Covariate Values Research Articles

Applying Neural ODEs to Derive a Mechanism-Based Model for Characterizing Maturation-Related Serum Creatinine Dynamics in Preterm Newborns.

Serum creatinine in neonates follows complex dynamics due to maturation processes, most pronounced in the first few weeks of life. The development of a mechanism-based model describing complex dynamics requires high expertise in pharmacometric (PMX) modeling and substantial model development time. A recently published machine learning (ML) approach of low-dimensional neural ordinary differential equations (NODEs) is capable of modeling such data from newborns automatically. However, this efficient data-driven approach in itself does not result in a clinically interpretable model. In this work, an approach to deriving an interpretable model with reasonable PMX-type functions is presented. This "translation" was applied to derive a PMX model for serum creatinine in neonates considering maturation processes and covariates. The developed model was compared to a previously published mechanism-based PMX model whereas both models had similar mechanistic structures. The developed model was then utilized to simulate serum creatinine concentrations in the first few weeks of life considering different covariate values for gestational age and birth weight. The reference serum creatinine values derived from these simulations are consistent with observed serum creatinine values and previously published reference values. Thus, the presented NODE-based ML approach to model complex serum creatinine dynamics in newborns and derive interpretable, mathematical-statistical components similar to those in a conventional PMX model demonstrates a novel, viable approach to facilitate the modeling of complex dynamics in clinical settings and pediatric drug development.

Self-normalized inference for stationarity of irregular spatial data

Self-normalized inference for stationarity of irregular spatial data

Pointwise Nonparametric Estimation of Odds Ratio Curves with R: Introducing the flexOR Package

The analysis of odds ratio curves is a valuable tool in understanding the relationship between continuous predictors and binary outcomes. Traditional parametric regression approaches often assume specific functional forms, limiting their flexibility and applicability to complex data. To address this limitation and introduce more flexibility, several smoothing methods may be applied, and approaches based on splines are the most frequently considered in this context. To better understand the effects that each continuous covariate has on the outcome, results can be expressed in terms of splines-based odds ratio (OR) curves, taking a specific covariate value as reference. In this paper, we introduce an R package, flexOR, which provides a comprehensive framework for pointwise nonparametric estimation of odds ratio curves for continuous predictors. The package can be used to estimate odds ratio curves without imposing rigid assumptions about their underlying functional form while considering a reference value for the continuous covariate. The package offers various options for automatically choosing the degrees of freedom in multivariable models. It also includes visualization functions to aid in the interpretation and presentation of the estimated odds ratio curves. flexOR offers a user-friendly interface, making it accessible to researchers and practitioners without extensive statistical backgrounds.

Intrinsic Information-Theoretic Models.

With this follow-up paper, we continue developing a mathematical framework based on information geometry for representing physical objects. The long-term goal is to lay down informational foundations for physics, especially quantum physics. We assume that we can now model information sources as univariate normal probability distributions N (μ, σ0), as before, but with a constant σ0 not necessarily equal to 1. Then, we also relaxed the independence condition when modeling m sources of information. Now, we model m sources with a multivariate normal probability distribution Nm(μ,Σ0) with a constant variance-covariance matrix Σ0 not necessarily diagonal, i.e., with covariance values different to 0, which leads to the concept of modes rather than sources. Invoking Schrödinger's equation, we can still break the information into m quantum harmonic oscillators, one for each mode, and with energy levels independent of the values of σ0, altogether leading to the concept of "intrinsic". Similarly, as in our previous work with the estimator's variance, we found that the expectation of the quadratic Mahalanobis distance to the sample mean equals the energy levels of the quantum harmonic oscillator, being the minimum quadratic Mahalanobis distance at the minimum energy level of the oscillator and reaching the "intrinsic" Cramér-Rao lower bound at the lowest energy level. Also, we demonstrate that the global probability density function of the collective mode of a set of m quantum harmonic oscillators at the lowest energy level still equals the posterior probability distribution calculated using Bayes' theorem from the sources of information for all data values, taking as a prior the Riemannian volume of the informative metric. While these new assumptions certainly add complexity to the mathematical framework, the results proven are invariant under transformations, leading to the concept of "intrinsic" information-theoretic models, which are essential for developing physics.

