPs distant from the gene whose expression levels they are associated with (i.e., trans-eQTLs) have been much more difficult to discover, even though they account for a majority of the heritability in gene expression levels. A major impediment to the identification of more trans-eQTLs is the lack of statistical methods that are powerful enough to overcome the obstacles of small effect sizes and large multiple testing burden of trans-eQTL mapping. Here, we propose ADELLE, a powerful statistical testing framework that requires only summary statistics and is designed to be most sensitive to SNPs that are associated with multiple gene expression levels, a characteristic of many trans-eQTLs. In simulations, we show that ADELLE is more powerful than other methods at detecting SNPs that are associated with 0.2-2% of the traits. We apply ADELLE to a mouse advanced intercross line data set and show its ability to find trans-eQTLs that were not significant under a standard analysis. This demonstrates that ADELLE is a powerful tool at uncovering trans regulators of genetic expression.

Integrated multiomic analysis reveals disulfidptosis subtypes in glioblastoma: implications for immunotherapy, targeted therapy, and chemotherapy

Introduction

Glioblastoma (GBM) presents significant challenges due to its malignancy and limited treatment options. Precision treatment requires subtyping patients based on prognosis. Disulfidptosis, a novel cell death mechanism, is linked to aberrant glucose metabolism and disulfide stress, particularly in tumors expressing high levels of SLC7A11. The exploration of disulfidptosis may provide a new perspective for precise diagnosis and treatment of glioblastoma.

Methods

Transcriptome sequencing was conducted on samples from GBM patients treated at Tiantan Hospital (January 2022 - December 2023). Data from CGGA and TCGA databases were collected. Consensus clustering based on disulfidptosis features categorized GBM patients into two subtypes (DRGclusters). Tumor immune microenvironment, response to immunotherapy, and drug sensitivity were analyzed. An 8-gene disulfidptosis-based subtype predictor was developed using LASSO machine learning algorithm and validated on CGGA dataset.

Results

Patients in DRGcluster A exhibited improved overall survival (OS) compared to DRGcluster B. DRGcluster subtypes showed differences in tumor immune microenvironment and response to immunotherapy. The predictor effectively stratified patients into high and low-risk groups. Significant differences in IC50 values for chemotherapy and targeted therapy were observed between risk groups.

Discussion

Disulfidptosis-based classification offers promise as a prognostic predictor for GBM. It provides insights into tumor immune microenvironment and response to therapy. The predictor aids in patient stratification and personalized treatment selection, potentially improving outcomes for GBM patients.

Identifying Predictors of Anal HPV Status in HPV-Vaccinated MSM: A Machine Learning Approach

Anal human papillomavirus (HPV) infection has a high prevalence in men who have sex with men (MSM), resulting in an increased risk for anal cancer. The present work aimed to identify factors associated with HPV in a prospective cohort of HPV-vaccinated MSM using a random forest (RF) approach. This observational study enrolled MSM patients admitted to an Italian (sexually transmitted infection) STI-AIDS Unit. For each patient, rectal swabs for 28 different HPV genotype detection were collected. Two RF algorithms were applied to evaluate predictors that were most associated with HPV. The cohort included 135 MSM, 49% of whom were HIV-positive with a median age of 39 years. In model 1 (baseline information), age, age sexual debut, HIV, number of lifetime sex partners, STIs, were most associated with the HPV. In model 2 (follow-up information), age, age sexual debut, HIV, STI class, and follow-up. The RF algorithm exhibited good performances with 61% and 83% accuracy for models 1 and 2, respectively. Traditional risk factors for anal HPV infection, such as drug use, receptive anal intercourse, and multiple sexual partner, were found to have low importance in predicting HPV status. The present results suggest the need to focus on HPV prevention campaigns.

Multiply robust estimation of natural indirect effects with multiple ordered mediators

Multiple mediation analysis is a powerful methodology to assess causal effects in the presence of multiple mediators. Several methodologies, such as G-computation and inverse-probability-weighting, have been widely used to draw inferences about natural indirect effects (NIEs). However, a limitation of these methods is their potential for model misspecification. Although powerful semiparametric methods with high robustness and consistency have been developed for inferring average causal effects and for analyzing the effects of a single mediator, a comparably robust method for multiple mediation analysis is still lacking. Therefore, this theoretical study proposes a method of using multiply robust estimators of NIEs in the presence of multiple ordered mediators. We show that the proposed estimators not only enjoy the multiply robustness to model misspecification, they are also consistent and asymptotically normal under regular conditions. We also performed simulations for empirical comparisons of the finite-sample properties between our multiply robust estimators and existing methods. In an illustrative example, a dataset for liver disease patients in Taiwan is used to examine the mediating roles of liver damage and liver cancer in the pathway from hepatitis B/C virus infection to mortality. The model is implemented in the open-source R package "MedMR."

