SpaceANOVA: Spatial Co-occurrence Analysis of Cell Types in Multiplex Imaging Data Using Point Process and Functional ANOVA.

Multiplex imaging platforms have enabled the identification of the spatial organization of different types of cells in complex tissue or the tumor microenvironment. Exploring the potential variations in the spatial co-occurrence or colocalization of different cell types across distinct tissue or disease classes can provide significant pathological insights, paving the way for intervention strategies. However, the existing methods in this context either rely on stringent statistical assumptions or suffer from a lack of generalizability. We present a highly powerful method to study differential spatial co-occurrence of cell types across multiple tissue or disease groups, based on the theories of the Poisson point process and functional analysis of variance. Notably, the method accommodates multiple images per subject and addresses the problem of missing tissue regions, commonly encountered due to data-collection complexities. We demonstrate the superior statistical power and robustness of the method in comparison with existing approaches through realistic simulation studies. Furthermore, we apply the method to three real data sets on different diseases collected using different imaging platforms. In particular, one of these data sets reveals novel insights into the spatial characteristics of various types of colorectal adenoma.

Phthalate and phthalate replacement concentrations in relationship to adiposity in a multi-racial cohort of children.

BACKGROUND/OBJECTIVE: Phthalates and phthalate replacements are used in multiple everyday products, making many of them bioavailable to children. Experimental studies suggest that phthalates and their replacements may be obesogenic, however, epidemiologic studies remain inconsistent. Therefore, our objective was to examine the association between phthalates, phthalate replacements and childhood adiposity/obesity markers in children. SUBJECTS/METHODS: A cross-sectional study was conducted in 630 racial/ethnically diverse children ages 4-8 years. Urinary oxidative metabolites of DINCH and DEHTP, three low molecular weight (LMW) phthalates, and eleven high molecular weight (HMW) phthalates were measured. Weight, height, waist circumference and % body fat were measured. Composite molar sum groups (nmol/ml) were natural log-transformed. Linear regression models adjusted for urine specific gravity, sex, age, race-ethnicity, birthweight, breastfeeding, reported activity level, mother's education and pre-pregnancy BMI. RESULTS: All children had LMW and HMW phthalate metabolites and 88% had DINCH levels above the limit of detection. One unit higher in the log of DINCH was associated with 0.106 units lower BMI z-score [ß = -0.106 (95% CI: -0.181, -0.031)], 0.119 units lower waist circumference z-score [ß = -0.119 (95% CI: -0.189, -0.050)], and 0.012 units lower percent body fat [ß = -0.012 (95% CI: -0.019, -0.005)]. LMW and HMW group values were not associated with adiposity/obesity. CONCLUSIONS: We report an inverse association between child urinary DINCH levels, a non-phthalate plasticizer that has replaced DEHP in several applications, and BMI z-score, waist circumference z-score and % body fat in children. Few prior studies of phthalates and their replacements in children have been conducted in diverse populations. Moreover, DINCH has not received a great deal of attention or regulation, but it is a common exposure. In summary, understanding the ubiquitous nature of these chemical exposures and ultimately their sources will contribute to our understanding of their relationship with obesity.

A Bayesian zero-inflated beta-binomial model for longitudinal data with group-specific changepoints.

Timeline followback (TLFB) is often used in addiction research to monitor recent substance use, such as the number of abstinent days in the past week. TLFB data usually take the form of binomial counts that exhibit overdispersion and zero inflation. Motivated by a 12-week randomized trial evaluating the efficacy of varenicline tartrate for smoking cessation among adolescents, we propose a Bayesian zero-inflated beta-binomial model for the analysis of longitudinal, bounded TLFB data. The model comprises a mixture of a point mass that accounts for zero inflation and a beta-binomial distribution for the number of days abstinent in the past week. Because treatment effects appear to level off during the study, we introduce random changepoints for each study group to reflect group-specific changes in treatment efficacy over time. The model also includes fixed and random effects that capture group- and subject-level slopes before and after the changepoints. Using the model, we can accurately estimate the mean trend for each study group, test whether the groups experience changepoints simultaneously, and identify critical windows of treatment efficacy. For posterior computation, we propose an efficient Markov chain Monte Carlo algorithm that relies on easily sampled Gibbs and Metropolis-Hastings steps. Our application shows that the varenicline group has a short-term positive effect on abstinence that tapers off after week 9.