Selecting intervals to optimize the design of observational studies subject to fine balance constraints

Motivated by designing observational studies using matching methods subject to fine balance constraints, we introduce a new optimization problem. This problem consists of two phases. In the first phase, the goal is to cluster the values of a continuous covariate into a limited number of intervals. In the second phase, we find the optimal matching subject to fine balance constraints with respect to the new covariate we obtained in the first phase. We show that the resulting optimization problem is NP-hard. However, it admits an FPT algorithm with respect to a natural parameter. This FPT algorithm also translates into a polynomial time algorithm for the most natural special cases of the problem.

Optimal Predictors of General Small Area Parameters Under an Informative Sample Design Using Parametric Sample Distribution Models

Abstract Two challenges in small area estimation occur when (i) the sample selection mechanism depends on the outcome variable and (ii) the parameter of interest is a nonlinear function of the response variable in the assumed model. If, given the values of the model covariates, the sample selection mechanism depends on the model response variable, the design is said to be informative for the model. Pfeffermann and Sverchkov (2007) develop a small area estimation procedure for informative sampling, focusing on the prediction of small area means. Molina and Rao (2010) develop a small area estimation procedure for general parameters that are nonlinear functions of the model response variable. The method of Molina and Rao assumes noninformative sampling. We combine these two approaches to develop a procedure for the estimation of general parameters in small areas under informative sampling. We introduce a parametric bootstrap MSE estimator that is appropriate for an informative sample design. We evaluate the validity of the proposed procedures through extensive simulation studies and illustrate the procedures utilizing Mexico’s income data.

Abstract 2390: Longitudinal assessment of circulating tumor DNA in patients with advanced colorectal cancer: A proposed general statistical framework and visualization tool

Abstract Background: Understanding how ctDNA levels change over time can act as a surrogate marker for disease progression. Since ctDNA evolution is complicated—exhibiting variability within and between patients—we propose a rigorous and flexible statistical framework that comprehensively represents these types of time-varying biomarkers. We propose employing a hierarchal random effects cubic spline model due to its advantages over traditional longitudinal modeling approaches such as the former’s ability to incorporate patient characteristics and create patient-specific results. Visualization of individual patient trajectories may provide clinical utility in precision oncology settings. Methods: 167 patients with CRC were selected from GuardantINFORM, a real-world database linking genomic and claims data. All patients received chemotherapy and had at least three serial liquid biopsy tests completed via Guardant360. To meet model assumptions, ctDNA levels, measured by maximum variant allele frequency on each test, were transformed into logits. Due to the model’s hierarchal structure, an unconditional cubic spline model was fit first, producing an estimated response pattern for the cohort. Next, as patient-level results are of interest, the unconditional model was built upon by fitting a conditional model that incorporated covariates consisting of demographic, health status, and mortality information, which provided numerous patient-level response patterns. The best fitting conditional model was guided by Akaike’s information criteria. Results: Since model parameter estimates are uninterpretable, and because numerous patient-level projections are generated (each covariate value combination produces a unique projection), an R-Shiny application was developed to visually present and compare results in an intuitive interactive fashion. Additionally, to enhance the understanding of patient response patterns, velocity plots, which provides the instantaneous rate of change in ctDNA levels at different time points, are also provided. Conclusions: We demonstrate that the proposed method can successfully be applied to genomic data to describe and explore complex patient-level temporal ctDNA patterns while accounting for the impact of covariate values have on these patterns. We implemented the proposed methodology as a visualization tool that can be used in a wide variety of settings, ranging from hypothesis testing in clinical trials to patient monitoring. Results from the model can further our basic conceptualization of ctDNA dynamics and enhance our ability to integrate these results into targeted, patient centric, clinical decision-making. Citation Format: Christopher R. Pretz, Jiemin Liao, Leylah Drusbosky, Amar Das. Longitudinal assessment of circulating tumor DNA in patients with advanced colorectal cancer: A proposed general statistical framework and visualization tool [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 2390.

Mortality in patients with Dupuytren's disease in the first 5 years after diagnosis: a population-based survival analysis.