Smoothed quantile residual life regression analysis with application to the Korea HIV/AIDS cohort study

BACKGROUND

The residual life of a patient with human immunodeficiency virus (HIV) is of major interest to patients and their physicians. While existing analyses of HIV patient survival focus mostly on data collected at baseline, residual life analysis allows for dynamic analysis based on additional data collected over a period of time. As survival times typically exhibit a right-skewed distribution, the median provides a more useful summary of the underlying distribution than the mean. In this paper, we propose an efficient inference procedure that fits a semiparametric quantile regression model assessing the effect of longitudinal biomarkers on the residual life of HIV patients until the development of dyslipidemia, a disease becoming more prevalent among those with HIV.

METHODS

For estimation of model parameters, we propose an induced smoothing method that smooths nonsmooth estimating functions based on check functions. For variance estimation, we propose an efficient resampling-based estimator. The proposed estimators are theoretically justified. Simulation studies are used to evaluate their finite sample performances, including their prediction accuracy. We analyze the Korea HIV/AIDS cohort study data to examine the effects of CD4 (cluster of differentiation 4) cell count on the residual life of HIV patients to the onset of dyslipidemia.

RESULTS

The proposed estimator is shown to be consistent and normally distributed asymptotically. Under various simulation settings, our estimates are approximately unbiased. Their variances estimates are close to the empirical variances and their computational efficiency is superior to that of the nonsmooth counterparts. Two measures of prediction performance indicate that our method adequately reflects the dynamic character of longitudinal biomarkers and residual life. The analysis of the Korea HIV/AIDS cohort study data shows that CD4 cell count is positively associated with residual life to the onset of dyslipidemia but the effect is not statistically significant.

CONCLUSIONS

Our method enables direct prediction of residual lifetimes with a dynamic feature that accommodates data accumulated at different times. Our estimator significantly improves computational efficiency in variance estimation compared to the existing nonsmooth estimator. Analysis of the HIV/AIDS cohort study data reveals dynamic effects of CD4 cell count on the residual life to the onset of dyslipidemia.

Predicting proximal tubule failed repair drivers through regularized regression analysis of single cell multiomic sequencing

Renal proximal tubule epithelial cells have considerable intrinsic repair capacity following injury. However, a fraction of injured proximal tubule cells fails to undergo normal repair and assumes a proinflammatory and profibrotic phenotype that may promote fibrosis and chronic kidney disease. The healthy to failed repair change is marked by cell state-specific transcriptomic and epigenomic changes. Single nucleus joint RNA- and ATAC-seq sequencing offers an opportunity to study the gene regulatory networks underpinning these changes in order to identify key regulatory drivers. We develop a regularized regression approach to construct genome-wide parametric gene regulatory networks using multiomic datasets. We generate a single nucleus multiomic dataset from seven adult human kidney samples and apply our method to study drivers of a failed injury response associated with kidney disease. We demonstrate that our approach is a highly effective tool for predicting key cis- and trans-regulatory elements underpinning the healthy to failed repair transition and use it to identify NFAT5 as a driver of the maladaptive proximal tubule state.

Estimation of optimal treatment regimes with electronic medical record data using the residual life value estimator

Clinicians and patients must make treatment decisions at a series of key decision points throughout disease progression. A dynamic treatment regime is a set of sequential decision rules that return treatment decisions based on accumulating patient information, like that commonly found in electronic medical record (EMR) data. When applied to a patient population, an optimal treatment regime leads to the most favorable outcome on average. Identifying optimal treatment regimes that maximize residual life is especially desirable for patients with life-threatening diseases such as sepsis, a complex medical condition that involves severe infections with organ dysfunction. We introduce the residual life value estimator (ReLiVE), an estimator for the expected value of cumulative restricted residual life under a fixed treatment regime. Building on ReLiVE, we present a method for estimating an optimal treatment regime that maximizes expected cumulative restricted residual life. Our proposed method, ReLiVE-Q, conducts estimation via the backward induction algorithm Q-learning. We illustrate the utility of ReLiVE-Q in simulation studies, and we apply ReLiVE-Q to estimate an optimal treatment regime for septic patients in the intensive care unit using EMR data from the Multiparameter Intelligent Monitoring Intensive Care database. Ultimately, we demonstrate that ReLiVE-Q leverages accumulating patient information to estimate personalized treatment regimes that optimize a clinically meaningful function of residual life.

Genomic prediction using machine learning: a comparison of the performance of regularized regression, ensemble, instance-based and deep learning methods on synthetic and empirical data

BACKGROUND

The accurate prediction of genomic breeding values is central to genomic selection in both plant and animal breeding studies. Genomic prediction involves the use of thousands of molecular markers spanning the entire genome and therefore requires methods able to efficiently handle high dimensional data. Not surprisingly, machine learning methods are becoming widely advocated for and used in genomic prediction studies. These methods encompass different groups of supervised and unsupervised learning methods. Although several studies have compared the predictive performances of individual methods, studies comparing the predictive performance of different groups of methods are rare. However, such studies are crucial for identifying (i) groups of methods with superior genomic predictive performance and assessing (ii) the merits and demerits of such groups of methods relative to each other and to the established classical methods. Here, we comparatively evaluate the genomic predictive performance and informally assess the computational cost of several groups of supervised machine learning methods, specifically, regularized regression methods, deep, ensemble and instance-based learning algorithms, using one simulated animal breeding dataset and three empirical maize breeding datasets obtained from a commercial breeding program.

