Simultaneous monitoring for regression coefficients and baseline hazard profile in Cox modeling of time-to-event data.

The Cox model is the most popular tool for analyzing time-to-event data. The nonparametric baseline hazard function can be as important as the regression coefficients in practice, especially when prediction is needed. In the context of stochastic process control, we propose a simultaneous monitoring method that combines a multivariate control chart for the regression coefficients and a profile control chart for the cumulative baseline hazard function that allows for data blocks of possibly different censoring rates and sample sizes. The method can detect changes in either the parametric or the nonparametric part of the Cox model. In simulation studies, the proposed method maintains its size and has substantial power in detecting changes in either part of the Cox model. An application in lymphoma survival analysis in which patients were enrolled by 2-month intervals in the Surveillance, Epidemiology, and End Results program identifies data blocks with structural model changes.

Intuitive eating buffers the link between internalized weight stigma and body mass index in stressed adults.

Internalized weight stigma (IWS) is independently associated with less intuitive eating (i.e., eating based on endogenous hunger/satiety cues) and higher Body Mass Index (BMI), and intuitive eating training is commonly conceptualized as protective against the effects of IWS on poor behavioral health. The 3-way relationship between IWS, intuitive eating, and BMI has yet to be examined, and it is unclear whether the link between IWS and BMI is buffered by high intuitive eating. This secondary preliminary analysis examined baseline data of stressed adults with poor diet (N = 75, 70% female, 64.1% White, 42.7% with overweight/obesity) in a parent clinical trial that tested the effects of yoga on diet and stress. Validated self-report surveys of IWS and intuitive eating were analyzed with objectively-assessed BMI. Moderated regression analyses using the SPSS PROCESS macro tested whether intuitive eating moderated the IWS-BMI link. The analysis revealed IWS was positively associated with BMI except among people with high intuitive eating. Results extend observational findings linking intuitive eating to lower BMI, and offer preliminary support for the hypothesis that this link may hold even among those with greater IWS. It's possible that individuals with lower BMI and greater IWS may gravitate more towards intuitive eating than those with greater BMI, and/or intuitive eating may be an important target for ameliorating the adverse association of IWS with behavioral and physical health indicators linked to BMI. Continued work is warranted in larger, more generalizable samples using causal and prospective designs.

An online updating approach for testing the proportional hazards assumption with streams of survival data.

The Cox model-which remains the first choice for analyzing time-to-event data, even for large data sets-relies on the proportional hazards (PH) assumption. When survival data arrive sequentially in chunks, a fast and minimally storage intensive approach to test the PH assumption is desirable. We propose an online updating approach that updates the standard test statistic as each new block of data becomes available and greatly lightens the computational burden. Under the null hypothesis of PH, the proposed statistic is shown to have the same asymptotic distribution as the standard version computed on an entire data stream with the data blocks pooled into one data set. In simulation studies, the test and its variant based on most recent data blocks maintain their sizes when the PH assumption holds and have substantial power to detect different violations of the PH assumption. We also show in simulation that our approach can be used successfully with "big data" that exceed a single computer's computational resources. The approach is illustrated with the survival analysis of patients with lymphoma cancer from the Surveillance, Epidemiology, and End Results Program. The proposed test promptly identified deviation from the PH assumption, which was not captured by the test based on the entire data.

A new Bayesian joint model for longitudinal count data with many zeros, intermittent missingness, and dropout with applications to HIV prevention trials.

In longitudinal clinical trials, it is common that subjects may permanently withdraw from the study (dropout), or return to the study after missing one or more visits (intermittent missingness). It is also routinely encountered in HIV prevention clinical trials that there is a large proportion of zeros in count response data. In this paper, a sequential multinomial model is adopted for dropout and subsequently a conditional model is constructed for intermittent missingness. The new model captures the complex structure of missingness and incorporates dropout and intermittent missingness simultaneously. The model also allows us to easily compute the predictive probabilities of different missing data patterns. A zero-inflated Poisson mixed-effects regression model is assumed for the longitudinal count response data. We also propose an approach to assess the overall treatment effects under the zero-inflated Poisson model. We further show that the joint posterior distribution is improper if uniform priors are specified for the regression coefficients under the proposed model. Variations of the g-prior, Jeffreys prior, and maximally dispersed normal prior are thus established as remedies for the improper posterior distribution. An efficient Gibbs sampling algorithm is developed using a hierarchical centering technique. A modified logarithm of the pseudomarginal likelihood and a concordance based area under the curve criterion are used to compare the models under different missing data mechanisms. We then conduct an extensive simulation study to investigate the empirical performance of the proposed methods and further illustrate the methods using real data from an HIV prevention clinical trial.