Previous studies suggest that Dupuytren's disease is associated with increased mortality, but most studies failed to account for important confounders. In this population-based cohort study, general practitioners' (GP) data were linked to Statistics Netherlands to register all-cause and disease-specific mortality. Patients with Dupuytren's disease were identified using the corresponding diagnosis code and assessing free-text fields from GP consultations. Multiple imputations were performed to estimate missing values of covariates, followed by 1:7 propensity score matching to balance cases with controls on confounding factors. A frailty proportional hazard model was used to compare mortality between both groups. Out of 209,966 individuals, 2561 patients with Dupuytren's disease were identified and matched to at least four controls. After a median follow-up of 5 years, mortality was found to be actually reduced in patients with Dupuytren's disease. There was no difference in mortality secondary to cancer or cardiovascular disease. Future studies with longer average follow-up using longitudinal data should clarify these associations in the longer term.Level of evidence: III.

Handling Overlapping Asymmetric Data Sets—A Twice Penalized P-Spline Approach

Aims: Overlapping asymmetric data sets are where a large cohort of observations have a small amount of information recorded, and within this group there exists a smaller cohort which have extensive further information available. Missing imputation is unwise if cohort size differs substantially; therefore, we aim to develop a way of modelling the smaller cohort whilst considering the larger. Methods: Through considering traditionally once penalized P-Spline approximations, we create a second penalty term through observing discrepancies in the marginal value of covariates that exist in both cohorts. Our now twice penalized P-Spline is designed to firstly prevent over/under-fitting of the smaller cohort and secondly to consider the larger cohort. Results: Through a series of data simulations, penalty parameter tunings, and model adaptations, our twice penalized model offers up to a 58% and 46% improvement in model fit upon a continuous and binary response, respectively, against existing B-Spline and once penalized P-Spline methods. Applying our model to an individual’s risk of developing steatohepatitis, we report an over 65% improvement over existing methods. Conclusions: We propose a twice penalized P-Spline method which can vastly improve the model fit of overlapping asymmetric data sets upon a common predictive endpoint, without the need for missing data imputation.

Abstract A003: Decomposing racial and ethnic disparities in endometrial cancer survival

Abstract Background: Black-White mortality disparities are well-established and continue to widen over time among women diagnosed with endometrial cancer (EC). Yet, few studies have explored the existence or sources of racial/ethnic disparities in EC survival among Hispanic, American Indian/Alaska Native (AI/AN), Asian, and Native-Hawaiian/Pacific Islander (NH/PI) women. EC survival disparities across racial and ethnic groups have been suggested to be linked to differences in sociodemographic, clinical, and access to care characteristics. We examined how these individual- and area-level factors influence the presence of racial/ethnic disparities among women with EC. Methods: Participants were diagnosed between 2004 and 2019 with stages 1A through 4B endometroid and non-endometroid EC in the National Cancer Database (NCDB). Race was categorized as non-Hispanic white (NHW), non-Hispanic black (NHB), Hispanic, Asian, NH/PI, and AI/AN. We performed Oaxaca-Blinder decompositions of the log overall survival (OS) time assuming an accelerated failure time model with Weibull error, which decomposes observed differences in log survival time into parts potentially explained by differences in measured prognostic factors and differences not explained by measured variables. We examined the following classes of prognostic factors: age at diagnosis, zip code-level demographics (educational attainment, average income, and rurality), access to care (insurance status, facility type and location, treatment delay), Charlson comorbidity score, tumor characteristics (stage and histology), and guideline-concordant treatment (GCT). White women were the comparison group in all decompositions. Results: Compared with NHW women (N=121,069; estimated overall survival for average covariates values (eOS)=19.7 years) eOS was significantly shorter among NHB women (N=15,855; eOS=9.9 years); longer among Asian women (N=5,098; eOS=42.7 years); longer among Hispanic women (N=9,671; eOS=25.9 years); and not significantly different in NH/PI (N=536; eOS=24.8 years) or AI/AN women (N=551; eOS=17.0 years). Measured covariates explained 76.7% of the difference for NHB women (p<0.001); 19.7% for Asian women (p<0.001); and did not significantly explain the difference for Hispanic women (8.4%; p=0.14). Factors significantly contributing to the observed differences for NHB women were tumor characteristics (61.5%); access to care (6.6%); zip code-level demographics (5.1%); comorbidity (3.9%); age (-1.8%); and GCT (1.5%). Significant factors for Asian women were age (16.6%); tumor characteristics (-6.0%); zip code-level demographics (4.7%); access to care (3.0%); and comorbidity (1.6%). Significant factors for Hispanic women were age (45.2%); tumor characteristics (-17.9%); access to care (-13.4%); comorbidity (-2.0%); and GCT (-1.0%). Conclusions: Contributors to EC survival disparities vary by race/ethnicity and strongly implicate the need for research to identify how disparities arise to tailor interventions to reduce survival gaps. Citation Format: Jordyn A. Brown, Jennifer A. Sinnott, Ziyu Gao, Caitlin E. Meade, Macarius M. Donneyong, Tasleem J. Padamsee, Ashley S. Felix. Decomposing racial and ethnic disparities in endometrial cancer survival [abstract]. In: Proceedings of the AACR Special Conference on Endometrial Cancer: Transforming Care through Science; 2023 Nov 16-18; Boston, Massachusetts. Philadelphia (PA): AACR; Clin Cancer Res 2024;30(5_Suppl):Abstract nr A003.