RESULTS

Our results show that the relative predictive performance and computational expense of the groups of machine learning methods depend upon both the data and target traits and that for classical regularized methods, increasing model complexity can incur huge computational costs but does not necessarily always improve predictive accuracy. Thus, despite their greater complexity and computational burden, neither the adaptive nor the group regularized methods clearly improved upon the results of their simple regularized counterparts. This rules out selection of one procedure among machine learning methods for routine use in genomic prediction. The results also show that, because of their competitive predictive performance, computational efficiency, simplicity and therefore relatively few tuning parameters, the classical linear mixed model and regularized regression methods are likely to remain strong contenders for genomic prediction.

CONCLUSIONS

The dependence of predictive performance and computational burden on target datasets and traits call for increasing investments in enhancing the computational efficiency of machine learning algorithms and computing resources.

Recent Methodological Trends in Epidemiology: No Need for Data-Driven Variable Selection?

Variable selection in regression models is a particularly important issue in epidemiology, where one usually encounters observational studies. In contrast to randomized trials or experiments, confounding is often not controlled by the study design, but has to be accounted for by suitable statistical methods. For instance, when risk factors should be identified with unconfounded effect estimates, multivariable regression techniques can help to adjust for confounders. We investigated the current practice of variable selection in 4 major epidemiologic journals in 2019 and found that the majority of articles used subject-matter knowledge to determine a priori the set of included variables. In comparison with previous reviews from 2008 and 2015, fewer articles applied data-driven variable selection. Furthermore, for most articles the main aim of analysis was hypothesis-driven effect estimation in rather low-dimensional data situations (i.e., large sample size compared with the number of variables). Based on our results, we discuss the role of data-driven variable selection in epidemiology.

An interpretable XGBoost-based approach for Arctic navigation risk assessment

The Northern Sea Route (NSR) makes travel between Europe and Asia shorter and quicker than a southern transit via the Strait of Malacca and Suez Canal. It provides greater access to Arctic resources such as oil and gas. As global warming accelerates, melting Arctic ice caps are likely to increase traffic in the NSR and enhance its commercial viability. Due to the harsh Arctic environment imposing threats to the safety of ship navigation, it is necessary to assess Arctic navigation risk to maintain shipping safety. Currently, most studies are focused on the conventional assessment of the risk, which lacks the validation based on actual data. In this study, actual data about Arctic navigation environment and related expert judgments were used to generate a structured data set. Based on the structured data set, extreme gradient boosting (XGBoost) and alternative methods were used to establish models for the assessment of Arctic navigation risk, which were validated using cross-validation. The results show that compared with alternative models, XGBoost models have the best performance in terms of mean absolute errors and root mean squared errors. The XGBoost models can learn and reproduce expert judgments and knowledge for the assessment of Arctic navigation risk. Feature importance (FI) and shapley additive explanations (SHAP) are used to further interpret the relationship between input data and predictions. The application of XGBoost, FI, and SHAP is aimed to improve the safety of Arctic shipping using advanced artificial intelligence techniques. The validated assessment enhances the quality and robustness of assessment.

Investigating sensitivity through the lens of parents: validation of the parent-report version of the Highly Sensitive Child scale

Children differ in their environmental sensitivity (ES), which can be measured observationally or by self-report questionnaire. A parent-report scale represents an important tool for investigating ES in younger children but has to be psychometrically robust and valid. In the current multistudy, we validated the parent-report version of the Highly Sensitive Child (HSC-PR) scale in Italian children, evaluating its factorial structure (Study 1, N = 1,857, 6.2 years, age range: 2.6-14 years) through a multigroup Confirmatory Factory Analysis in preschoolers (n = 1,066, 4.2 years) and school-age children (n = 791, 8.8 years). We then investigated the HSC-PR relationship with established temperament traits (Study 2, N = 327, 4.3 years), before exploring whether the scale moderates the effects of parenting stress on children's emotion regulation (Study 3, N = 112, 6.5 years). We found support for a bi-factor structure in both groups, though in preschoolers minor adaptations were suggested for one item. Importantly, the HSC-PR did not fully overlap with common temperament traits and moderated the effects of parenting stress on children emotion regulation. To conclude, the HSC-PR performs well and appears to capture ES in children.

Epidemiology of patellar dislocations in the United States from 2001 to 2020: results of a national emergency department database

OBJECTIVES

Recent studies have shown an increasing incidence of patellar dislocations among children and adolescents. Updated, population-based studies of all patellar dislocations in the United States (US), however, are lacking. This study investigated recent trends in injury rates and demographics among patients sustaining patellar dislocations in the US from 2001 to 2020.