Bayesian Modeling and Inference for Nonignorably Missing Longitudinal Binary Response Data with Applications to HIV Prevention Trials.

Missing data are frequently encountered in longitudinal clinical trials. To better monitor and understand the progress over time, one must handle the missing data appropriately and examine whether the missing data mechanism is ignorable or nonignorable. In this article, we develop a new probit model for longitudinal binary response data. It resolves a challenging issue for estimating the variance of the random effects, and substantially improves the convergence and mixing of the Gibbs sampling algorithm. We show that when improper uniform priors are specified for the regression coefficients of the joint multinomial model via a sequence of one-dimensional conditional distributions for the missing data indicators under nonignorable missingness, the joint posterior distribution is improper. A variation of Jeffreys prior is thus established as a remedy for the improper posterior distribution. In addition, an efficient Gibbs sampling algorithm is developed using a collapsing technique. Two model assessment criteria, the deviance information criterion (DIC) and the logarithm of the pseudomarginal likelihood (LPML), are used to guide the choices of prior specifications and to compare the models under different missing data mechanisms. We report on extensive simulations conducted to investigate the empirical performance of the proposed methods. The proposed methodology is further illustrated using data from an HIV prevention clinical trial.

Online updating method with new variables for big data streams.

For big data arriving in streams, online updating is an important statistical method that breaks the storage barrier and the computational barrier under certain circumstances. In the regression context, online updating algorithms assume that the set of predictor variables does not change, and consequently cannot incorporate new variables that may become available midway through the data stream. A naive approach would be to discard all previous information and start updating with new variables from scratch. We propose a method that utilizes the information from earlier data in the online updating algorithm with bias corrections to improve efficiency. The method is developed for linear models first, and then extended to estimating equations for generalized linear models. Closed-form expressions for the efficiency gain over the naive approach are derived in a particular linear model setting. We compare the performance of our proposed bias-correcting approach and the naive approach in simulation studies with data generated from a normal linear model and a logistic regression model. The method is applied to a study on airline delay, where reasons for delays were only available more recently, starting in 2003.

Weighted pseudolikelihood for SNP set analysis with multiple secondary outcomes in case-control genetic association studies.

We propose a weighted pseudolikelihood method for analyzing the association of a SNP set, example, SNPs in a gene or a genetic pathway or network, with multiple secondary phenotypes in case-control genetic association studies. To boost analysis power, we assume that the SNP-specific effects are shared across all secondary phenotypes using a scaled mean model. We estimate regression parameters using Inverse Probability Weighted (IPW) estimating equations obtained from the weighted pseudolikelihood, which accounts for case-control sampling to prevent potential ascertainment bias. To test the effect of a SNP set, we propose a weighted variance component pseudo-score test. We also propose a penalized IPW pseudolikelihood method for selecting a subset of SNPs that are associated with the multiple secondary phenotypes. We show that the proposed variable selection procedure has the oracle properties and is robust to misspecification of the correlation structure among secondary phenotypes. We select the tuning parameter using a weighted Bayesian Information-like Criterion (wBIC). We evaluate the finite sample performance of the proposed methods via simulations, and illustrate the methods by the analysis of the multiple secondary smoking behavior outcomes in a lung cancer case-control genetic association study.

Health Care Transition for Young Adults With Type 1 Diabetes: Stakeholder Engagement for Defining Optimal Outcomes.