Adaptive unscented Kalman filter methods for identifying time‐variant parameters via state covariance re‐updating

AbstractThe conventional parameter identification process generally assumes that parameters remain constant. However, under extreme loading conditions, structures may exhibit nonlinear behavior, and parameters could demonstrate time‐variant characteristics. The unscented Kalman filter (UKF), as an efficient online recursive estimator, is widely used for identifying parameters of nonlinear systems. Nevertheless, it exhibits limitations when attempting to identify time‐variant parameters. To address this issue, this paper proposes a covariance matching technique that produces an array of adaptive UKF algorithms. Firstly, the sensitivity parameterηis defined to identify the instant when the parameter change occurs, and its threshold is calculated based on the sensitivity parameter time history curve. Secondly, an adaptive forgetting factor is introduced to simultaneously update the innovation, cross, and state covariance matrices when thekth‐step sensitive parameter surpasses the threshold. Finally, a secondary correction forgetting factor (SCFF) is employed to further re‐update the state covariance values at the identified damage locations. This creative step enhances the adaptive capability and optimizes the identification accuracy of the proposed algorithms. Both the numerical simulations and shaking table test demonstrate that the proposed adaptive algorithms can efficiently identify the time‐variant stiffness‐type parameters, and accurately capture their time‐variant characteristics.

Covariate adjustment in Bayesian adaptive randomized controlled trials.

In conventional randomized controlled trials, adjustment for baseline values of covariates known to be at least moderately associated with the outcome increases the power of the trial. Recent work has shown a particular benefit for more flexible frequentist designs, such as information adaptive and adaptive multi-arm designs. However, covariate adjustment has not been characterized within the more flexible Bayesian adaptive designs, despite their growing popularity. We focus on a subclass of these which allow for early stopping at an interim analysis given evidence of treatment superiority. We consider both collapsible and non-collapsible estimands and show how to obtain posterior samples of marginal estimands from adjusted analyses. We describe several estimands for three common outcome types. We perform a simulation study to assess the impact of covariate adjustment using a variety of adjustment models in several different scenarios. This is followed by a real-world application of the compared approaches to a COVID-19 trial with a binary endpoint. For all scenarios, it is shown that covariate adjustment increases power and the probability of stopping the trials early, and decreases the expected sample sizes as compared to unadjusted analyses.

Impact of jittering on raster- and distance-based geostatistical analyses of DHS data

Fine-scale covariate rasters are routinely used in geostatistical models for mapping demographic and health indicators based on household surveys from the Demographic and Health Surveys (DHS) program. However, the geostatistical analyses ignore the fact that GPS coordinates in DHS surveys are jittered for privacy purposes. We demonstrate the need to account for this jittering, and we propose a computationally efficient approach that can be routinely applied. We use the new method to analyse the prevalence of completion of secondary education for 20-49 year old women in Nigeria in 2018 based on the 2018 DHS survey. The analysis demonstrates substantial changes in the estimates of spatial range and fixed effects compared to when we ignore jittering. Through a simulation study that mimics the dataset, we demonstrate that accounting for jittering reduces attenuation in the estimated coefficients for covariates and improves predictions. The results also show that the common approach of averaging covariate values in windows around the observed locations does not lead to the same improvements as accounting for jittering.