METHODS

This descriptive epidemiologic study retrospectively analyzed the National Electronic Injury Surveillance System (NEISS) database to identify cases of acute patellar dislocations presenting to US Emergency Departments (EDs) from 2001 to 2020. Annual, overall, and age-adjusted incidence rates (IRs, expressed per 100,000 at-risk person-years) and additional patient/injury characteristics were analyzed. Average annual percent change (AAPC) estimates are presented to indicate the magnitude/direction of trends in annual injury rates.

RESULTS

An estimated total of 159,529 patellar dislocations occurred over the study period for an overall IR of 2.58 (95% CI = 2.04-3.12). Accounting for population growth, the overall annual incidence increased significantly from 2.61 in 2001 to 3.0 in 2020 (AAPC = 2.8, p < 0.0001). When considering sex and age, statistically significant increases in annual IRs were observed among males aged 10-19 years (AAPC = 3.8, p < 0.0001), females aged 10-19 years (AAPC = 5.3, p < 0.0001), and females aged 20-29 years (AAPC = 3.5, p = 0.0152), while no significant changes were observed in any other age groups. Two-thirds of patellar dislocations involved sports-related injury mechanisms. The annual incidence of both sports-related and non-sports-related injuries increased significantly over the study period (sports-related: AAPC = 2.6, p = 0.0001; non-sports-related: AAPC = 3.4, p = 0.0001). Athletic patellar dislocations occurred most commonly in basketball and dance.

CONCLUSION

The number of patients sustaining patellar dislocations is increasing in the US. Similar increasing trends were observed in both males and females aged 10-19 years, whereas injury rates increased in the third decade only among females. A large percentage of injuries occur during athletic activity, but both sports- and non-sports-related patellar dislocations are on the rise.

RANK-BASED INDICES FOR TESTING INDEPENDENCE BETWEEN TWO HIGH-DIMENSIONAL VECTORS

To test independence between two high-dimensional random vectors, we propose three tests based on the rank-based indices derived from Hoeffding's D , Blum-Kiefer-Rosenblatt's R and Bergsma-Dassios-Yanagimoto's τ * . Under the null hypothesis of independence, we show that the distributions of the proposed test statistics converge to normal ones if the dimensions diverge arbitrarily with the sample size. We further derive an explicit rate of convergence. Thanks to the monotone transformation-invariant property, these distribution-free tests can be readily used to generally distributed random vectors including heavily tailed ones. We further study the local power of the proposed tests and compare their relative efficiencies with two classic distance covariance/correlation based tests in high dimensional settings. We establish explicit relationships between D , R , τ * and Pearson's correlation for bivariate normal random variables. The relationships serve as a basis for power comparison. Our theoretical results show that under a Gaussian equicorrelation alternative, (i) the proposed tests are superior to the two classic distance covariance/correlation based tests if the components of random vectors have very different scales; (ii) the asymptotic efficiency of the proposed tests based on D , τ * and R are sorted in a descending order.

Predicting Acoustic Hearing Preservation Following Cochlear Implant Surgery Using Machine Learning

OBJECTIVES

The aim of the study was to train and test supervised machine-learning classifiers to predict acoustic hearing preservation after CI using preoperative clinical data.

STUDY DESIGN

Retrospective predictive modeling study of prospectively collected single-institution CI dataset.

METHODS

One hundred and seventy-five patients from a REDCap database including 761 patients >18 years who underwent CI and had audiometric testing preoperatively and one month after surgery were included. The primary outcome variable was the lowest quartile change in acoustic hearing at one month after CI using various formulae (standard pure tone average, SPTA; low-frequency PTA, LFPTA). Analysis involved applying multivariate logistic regression to detect statistical associations and training and testing supervised learning classifiers. Classifier performance was assessed with numerous metrics including area under the receiver operating characteristic curve (AUC) and Matthews correlation coefficient (MCC).

RESULTS

Lowest quartile change (indicating hearing preservation) in SPTA was positively associated with a history of meningitis, preoperative LFPTA, and preoperative SPTA. Lowest quartile change in SPTA was negatively associated with sudden hearing loss, noise exposure, aural fullness, and abnormal anatomy. Lowest quartile change in LFPTA was positively associated with preoperative LFPTA. Lowest quartile change in LFPTA was negatively associated with tobacco use. Random forest demonstrated the highest mean classification performance on the validation dataset when predicting each of the outcome variables.

CONCLUSIONS

Machine learning demonstrated utility for predicting preservation of residual acoustic hearing in patients undergoing CI surgery, and the detected associations facilitated the interpretation of our machine-learning models. The models and statistical associations together may be used to facilitate improvements in shared clinical decision-making and patient outcomes.