Objectives: Research on the transition to adult care for young adults with type 1 diabetes (T1D) emphasizes transition readiness, with less emphasis on transition outcomes. The relatively few studies that focus on outcomes use a wide variety of measures with little reliance on stakeholder engagement for measure selection. Methods: This study engaged multiple stakeholders (i.e., young adults with T1D, parents, pediatric and adult health care providers, and experts) in qualitative interviews to identify the content domain for developing a multidimensional measure of health care transition (HCT) outcomes. Results: The following constructs were identified for a planned measure of HCT outcomes: biomedical markers of T1D control; T1D knowledge/skills; navigation of a new health care system; integration of T1D into emerging adult roles; balance of parental involvement with autonomy; and "ownership" of T1D self-management. Discussion: The results can guide creation of an initial item pool for a multidimensional profile of HCT outcomes.

The Ups and Downs of Parenting Young Children With Type 1 Diabetes: A Crowdsourcing Study.

Objectives: Parenting young children with type 1 diabetes (YC-T1D) entails pervasive challenges; parental coping may influence child and parent outcomes. This study used a qualitative descriptive design to describe these challenges comprehensively to inform the user-centered design of an Internet coping resource for parents. Methods: A "Parent Crowd" of 153 parents of children with T1D onset at ≤ 5 years old submitted textual responses online to open-ended questions about parenting YC-T1D. Systematic coding organized responses into domains, themes, and examples. A supplemental focus group of racial/ethnic minority parents enhanced the sample's diversity and validated findings from the Parent Crowd. Results: Similar domains and themes emerged from responses of crowdsourcing and focus group participants. In each domain, parenting YC-T1D was challenging, but there was also substantial evidence of positive coping strategies and adaptability. Conclusions: The study yielded rich data to inform user-centered design of an Internet resource for parents of YC-T1D.

Genome-wide association analysis for multiple continuous secondary phenotypes.

There is increasing interest in the joint analysis of multiple phenotypes in genome-wide association studies (GWASs), especially for the analysis of multiple secondary phenotypes in case-control studies and in detecting pleiotropic effects. Multiple phenotypes often measure the same underlying trait. By taking advantage of similarity across phenotypes, one could potentially gain statistical power in association analysis. Because continuous phenotypes are likely to be measured on different scales, we propose a scaled marginal model for testing and estimating the common effect of single-nucleotide polymorphism (SNP) on multiple secondary phenotypes in case-control studies. This approach improves power in comparison to individual phenotype analysis and traditional multivariate analysis when phenotypes are positively correlated and measure an underlying trait in the same direction (after transformation) by borrowing strength across outcomes with a one degree of freedom (1-DF) test and jointly estimating outcome-specific scales along with the SNP and covariate effects. To account for case-control ascertainment bias for the analysis of multiple secondary phenotypes, we propose weighted estimating equations for fitting scaled marginal models. This weighted estimating equation approach is robust to departures from normality of continuous multiple phenotypes and the misspecification of within-individual correlation among multiple phenotypes. Statistical power improves when the within-individual correlation is correctly specified. We perform simulation studies to show the proposed 1-DF common effect test outperforms several alternative methods. We apply the proposed method to investigate SNP associations with smoking behavior measured with multiple secondary smoking phenotypes in a lung cancer case-control GWAS and identify several SNPs of biological interest.

A-clustering: a novel method for the detection of co-regulated methylation regions, and regions associated with exposure.

MOTIVATION: DNA methylation is a heritable modifiable chemical process that affects gene transcription and is associated with other molecular markers (e.g. gene expression) and biomarkers (e.g. cancer or other diseases). Current technology measures methylation in hundred of thousands, or millions of CpG sites throughout the genome. It is evident that neighboring CpG sites are often highly correlated with each other, and current literature suggests that clusters of adjacent CpG sites are co-regulated. RESULTS: We develop the Adjacent Site Clustering (A-clustering) algorithm to detect sets of neighboring CpG sites that are correlated with each other. To detect methylation regions associated with exposure, we propose an analysis pipeline for high-dimensional methylation data in which CpG sites within regions identified by A-clustering are modeled as multivariate responses to environmental exposure using a generalized estimating equation approach that assumes exposure equally affects all sites in the cluster. We develop a correlation preserving simulation scheme, and study the proposed methodology via simulations. We study the clusters detected by the algorithm on high dimensional dataset of peripheral blood methylation of pesticide applicators. AVAILABILITY: We provide the R package Aclust that efficiently implements the A-clustering and the analysis pipeline, and produces analysis reports. The package is found on http://www.hsph.harvard.edu/tamar-sofer/packages/ CONTACT: tsofer@hsph.harvard.edu

Mismatch between asthma symptoms and spirometry: implications for managing asthma in children.