An Analysis of the Effect of Streaming on Civic Participation Through a Causal Hidden Markov Model

We examine the effect of streaming based on ability levels on individuals’ civic participation throughout their adult life. The hypothesis we test is that ability grouping influences individuals’ general self-concept and, consequently, their civic participation choices across the life course. We employ data from the British National Child Development Study, which follows all UK citizens born during a certain week in 1958. Six binary variables observed at 33, 42, and 51 years of age are considered to measure civic participation. Our approach defines causal estimands with multiple treatments referring to the evolution of civic engagement over time in terms of potential versions of a sequence of latent variables assumed to follow a Markov chain with initial and transition probabilities depending on posttreatment time-varying covariates. The model also addresses partially or entirely missing data on one or more indicators at a given time occasion and missing posttreatment covariate values using dummy indicators. The model is estimated by maximizing a weighted log-likelihood function with weights corresponding to the inverse probability of the received treatment obtained from a multinomial logit model based on pretreatment covariates. Our results show that ability grouping affects the civic participation of high-ability individuals when they are 33 years old with respect to participation in general elections.

Clustering blood donors via mixtures of product partition models with covariates.

Motivated by the problem of accurately predicting gap times between successive blood donations, we present here a general class of Bayesian nonparametric models for clustering. These models allow for the prediction of new recurrences, accommodating covariate information that describes the personal characteristics of the sample individuals. We introduce a prior for the random partition of the sample individuals, which encourages two individuals to be co-clustered if they have similar covariate values. Our prior generalizes product partition models with covariates (PPMx) models in the literature, which are defined in terms of cohesion and similarity functions. We assume cohesion functions that yield mixtures of PPMx models, while our similarity functions represent the denseness of a cluster. We show that including covariate information in the prior specification improves the posterior predictive performance and helps interpret the estimated clusters in terms of covariates in the blood donation application.

RISK ANALYSIS OF SHARIA STOCKS IN THE INFRASTRUCTURE, UTILITIES AND TRANSPORTATION SECTOR LISTED ON THE JAKARTA ISLAMIC INDEX (JII) 2015-2023

This research is intended to analyze the risk comparison using the Value at Risk (VaR) Variance Covariance and Value at Risk (VaR) Historical Simulation models in the Infrastructure, Utilities and Transportation subsectors. The development of infrastructure, utilities, and transportation plays a very important role in national development and is the main driver of regional growth and the industrial sector. Improvements in the regulatory and investment policy framework are expected to significantly increase the availability of infrastructure facilities and services. The population involved in this study includes 9 companies listed in the JII in the Infrastructure, Utilities, and Transportation sector, which are used as samples. The data used is secondary data obtained from www.yahoo.finance.com. Data analysis was carried out using a two-sample average test. The test results of the Value at Risk (VaR) Historical Simulation and Variance Covariance values are largest in INDX stocks, followed by other stocks, and those with the lowest risk level are TLKM stocks for the upcoming 5-day, 7-day, and 15-day periods with good values for alpha 1%, 5%, and 10%. The results concluded that the comparison between VaR Variance Covariance and VaR Historical Simulation produces a good level of risk and helps in determining sectors that are worth investing in the future, and can describe the fundamental strength of each sector.

Factors associated with disease control failure in acromegaly patients treated with pegvisomant: an ACROSTUDY analysis.