LEVEL OF EVIDENCE

3 Laryngoscope, 134:926-936, 2024.

Emotional Engagement and Caring Relationships: The Assessment of Emotion Regulation Repertoires of Nurses

In spite of the importance of emotion regulation for nurses' well-being, little is known about which strategies nurses habitually use, how these strategies combine in order to regulate their emotional distress, and how these are related to their caregiving orientations. The current study aimed to explore the emotion regulation repertoires that characterize health-care providers and to investigate the association between these repertoires and caregiving orientations in a sample of nurses. Firstly, a confirmatory factor analyses was run to test the suitability of the Regulation of Emotion System Survey for the assessment of six emotion regulation strategies among health-care providers. Subsequently, the latent profiles analysis was employed to explore emotion regulation repertoires. Three repertoires emerged: The Average, the Suppression Propensity and the Engagement Propensity profiles. The participants of the last two groups relied on Expressive Suppression and Engagement, respectively, more often than others. Nurses were more likely to be placed within the Engagement Propensity group when compared to the first responders, and higher levels of hyperactivation of the Caregiving System were associated with this repertoire. A greater reliance on Expressive Engagement among nurses was discussed in terms of the fact that nurses usually have a longer and more care-oriented relationships with patients than first responders.

Association between food control inspection grades and regional incidence of infectious foodborne diseases in Finland

We studied regional food control inspection grades and their relation to regional incidence of domestically acquired foodborne diseases (caused by Campylobacter spp. Salmonella spp. enterohemorrhagic Escherichia coli (EHEC), and Listeria monocytogenes) using food control inspection data of local food business operators and infectious disease data from 2014 to 2019 from Finland. We observed that inferior overall inspection grades were associated with increased incidence of Salmonella infections (p=0.02). Specifically, inferior grades on cleanliness of facilities, surfaces, and equipment were associated with increased incidence of Salmonella infections (p=0.04). For this topical inspection area, a high effect size was also seen for Campylobacter infections (p=0.06). Of the individual inspection items, an association between increased incidence of Campylobacter infections and inferior grades on storage of foodstuffs (p=0.01) and verification of hygiene proficiency (p=0.03) was observed. These results suggest that food control recognizes non-compliances that may predispose to foodborne diseases.

False Correlation Reduction for Offline Reinforcement Learning

Offline reinforcement learning (RL) harnesses the power of massive datasets for resolving sequential decision problems. Most existing papers only discuss defending against out-of-distribution (OOD) actions while we investigate a broader issue, the false correlations between epistemic uncertainty and decision-making, an essential factor that causes suboptimality. In this paper, we propose falSe COrrelation REduction (SCORE) for offline RL, a practically effective and theoretically provable algorithm. We empirically show that SCORE achieves the SoTA performance with 3.1x acceleration on various tasks in a standard benchmark (D4RL). The proposed algorithm introduces an annealing behavior cloning regularizer to help produce a high-quality estimation of uncertainty which is critical for eliminating false correlations from suboptimality. Theoretically, we justify the rationality of the proposed method and prove its convergence to the optimal policy with a sublinear rate under mild assumptions.

Plasma SOMAmer proteomics of postoperative delirium

BACKGROUND

Postoperative delirium is prevalent in older adults and has been shown to increase the risk of long-term cognitive decline. Plasma biomarkers to identify the risk for postoperative delirium and the risk of Alzheimer's disease and related dementias are needed.

METHODS

This biomarker discovery case-control study aimed to identify plasma biomarkers associated with postoperative delirium. Patients aged ≥65 years undergoing major elective noncardiac surgery were recruited. The preoperative plasma proteome was interrogated with SOMAmer-based technology targeting 1433 biomarkers.

RESULTS

In 40 patients (20 with vs. 20 without postoperative delirium), a preoperative panel of 12 biomarkers discriminated patients with postoperative delirium with an accuracy of 97.5%. The final model of five biomarkers delivered a leave-one-out cross-validation accuracy of 80%. Represented biological pathways included lysosomal and immune response functions.

CONCLUSION

In older patients who have undergone major surgery, plasma SOMAmer proteomics may provide a relatively non-invasive benchmark to identify biomarkers associated with postoperative delirium.

Subjective awareness of sleepiness while driving in younger and older adults

Understanding whether drivers can accurately assess sleepiness is essential for educational campaigns advising drivers to stop driving when feeling sleepy. However, few studies have examined this in real-world driving environments, particularly among older drivers who comprise a large proportion of all road users. To examine the accuracy of subjective sleepiness ratings in predicting subsequent driving impairment and physiological drowsiness, 16 younger (21-33 years) and 17 older (50-65 years) adults drove an instrumented vehicle for 2 h on closed loop under two conditions: well-rested and 29 h sleep deprivation. Sleepiness ratings (Karolinska Sleepiness Scale, Likelihood of Falling Asleep scale, Sleepiness Symptoms Questionnaire) were obtained every 15min, alongside lane deviations, near crash events, and ocular indices of drowsiness. All subjective sleepiness measures increased with sleep deprivation for both age groups (p < 0.013). While most subjective sleepiness ratings significantly predicted driving impairment and drowsiness in younger adults (OR: 1.7-15.6, p < 0.02), this was only apparent for KSS, likelihood of falling asleep, and "difficulty staying in the lane for the older adults" (OR: 2.76-2.86, p = 0.02). This may be due to an altered perception of sleepiness in older adults, or due to lowered objective signs of impairment in the older group. Our data suggest that (i) younger and older drivers are aware of sleepiness; (ii) the best subjective scale may differ across age groups; and (iii) future research should expand on the best subjective measures to inform of crash risk in older adults to inform tailored educational road safety campaigns on signs of sleepiness.