OBJECTIVES: To examine the concordance between spirometry and asthma symptoms in assessing asthma severity and beginning therapy by the general pediatrician. STUDY DESIGN: Between 2008 and 2012, spirometry testing was satisfactorily performed in 894 children (ages 5-19 years) whose asthma severity had been determined by their pediatrician using asthma guideline-based clinical criteria. Spirometry-determined asthma severity using national asthma guidelines and clinician-determined asthma severity were compared for concordance using weighted Kappa coefficients. RESULTS: Thirty percent of participants had clinically determined intermittent asthma; 32%, 33%, and 5% had mild, moderate, and severe, persistent asthma, respectively. Increasing disease severity was associated with decreases in the forced expiratory volume in 1 second/forced vital capacity (FVC) ratio (P < .001), the forced expiratory volume in 1 second/FVC% predicted (P < .0001), and the FVC% predicted (P < .01). In 319 children (36%), clinically determined asthma severity was lower than spirometry-determined severity. Concordance was 0.16 (95% CI 0.10, 0.23), and when adjusted for bias and prevalence, was 0.20 (95% CI 0.17, 0.23). When accounting for age, sex, exposure to smoke, and insurance type, only spirometry-determined asthma severity was a significant predictor of agreement (P < .0001), with worse agreement as spirometry-determined severity increased. CONCLUSIONS: Concordance between spirometry and asthma symptoms in determining asthma severity is low even when guideline-based clinical assessment tools are used. Because appropriate therapy reduces asthma morbidity and is guided by disease severity, results from spirometry testing could better guide pediatricians in determining appropriate therapy for their patients with asthma.

SNP set association analysis for familial data.

Genome-wide association studies (GWAS) are a popular approach for identifying common genetic variants and epistatic effects associated with a disease phenotype. The traditional statistical analysis of such GWAS attempts to assess the association between each individual single-nucleotide polymorphism (SNP) and the observed phenotype. Recently, kernel machine-based tests for association between a SNP set (e.g., SNPs in a gene) and the disease phenotype have been proposed as a useful alternative to the traditional individual-SNP approach, and allow for flexible modeling of the potentially complicated joint SNP effects in a SNP set while adjusting for covariates. We extend the kernel machine framework to accommodate related subjects from multiple independent families, and provide a score-based variance component test for assessing the association of a given SNP set with a continuous phenotype, while adjusting for additional covariates and accounting for within-family correlation. We illustrate the proposed method using simulation studies and an application to genetic data from the Genetic Epidemiology Network of Arteriopathy (GENOA) study.

Online Updating of Survival Analysis.

When large amounts of survival data arrive in streams, conventional estimation methods become computationally infeasible since they require access to all observations at each accumulation point. We develop online updating methods for carrying out survival analysis under the Cox proportional hazards model in an online-update framework. Our methods are also applicable with time-dependent covariates. Specifically, we propose online-updating estimators as well as their standard errors for both the regression coefficients and the baseline hazard function. Extensive simulation studies are conducted to investigate the empirical performance of the proposed estimators. A large colon cancer data set from the Surveillance, Epidemiology, and End Results (SEER) program and a large venture capital (VC) data set with time-dependent covariates are analyzed to demonstrate the utility of the proposed methodologies.

An analysis of the perception of stop consonants in bilinguals and monolinguals in different phonetic contexts: A range-based language cueing approach.

Bilinguals' observed perceptual shift across language contexts for shared acoustic properties between their languages supports the idea that bilinguals, but not monolinguals, develop two phonemic representations for the same acoustic property. This phenomenon is known as the double phonemic boundary. This investigation replicated previous findings of bilinguals' double phonemic boundary across a series of go/no-go tasks while controlling for known confounding effects in speech perception (i.e., contrast effects) and differences in resource allocation between bilinguals and monolinguals (i.e., left-hand or right-hand response). Using a range-base language cueing approach, we designed 2 experiments. The first experiment tested whether a voice onset time (VOT) range representative of either Spanish or English phonetic categories can cue bilinguals, but not monolinguals, to use language-specific perceptual routines. The second experiment tested a VOT range with a mixture of Spanish and English phonetic categories to determine whether directing attention to a specific phonetic category can disambiguate the competition of the nonattended category. The results for Experiment 1 showed that bilinguals can rely on the distributional patterns of their native phonetic categories to activate specific language modes. Experiment 2 showed that attention can change the weight given to a native phonetic distinction. However, this process is restricted by the internal phonetic composition of the native language(s).