Bayesian hierarchical profile regression for binary covariates.

Dysphagia, a common result of other medical conditions, is caused by malfunctions in swallowing physiology resulting in difficulty eating and drinking. The Modified Barium Swallow Study (MBSS), the most commonly used diagnostic tool for evaluating dysphagia, can be assessed using the Modified Barium Swallow Impairment Profile (MBSImP™). The MBSImP assessment tool consists of a hierarchical grouped data structure with multiple domains, a set of components within each domain which characterize specific swallowing physiologies, and a set of tasks scored on a discrete scale within each component. We lack sophisticated approaches to extract patterns of physiologic swallowing impairment from the MBSImP task scores within a component while still recognizing the nested structure of components within a domain. We propose a Bayesian hierarchical profile regression model, which uses a Bayesian profile regression model in conjunction with a hierarchical Dirichlet process mixture model to (1) cluster subjects into impairment profile patterns while respecting the hierarchical grouped data structure of the MBSImP, and (2) simultaneously determine associations between latent profile cluster membership for all components and the outcome of dysphagia severity. We apply our approach to a cohort of patients referred for an MBSS and assessed using the MBSImP. Our research results can be used to inform appropriate intervention strategies, and provide tools for clinicians to make better multidimensional management and treatment decisions for patients with dysphagia.

Children's physical activity in family child care homes: Influence of quality status, environment and policy features, and child characteristics.

BACKGROUND: About half of preschool-age children are not meeting recommendations of 15 min/h of physical activity (PA), and nearly one out of seven children between the ages of 2-5 years are living with obesity. Furthermore, children attending family child care homes (FCCHs), compared with larger child care centers, engage in lower levels of PA and appear to be at a higher risk of obesity. Therefore, examining PA and multi-level factors that influence PA in children who attend FCCHs is essential. METHODS: The Childcare Home Eating and Exercise Study (CHEER) examined PA behaviors of 184 children enrolled in 56 FCCHs and FCCH quality status, environment and policy features, and child characteristics. PA was assessed by accelerometer, and FCCH environment and policy was assessed via structured observation. Multiple linear regression was used to model associations between school day total PA and FCCH quality status, environment and policy features, and child characteristics. RESULTS: Child participants were on average 3.1 years old; participants were non-Hispanic Black (47.3%), Non-Hispanic White (42.9%), other race/ethnicity (7.1%), and Hispanic/Latin (2.7%). Children in FCCH settings participated in 11.2 min/h of total PA, which is below the recommended 15 min per hour. The PA environment and policy observation yielded a score of 11.8 out of a possible 30, which is not supportive of child PA. There were no associations between total child PA and FCCH quality status, environment and policy features, and child characteristics in these FCCH settings. CONCLUSIONS: This study was unique in its examination of PA and a comprehensive set of factors that may influence PA at the individual, organizational, environmental, and policy levels in a diverse sample of children attending FCCHs in South Carolina. Additional research is needed to better understand how to increase children's physical activity while they are in the FCCH setting. This research should use multi-level frameworks and apply longitudinal study designs.

A Bayesian multivariate mixture model for high throughput spatial transcriptomics.