Markov Subsampling Based on Huber Criterion

Subsampling is an important technique to tackle the computational challenges brought by big data. Many subsampling procedures fall within the framework of importance sampling, which assigns high sampling probabilities to the samples appearing to have big impacts. When the noise level is high, those sampling procedures tend to pick many outliers and thus often do not perform satisfactorily in practice. To tackle this issue, we design a new Markov subsampling strategy based on Huber criterion (HMS) to construct an informative subset from the noisy full data; the constructed subset then serves as refined working data for efficient processing. HMS is built upon a Metropolis-Hasting procedure, where the inclusion probability of each sampling unit is determined using the Huber criterion to prevent over scoring the outliers. Under mild conditions, we show that the estimator based on the subsamples selected by HMS is statistically consistent with a sub-Gaussian deviation bound. The promising performance of HMS is demonstrated by extensive studies on large-scale simulations and real data examples.

Well-functioning bicuspid aortic valves should be preserved during aortic replacement for the ascending aortopathy phenotype

OBJECTIVES

Consensus has not been reached on whether or not to replace or preserve a well-functioning bicuspid aortic valve (BAV) in patients undergoing aortic replacement for the ascending phenotype of BAV aortopathy. We characterize morphology, evaluate progression of aortic regurgitation or aortic stenosis, and investigate the need for aortic valve replacement in patients whose well-functioning BAV was preserved during ascending aortic replacement ≥10 years prior.

METHODS

From January 1991 to August 2011, 191 patients with a well-functioning BAV underwent supracoronary aortic replacement (113 valves were minimally repaired). Aortic morphology was evaluated, aortic regurgitation grade and transvalvular aortic gradient modeled parametrically, and survival assessed by the Kaplan-Meier method. Median follow-up was 10 years.

RESULTS

Mean aortic diameter was 2.9 ± 0.53 cm at the annulus and 4.2 ± 0.55 cm at the sinuses. Mean maximum ascending diameter was 5.1 ± 0.49 cm. All patients exhibited a cusp-fusion BAV phenotype. Fifteen-year progression to severe aortic regurgitation was 3.2%. Mean aortic valve gradient began to rise 5 years postoperatively to 27 mm Hg by 14 years. Freedom from aortic valve replacement at 1, 5, 10, and 15 years was 100%, 95%, 83%, and 63%, respectively. Minimal valve repair was not associated with late aortic valve replacement. Fifteen-year survival was 74%.

CONCLUSIONS

Preserving a well-functioning BAV should be considered in carefully selected patients undergoing aortic replacement for the ascending phenotype of BAV aortopathy. The valves remain durable in the long term, with slow progression of regurgitation or stenosis, and low probability of aortic valve replacement through 10 years.

Efficient computation of high-dimensional penalized generalized linear mixed models by latent factor modeling of the random effects

Modern biomedical datasets are increasingly high-dimensional and exhibit complex correlation structures. Generalized linear mixed models (GLMMs) have long been employed to account for such dependencies. However, proper specification of the fixed and random effects in GLMMs is increasingly difficult in high dimensions, and computational complexity grows with increasing dimension of the random effects. We present a novel reformulation of the GLMM using a factor model decomposition of the random effects, enabling scalable computation of GLMMs in high dimensions by reducing the latent space from a large number of random effects to a smaller set of latent factors. We also extend our prior work to estimate model parameters using a modified Monte Carlo Expectation Conditional Minimization algorithm, allowing us to perform variable selection on both the fixed and random effects simultaneously. We show through simulation that through this factor model decomposition, our method can fit high-dimensional penalized GLMMs faster than comparable methods and more easily scale to larger dimensions not previously seen in existing approaches.

Multiobjective tree-based reinforcement learning for estimating tolerant dynamic treatment regimes

A dynamic treatment regime (DTR) is a sequence of treatment decision rules that dictate individualized treatments based on evolving treatment and covariate history. It provides a vehicle for optimizing a clinical decision support system and fits well into the broader paradigm of personalized medicine. However, many real-world problems involve multiple competing priorities, and decision rules differ when trade-offs are present. Correspondingly, there may be more than one feasible decision that leads to empirically sufficient optimization. In this paper, we propose a concept of "tolerant regime," which provides a set of individualized feasible decision rules under a prespecified tolerance rate. A multiobjective tree-based reinforcement learning (MOT-RL) method is developed to directly estimate the tolerant DTR (tDTR) that optimizes multiple objectives in a multistage multitreatment setting. At each stage, MOT-RL constructs an unsupervised decision tree by modeling the counterfactual mean outcome of each objective via semiparametric regression and maximizing a purity measure constructed by the scalarized augmented inverse probability weighted estimators (SAIPWE). The algorithm is implemented in a backward inductive manner through multiple decision stages, and it estimates the optimal DTR and tDTR depending on the decision-maker's preferences. Multiobjective tree-based reinforcement learning is robust, efficient, easy-to-interpret, and flexible to different settings. We apply MOT-RL to evaluate 2-stage chemotherapy regimes that reduce disease burden and prolong survival for advanced prostate cancer patients using a dataset collected at MD Anderson Cancer Center.