Yoga participation associated with changes in dietary patterns and stress: A pilot study in stressed adults with poor diet.

BACKGROUND AND PURPOSE: Stress contributes to dietary patterns that impede health. Yoga is an integrative stress management approach associated with improved dietary patterns in burgeoning research. Yet, no research has examined change in dietary patterns, body mass index (BMI), and stress during a yoga intervention among stressed adults with poor diet. MATERIALS AND METHODS: Objectively-measured BMI and a battery of self-report questionnaires were collected at four time points during and following a 12-week yoga intervention (N = 78, 71% women, mean BMI = 25.69 kg/m2±4.59) - pre-treatment (T1), mid-treatment (6 weeks; T2), post-treatment (12 weeks; T3), and at 3-month follow-up (24 weeks; T4). RESULTS: T1 to T3 fruit and vegetable intake, BMI, and stress significantly declined in the overall sample. Reduction in vegetable intake was no longer significant after accounting for reductions in caloric intake, and reduction in caloric intake remained significant after accounting for reductions in stress. CONCLUSION: Findings may be interpreted as yoga either encouraging or adversely impacting healthy dietary patterns (i.e., minimizing likelihood of future weight gain vs. decreasing vegetable intake and overall caloric intake among individuals who may not need to lose weight, respectively). Continued research is warranted, utilizing causal designs.

Internalized weight stigma and intuitive eating among stressed adults during a mindful yoga intervention: associations with changes in mindfulness and self-compassion.

PURPOSE: Internalized weight stigma (IWS) is common in the United States of America across body weight categories, and is implicated in the development of distress and unhealthy eating behaviors (e.g. overeating, disordered eating) that can foster poor cardiometabolic health. While emerging intervention research shows early promise in reducing IWS, long-term efficacy is unclear and novel strategies remain needed. This analysis examined whether participation in a mindful yoga intervention was associated with reduced IWS and increased intuitive eating, an adaptive eating behavior, and whether these changes correlated with each other or with changes in mindfulness and self-compassion. METHODS: Participants were stressed adults with low fruit and vegetable intake (N = 78, 64.1% White, M. Body Mass Index 25.59 ± 4.45) enrolled in a parent clinical trial of a 12-week mindful yoga intervention. Validated self-report measures of IWS, intuitive eating, mindfulness, and self-compassion were administered at pre-treatment, mid-treatment (8 weeks), post-treatment (12 weeks), and 4-month follow-up (24 weeks). RESULTS: Linear mixed modeling revealed significant improvements in IWS and intuitive eating across the four timepoints (p < .001). Reduced IWS correlated with increased intuitive eating pre- to post-treatment (p = .01). Improved self-compassion and mindfulness correlated with intuitive eating (both p = . 04), but not IWS (p = .74 and p = .56, respectively). CONCLUSION: This study offers preliminary support for the hypothesis that mindful yoga may promote intuitive eating and reduce IWS among stressed adults with poor diet, and suggests that changes in these factors may co-occur over time. Further investigation with controlled designs is necessary to better understand the temporality and causality of these relationships.Trial registration: ClinicalTrials.gov identifier: NCT02098018.

BinomiRare: A robust test for association of a rare genetic variant with a binary outcome for mixed models and any case-control proportion.