High throughput spatial transcriptomics (HST) is a rapidly emerging class of experimental technologies that allow for profiling gene expression in tissue samples at or near single-cell resolution while retaining the spatial location of each sequencing unit within the tissue sample. Through analyzing HST data, we seek to identify sub-populations of cells within a tissue sample that may inform biological phenomena. Existing computational methods either ignore the spatial heterogeneity in gene expression profiles, fail to account for important statistical features such as skewness, or are heuristic-based network clustering methods that lack the inferential benefits of statistical modeling. To address this gap, we develop SPRUCE: a Bayesian spatial multivariate finite mixture model based on multivariate skew-normal distributions, which is capable of identifying distinct cellular sub-populations in HST data. We further implement a novel combination of Pólya-Gamma data augmentation and spatial random effects to infer spatially correlated mixture component membership probabilities without relying on approximate inference techniques. Via a simulation study, we demonstrate the detrimental inferential effects of ignoring skewness or spatial correlation in HST data. Using publicly available human brain HST data, SPRUCE outperforms existing methods in recovering expertly annotated brain layers. Finally, our application of SPRUCE to human breast cancer HST data indicates that SPRUCE can distinguish distinct cell populations within the tumor microenvironment. An R package spruce for fitting the proposed models is available through The Comprehensive R Archive Network.

A Bayesian framework for pathway-guided identification of cancer subgroups by integrating multiple types of genomic data.

In recent years, comprehensive cancer genomics platforms, such as The Cancer Genome Atlas (TCGA), provide access to an enormous amount of high throughput genomic datasets for each patient, including gene expression, DNA copy number alterations, DNA methylation, and somatic mutation. While the integration of these multi-omics datasets has the potential to provide novel insights that can lead to personalized medicine, most existing approaches only focus on gene-level analysis and lack the ability to facilitate biological findings at the pathway-level. In this article, we propose Bayes-InGRiD (Bayesian Integrative Genomics Robust iDentification of cancer subgroups), a novel pathway-guided Bayesian sparse latent factor model for the simultaneous identification of cancer patient subgroups (clustering) and key molecular features (variable selection) within a unified framework, based on the joint analysis of continuous, binary, and count data. By utilizing pathway (gene set) information, Bayes-InGRiD does not only enhance the accuracy and robustness of cancer patient subgroup and key molecular feature identification, but also promotes biological understanding and interpretation. Finally, to facilitate an efficient posterior sampling, an alternative Gibbs sampler for logistic and negative binomial models is proposed using Pólya-Gamma mixtures of normal to represent latent variables for binary and count data, which yields a conditionally Gaussian representation of the posterior. The R package "INGRID" implementing the proposed approach is currently available in our research group GitHub webpage (https://dongjunchung.github.io/INGRID/).

Evaluation of Bayesian spatiotemporal infectious disease models for prospective surveillance analysis.

BACKGROUND: COVID-19 brought enormous challenges to public health surveillance and underscored the importance of developing and maintaining robust systems for accurate surveillance. As public health data collection efforts expand, there is a critical need for infectious disease modeling researchers to continue to develop prospective surveillance metrics and statistical models to accommodate the modeling of large disease counts and variability. This paper evaluated different likelihoods for the disease count model and various spatiotemporal mean models for prospective surveillance. METHODS: We evaluated Bayesian spatiotemporal models, which are the foundation for model-based infectious disease surveillance metrics. Bayesian spatiotemporal mean models based on the Poisson and the negative binomial likelihoods were evaluated with the different lengths of past data usage. We compared their goodness of fit and short-term prediction performance with both simulated epidemic data and real data from the COVID-19 pandemic. RESULTS: The simulation results show that the negative binomial likelihood-based models show better goodness of fit results than Poisson likelihood-based models as deemed by smaller deviance information criteria (DIC) values. However, Poisson models yield smaller mean square error (MSE) and mean absolute one-step prediction error (MAOSPE) results when we use a shorter length of the past data such as 7 and 3 time periods. Real COVID-19 data analysis of New Jersey and South Carolina shows similar results for the goodness of fit and short-term prediction results. Negative binomial-based mean models showed better performance when we used the past data of 52 time periods. Poisson-based mean models showed comparable goodness of fit performance and smaller MSE and MAOSPE results when we used the past data of 7 and 3 time periods. CONCLUSION: We evaluate these models and provide future infectious disease outbreak modeling guidelines for Bayesian spatiotemporal analysis. Our choice of the likelihood and spatiotemporal mean models was influenced by both historical data length and variability. With a longer length of past data usage and more over-dispersed data, the negative binomial likelihood shows a better model fit than the Poisson likelihood. However, as we use a shorter length of the past data for our surveillance analysis, the difference between the Poisson and the negative binomial models becomes smaller. In this case, the Poisson likelihood shows robust posterior mean estimate and short-term prediction results.