Adaptive sequential surveillance with network and temporal dependence

Strategic test allocation is important for control of both emerging and existing pandemics (eg, COVID-19, HIV). It supports effective epidemic control by (1) reducing transmission via identifying cases and (2) tracking outbreak dynamics to inform targeted interventions. However, infectious disease surveillance presents unique statistical challenges. For instance, the true outcome of interest (positive infection status) is often a latent variable. In addition, presence of both network and temporal dependence reduces data to a single observation. In this work, we study an adaptive sequential design, which allows for unspecified dependence among individuals and across time. Our causal parameter is the mean latent outcome we would have obtained, if, starting at time t given the observed past, we had carried out a stochastic intervention that maximizes the outcome under a resource constraint. The key strength of the method is that we do not have to model network and time dependence: a short-term performance Online Super Learner is used to select among dependence models and randomization schemes. The proposed strategy learns the optimal choice of testing over time while adapting to the current state of the outbreak and learning across samples, through time, or both. We demonstrate the superior performance of the proposed strategy in an agent-based simulation modeling a residential university environment during the COVID-19 pandemic.

A Bayesian survival treed hazards model using latent Gaussian processes

Survival models are used to analyze time-to-event data in a variety of disciplines. Proportional hazard models provide interpretable parameter estimates, but proportional hazard assumptions are not always appropriate. Non-parametric models are more flexible but often lack a clear inferential framework. We propose a Bayesian treed hazards partition model that is both flexible and inferential. Inference is obtained through the posterior tree structure and flexibility is preserved by modeling the log-hazard function in each partition using a latent Gaussian process. An efficient reversible jump Markov chain Monte Carlo algorithm is accomplished by marginalizing the parameters in each partition element via a Laplace approximation. Consistency properties for the estimator are established. The method can be used to help determine subgroups as well as prognostic and/or predictive biomarkers in time-to-event data. The method is compared with some existing methods on simulated data and a liver cirrhosis dataset.

Merging or ensembling: integrative analysis in multiple neuroimaging studies

The aim of this paper is to systematically investigate merging and ensembling methods for spatially varying coefficient mixed effects models (SVCMEM) in order to carry out integrative learning of neuroimaging data obtained from multiple biomedical studies. The "merged" approach involves training a single learning model using a comprehensive dataset that encompasses information from all the studies. Conversely, the "ensemble" approach involves creating a weighted average of distinct learning models, each developed from an individual study. We systematically investigate the prediction accuracy of the merged and ensemble learners under the presence of different degrees of interstudy heterogeneity. Additionally, we establish asymptotic guidelines for making strategic decisions about when to employ either of these models in different scenarios, along with deriving optimal weights for the ensemble learner. To validate our theoretical results, we perform extensive simulation studies. The proposed methodology is also applied to 3 large-scale neuroimaging studies.

Development and validation of a federated learning framework for detection of subphenotypes of multisystem inflammatory syndrome in children

Background

Multisystem inflammatory syndrome in children (MIS-C) is a severe post-acute sequela of SARS-CoV-2 infection. The highly diverse clinical features of MIS-C necessities characterizing its features by subphenotypes for improved recognition and treatment. However, jointly identifying subphenotypes in multi-site settings can be challenging. We propose a distributed multi-site latent class analysis (dMLCA) approach to jointly learn MIS-C subphenotypes using data across multiple institutions.

Methods

We used data from the electronic health records (EHR) systems across nine U.S. children's hospitals. Among the 3,549,894 patients, we extracted 864 patients < 21 years of age who had received a diagnosis of MIS-C during an inpatient stay or up to one day before admission. Using MIS-C conditions, laboratory results, and procedure information as input features for the patients, we applied our dMLCA algorithm and identified three MIS-C subphenotypes. As validation, we characterized and compared more granular features across subphenotypes. To evaluate the specificity of the identified subphenotypes, we further compared them with the general subphenotypes identified in the COVID-19 infected patients.

Findings

Subphenotype 1 (46.1%) represents patients with a mild manifestation of MIS-C not requiring intensive care, with minimal cardiac involvement. Subphenotype 2 (25.3%) is associated with a high risk of shock, cardiac and renal involvement, and an intermediate risk of respiratory symptoms. Subphenotype 3 (28.6%) represents patients requiring intensive care, with a high risk of shock and cardiac involvement, accompanied by a high risk of >4 organ system being impacted. Importantly, for hospital-specific clinical decision-making, our algorithm also revealed a substantial heterogeneity in relative proportions of these three subtypes across hospitals. Properly accounting for such heterogeneity can lead to accurate characterization of the subphenotypes at the patient-level.