Whole-genome sequencing (WGS) and whole-exome sequencing studies have become increasingly available and are being used to identify rare genetic variants associated with health and disease outcomes. Investigators routinely use mixed models to account for genetic relatedness or other clustering variables (e.g., family or household) when testing genetic associations. However, no existing tests of the association of a rare variant with a binary outcome in the presence of correlated data control the type 1 error where there are (1) few individuals harboring the rare allele, (2) a small proportion of cases relative to controls, and (3) covariates to adjust for. Here, we address all three issues in developing a framework for testing rare variant association with a binary trait in individuals harboring at least one risk allele. In this framework, we estimate outcome probabilities under the null hypothesis and then use them, within the individuals with at least one risk allele, to test variant associations. We extend the BinomiRare test, which was previously proposed for independent observations, and develop the Conway-Maxwell-Poisson (CMP) test and study their properties in simulations. We show that the BinomiRare test always controls the type 1 error, while the CMP test sometimes does not. We then use the BinomiRare test to test the association of rare genetic variants in target genes with small-vessel disease (SVD) stroke, short sleep, and venous thromboembolism (VTE), in whole-genome sequence data from the Trans-Omics for Precision Medicine (TOPMed) program.

Family consequences of potentially traumatic pediatric medical events: Implications for trauma-informed care.

Extensive research demonstrates that pediatric medical events can be traumatic for patients, caregivers, and siblings, but the aftereffects of these potentially traumatic events for the family and its members are not well documented. Through focus groups with patients, caregivers, and siblings, this qualitative study examined the perceived consequences of potentially traumatic medical events for individual family members and the family as a whole. Sixteen focus groups (6 caregiver, 5 patient, 5 sibling) were conducted. Participants included 44 caregivers, 24 patients, and 14 siblings from 28 families with children treated in cardiology, endocrinology, oncology, orthopedics, or pulmonology. Constant-comparison and directed-content analysis were used to analyze the resulting data. Six themes regarding the family consequences of potentially traumatic medical events emerged: (a) family members experience strong emotional reactions and distressing thoughts, (b) family members experience trauma-related reactions and behaviors, (c) family patterns and routines change, (d) family conflict arises, (e) family members feel different from their peers and strive for normalcy, and (f) family members construct positive narratives about these events and experience positive consequences and emotions. These findings reveal the consequences of potentially traumatic medical events that extend beyond traumatic stress symptoms. Moreover, the impact of these consequences is seen within both individual family member responses and responses within the family system as a whole. Understanding both individual- and family-level consequences of medical events is important in order to provide family-centered, trauma-informed care for children with illness or injury and their family members. (PsycINFO Database Record (c) 2020 APA, all rights reserved).

Development and content validation of the Healthcare Transition Outcomes Inventory for young adults with type 1 diabetes.

BACKGROUND: The literature on the specification and measurement of the outcomes of the healthcare transition from pediatric to adult centered-care is scarce and methodologically weak. To address these gaps, we conducted a series of studies to develop a multidimensional, multi-informant (young adults, parents, and healthcare providers) measure of healthcare transition outcomes for young adults with type 1 diabetes (T1D), the Healthcare Transition Outcomes Inventory (HCTOI). The current study describes the development and refinement of the HCTOI item pool. METHODS: Following Patient Reported Outcomes Measurement Information System (PROMIS) standards, the research team conducted qualitative interviews to define six content domains of healthcare transition outcomes from the perspectives of multiple stakeholders, developed an initial item pool of the HCTOI based on the six domains, analyzed expert item ratings and feedback for content validation, and conducted cognitive interviews with informants (patients, parents, and healthcare providers) for further item pool refinement. RESULTS: Qualitative findings revealed six healthcare transition outcome domains: 1) Biomedical markers of T1D control; 2) Navigation of a new health care system; 3) Possession of T1D self-management skills and knowledge; 4) Integration of T1D care into emerging adult roles; 5) Balance of parental involvement with autonomy; and 6) Attainment of T1D "ownership." An initial pool of 88 items focused on the extent to which a young adult with T1D is successful on each of the six domains. Experts rated all content domains and all but six items as relevant. In addition to suggesting additional items, experts were concerned about the length of the measure, response burden, and whether every informant type would have sufficient knowledge to rate items in particular content domains. Cognitive interviews resulted in retaining all six content domains, but dropping some items and yielded fewer items for the healthcare provider version (47 items versus 54 items for the young adult- and parent-versions). CONCLUSIONS: Expert review and cognitive interviews confirmed that all six domains of HCT outcomes were relevant and both procedures resulted in retaining a sufficient number of clear and representative items for each content domain. The HCTOI represents the first multi-informant, rigorously developed item pool that comprehensively measures the multiple components of the transition from pediatric to adult specialty healthcare.