Differential prevalence of Adverse Childhood Experiences (ACEs) by gender and substance used in individuals with cannabis, cocaine, opioid, and tobacco use disorders.

Background: Adverse childhood experiences (ACEs) show a graded association with the development of substance use disorders (SUDs) and engagement in risky substance use behaviors. Women are overrepresented among individuals with more severe childhood adversity (≥4 types of ACEs) and may be at particular risk for aberrant substance use.Objectives: To assess the prevalence of ACEs among men and women with cannabis, opioid, cocaine, and tobacco use disorders.Methods: Non-treatment-seeking individuals participating in clinical addiction research at a single site completed the ACE questionnaire and provided a detailed substance use history. Data were analyzed using proportional odds models and logistic regression.Results: Most participants (424/565; 75%) reported at least one ACE, and more than one-quarter (156/565; 27%) reported severe childhood adversity. Relative to men (n = 283), women (n = 282) reported more ACEs (OR = 1.49; p = .01) and more experiences of emotional/physical abuse (OR = 1.52; p = .02), sexual abuse (OR = 4.08; p = .04), and neglect (OR = 2.30; p < .01). Participants in the cocaine (OR = 1.87; n = .01) and opioid (OR = 2.21; p = .01) use disorder, but not cannabis use disorder (OR = 1.46; p = .08), studies reported more severe adversity relative to the tobacco group. Relative to tobacco users, emotional/physical abuse (OR = 1.92; p = .02) and neglect (OR = 2.46; p = .01) scores were higher in cocaine users and household dysfunction scores were higher in opioid users (OR = 2.67; p = .01).Conclusion: The prevalence of ACEs differs with respect to both participant gender and primary substance used. Novel SUD treatment strategies that incorporate ACEs may be uniquely beneficial in specific subpopulations of people with SUDs.

Predictors of Treatment Completion among Women Receiving Integrated Treatment for Comorbid Posttraumatic Stress and Substance Use Disorders.

Background: Retention in treatment for individuals with comorbid posttraumatic stress disorder (PTSD) and substance use disorders (SUD) is an area of concern in treatment outcome studies. The current study explores key variables related to retention in a group of women with comorbid PTSD and SUD enrolled in community SUD treatment randomized to eight weekly sessions of a trauma adapted mindfulness-based relapse prevention (TA-MBRP) or an integrated coping skills (ICS) group intervention. Methods: Two unadjusted and adjusted logistic discrete failure time (DFT) models were fit to examine associations between participants and the time (in weeks) to treatment completion status. Key covariates of interest, including time-varying PTSD Symptom Scale-Self Report (PSS) total score, time-varying Five Factors Mindfulness Questionnaire (FFMQ) total score, group assignment, baseline endorsements of substance use and demographics such as age, race and employment status were fit into the model. Results: In the adjusted PSS model, increased levels of PTSD symptom severity (PSS) scores at week 5 and 7 (PSS OR: 1:06: OR 1.13, respectively) were associated with higher odds of non-completion. In the FFMQ model, increased levels of FFMQ scores at week 6 (OR: 0:92) were associated with lower odds of non-completion. In both models, assignment to the ICS control group and unemployment were associated with lower odds of completion and baseline use of cocaine and sedatives were associated with higher odds of completion. Conclusion: Monitoring PTSD symptom severity and measures of mindfulness can inform providers on strategies to enhance retention early in treatment for individuals with comorbid PTSD/SUD.ClinicalTrials.gov # NCT02755103.