Interpretation

Our identified three MIS-C subphenotypes have profound implications for personalized treatment strategies, potentially influencing clinical outcomes. Further, the proposed algorithm facilitates federated subphenotyping while accounting for the heterogeneity across hospitals.

Robustness and reproducibility for AI learning in biomedical sciences: RENOIR

Artificial intelligence (AI) techniques are increasingly applied across various domains, favoured by the growing acquisition and public availability of large, complex datasets. Despite this trend, AI publications often suffer from lack of reproducibility and poor generalisation of findings, undermining scientific value and contributing to global research waste. To address these issues and focusing on the learning aspect of the AI field, we present RENOIR (REpeated random sampliNg fOr machIne leaRning), a modular open-source platform for robust and reproducible machine learning (ML) analysis. RENOIR adopts standardised pipelines for model training and testing, introducing elements of novelty, such as the dependence of the performance of the algorithm on the sample size. Additionally, RENOIR offers automated generation of transparent and usable reports, aiming to enhance the quality and reproducibility of AI studies. To demonstrate the versatility of our tool, we applied it to benchmark datasets from health, computer science, and STEM (Science, Technology, Engineering, and Mathematics) domains. Furthermore, we showcase RENOIR's successful application in recently published studies, where it identified classifiers for SET2D and TP53 mutation status in cancer. Finally, we present a use case where RENOIR was employed to address a significant pharmacological challenge-predicting drug efficacy. RENOIR is freely available at https://github.com/alebarberis/renoir .

Exploring changes in brain function in IBD patients using SPCCA: a study of simultaneous EEG-fMRI

Research on functional changes in the brain of inflammatory bowel disease (IBD) patients is emerging around the world, which brings new perspectives to medical research. In this paper, the methods of canonical correlation analysis (CCA), kernel canonical correlation analysis (KCCA), and sparsity preserving canonical correlation analysis (SPCCA) were applied to the fusion of simultaneous EEG-fMRI data from 25 IBD patients and 15 healthy individuals. The CCA, KCCA and SPCCA fusion methods were used for data processing to compare the results obtained by the three methods. The results clearly show that there is a significant difference in the activation intensity between IBD and healthy control (HC), not only in the frontal lobe (p < 0.01) and temporal lobe (p < 0.01) regions, but also in the posterior cingulate gyrus (p < 0.01), gyrus rectus (p < 0.01), and amygdala (p < 0.01) regions, which are usually neglected. The mean difference in the SPCCA activation intensity was 60.1. However, the mean difference in activation intensity was only 36.9 and 49.8 by using CCA and KCCA. In addition, the correlation of the relevant components selected during the SPCCA calculation was high, with correlation components of up to 0.955; alternatively, the correlations obtained from CCA and KCCA calculations were only 0.917 and 0.926, respectively. It can be seen that SPCCA is indeed superior to CCA and KCCA in processing high-dimensional multimodal data. This work reveals the process of analyzing the brain activation state in IBD disease, provides a further perspective for the study of brain function, and opens up a new avenue for studying the SPCCA method and the change in the intensity of brain activation in IBD disease.

Optimizing the use of ketamine to reduce chronic postsurgical pain in women undergoing mastectomy for oncologic indication: study protocol for the KALPAS multicenter randomized controlled trial

BACKGROUND

Mastectomies are commonly performed and strongly associated with chronic postsurgical pain (CPSP), more specifically termed postmastectomy pain syndrome (PMPS), with 25-60% of patients reporting pain 3 months after surgery. PMPS interferes with function, recovery, and compliance with adjuvant therapy. Importantly, it is associated with chronic opioid use, as a recent study showed that 1 in 10 patients continue to use opioids at least 3 months after curative surgery. The majority of PMPS patients are women, and, over the past 10 years, women have outpaced men in the rate of growth in opioid dependence. Standard perioperative multimodal analgesia is only modestly effective in prevention of CPSP. Thus, interventions to reduce CPSP and PMPS are urgently needed. Ketamine is well known to improve pain and reduce opioid use in the acute postoperative period. Additionally, ketamine has been shown to control mood in studies of anxiety and depression. By targeting acute pain and improving mood in the perioperative period, ketamine may be able to prevent the development of CPSP.

METHODS

Ketamine analgesia for long-lasting pain relief after surgery (KALPAS) is a phase 3, multicenter, randomized, placebo-controlled, double-blind trial to study the effectiveness of ketamine in reducing PMPS. The study compares continuous perioperative ketamine infusion vs single-dose ketamine in the postanesthesia care unit vs placebo for reducing PMPS. Participants are followed for 1 year after surgery. The primary outcome is pain at the surgical site at 3 months after the index surgery as assessed with the Brief Pain Inventory-short form pain severity subscale.

DISCUSSION

This project is part of the NIH Helping to End Addiction Long-term (HEAL) Initiative, a nationwide effort to address the opioid public health crisis. This study can substantially impact perioperative pain management and can contribute significantly to combatting the opioid epidemic.

TRIAL REGISTRATION

ClinicalTrials.gov NCT05037123. Registered on September 8, 2021.