The aim of this study was to examine the probability of achieving acromegaly disease control according to several patient-, disease- and treatment-related factors longitudinally. We analyzed data from ACROSTUDY, an open-label, noninterventional, post-marketing safety surveillance study conducted in 15 countries. A total of 1546 patients with acromegaly and treated with pegvisomant, with available information on baseline IGF-1 level, were included. Factors influencing IGF-1 control were assessed up to 10 years of follow-up by mixed-effects logistic regression models, taking into account changing values of covariates at baseline and at yearly visits. Twenty-eight anthropometric, clinical and treatment-related covariates were examined through univariate and multivariate analyses. We tested whether the probability of non-control was different than 0.50 (50%) by computing effect sizes (ES) and the corresponding 95% CI. Univariate analysis showed that age <40 years, normal or overweight, baseline IGF-1 <300 µg/L or ranged between 300 and 500 µg/L, and all pegvisomant dose <20 mg/day were associated with a lower probability of acromegaly uncontrol. Consistently, in multivariate analyses, the probability of uncontrolled acromegaly was influenced by baseline IGF-1 value: patients with IGF-1 <300 µg/L had the lowest risk of un-controlled acromegaly (ES = 0.29, 95% CI: 0.23-0.36). The probability of acromegaly uncontrol was also lower for values 300-500 µg/L (ES = 0.37, 95% CI: 0.32-0.43), while it was higher for baseline IGF-1 values ≥700 µg/L (ES = 0.58, 95% CI: 0.53-0.64). Baseline IGF-l levels were a good predictor factor for long-term acromegaly control. On the contrary, our data did not support a role of age, sex, BMI and pegvisomant dose as predictors of long-term control of acromegaly. Among factors that could influence and predict the efficacy of pegvisomant therapy in controlling acromegaly, a central role of baseline IGF-1 values on the probability of achieving a biochemical control of acromegaly during the treatment with pegvisomant was identified, in a real-life setting.

Confidence sets for a level set in linear regression.

Regression modeling is the workhorse of statistics and there is a vast literature on estimation of the regression function. It has been realized in recent years that in regression analysis the ultimate aim may be the estimation of a level set of the regression function, ie, the set of covariate values for which the regression function exceeds a predefined level, instead of the estimation of the regression function itself. The published work on estimation of the level set has thus far focused mainly on nonparametric regression, especially on point estimation. In this article, the construction of confidence sets for the level set of linear regression is considered. In particular, level upper, lower and two-sided confidence sets are constructed for the normal-error linear regression. It is shown that these confidence sets can be easily constructed from the corresponding level simultaneous confidence bands. It is also pointed out that the construction method is readily applicable to other parametric regression models where the mean response depends on a linear predictor through a monotonic link function, which include generalized linear models, linear mixed models and generalized linear mixed models. Therefore, the method proposed in this article is widely applicable. Simulation studies with both linear and generalized linear models are conducted to assess the method and real examples are used to illustrate the method.

Multiple imputation in the functional linear model with partially observed covariate and missing values in the response

. Missing data problems are common and difficult to handle in data analysis. Ad hoc methods, such as simply removing cases with missing values, can lead to invalid analysis results. In this article, we consider a functional linear regression model with partially observed covariate and missing values in the response. We use a reconstruction operator that aims at recovering the missing parts of the explanatory curves, then we are interested in regression imputation method of missing data on the response variable, using functional principal component regression to estimate the functional coefficient of the model. We study the asymptotic behavior of the prediction error when missing values in an original dataset are imputed by multiple sets of plausible values. The method behavior is also evaluated in practice.

Analyzing the Nexus between Personal Values and Consumption Values of Consumers’ Preference for Fresh Fish

This study aimed to investigate the nexus between personal values and consumption values towards consumers' preferences for fresh fish. A cross-sectional data of 300 respondents was selected using a multi-stage random sampling procedure. A well-structured questionnaire and personal interview were used to collect data from the respondents. Factor analysis and canonical correlation analysis were conducted to achieve the study’s objectives. Findings showed that females (67.5%) were the majority of respondents, household size ranged between 4-6 persons, and many (76.7%) were married. The mean age was 41 years, and 83.3% were formally educated. The relationship between personal values and consumption values was positive and statistically significant at the 1% level. The study revealed that variables like benevolence, security and self-direction were strongly correlated with emotional and functional values of the first canonical covariates. The study concludes that high benevolence, security and self-direction evoke high emotional and functional values when consumers buy fresh fish. Based on these findings, the study recommends that fresh fish marketers should pay more attention to the price-quality relationship, the performance and content of the product and the quality of packaging, as most respondents place more value on what they consume.