Association between gestational PFAS exposure and Children's adiposity in a diverse population.

Perfluoroalkyl substances (PFAS) are widely distributed suspected obesogens that cross the placenta. However, few data are available to assess potential fetal effects of PFAS exposure on children's adiposity in diverse populations. To address the data gap, we estimated associations between gestational PFAS concentrations and childhood adiposity in a diverse mother-child cohort. We considered 6 PFAS in first trimester blood plasma, measured using ultra-high-performance liquid chromatography with tandem mass spectrometry, collected from non-smoking women with low-risk singleton pregnancies (n = 803). Body mass index (BMI), waist circumference (WC), fat mass, fat-free mass, and % body fat were ascertained in 4-8 year old children as measures of adiposity. We estimated associations of individual gestational PFAS with children's adiposity and overweight/obesity, adjusted for confounders. There were more non-Hispanic Black (31.7 %) and Hispanic (42.6 %) children with overweight/obesity, than non-Hispanic white (18.2 %) and Asian/Pacific Islander (16.4 %) children (p < 0.0001). Perfluorooctane sulfonate (PFOS; 5.3 ng/mL) and perfluorooctanoic acid (2.0 ng/mL) had the highest median concentrations in maternal blood. Among women without obesity (n = 667), greater perfluoroundecanoic acid (PFUnDA) was associated with their children having higher WC z-score (ß = 0.08, 95%CI: 0.01, 0.14; p = 0.02), fat mass (ß = 0.55 kg, 95%CI: 0.21, 0.90; p = 0.002), and % body fat (ß = 0.01 %; 95%CI: 0.003, 0.01; p = 0.004), although the association of PFUnDA with fat mass attenuated at the highest concentrations. Among women without obesity, the associations of PFAS and their children's adiposity varied significantly by self-reported race-ethnicity, although the direction of the associations was inconsistent. In contrast, among the children of women with obesity, greater, PFOS, perfluorononanoic acid, and perfluorodecanoic acid concentrations were associated with less adiposity (n = 136). Our results suggest that specific PFAS may be developmental obesogens, and that maternal race-ethnicity may be an important modifier of the associations among women without obesity.

A Bayesian multivariate mixture model for skewed longitudinal data with intermittent missing observations: An application to infant motor development.

In studies of infant growth, an important research goal is to identify latent clusters of infants with delayed motor development-a risk factor for adverse outcomes later in life. However, there are numerous statistical challenges in modeling motor development: the data are typically skewed, exhibit intermittent missingness, and are correlated across repeated measurements over time. Using data from the Nurture study, a cohort of approximately 600 mother-infant pairs, we develop a flexible Bayesian mixture model for the analysis of infant motor development. First, we model developmental trajectories using matrix skew-normal distributions with cluster-specific parameters to accommodate dependence and skewness in the data. Second, we model the cluster-membership probabilities using a Pólya-Gamma data-augmentation scheme, which improves predictions of the cluster-membership allocations. Lastly, we impute missing responses from conditional multivariate skew-normal distributions. Bayesian inference is achieved through straightforward Gibbs sampling. Through simulation studies, we show that the proposed model yields improved inferences over models that ignore skewness or adopt conventional imputation methods. We applied the model to the Nurture data and identified two distinct developmental clusters, as well as detrimental effects of food insecurity on motor development. These findings can aid investigators in targeting interventions during this critical early-life developmental window.

Exploring associations between prenatal exposure to multiple endocrine disruptors and birth weight with exposure continuum mapping.

BACKGROUND: Improved understanding of how prenatal exposure to environmental mixtures influences birth weight or other adverse outcomes is essential in protecting child health. OBJECTIVE: We illustrate a novel exposure continuum mapping (ECM) framework that combines the self-organizing map (SOM) algorithm with generalized additive modeling (GAM) in order to integrate spatially-correlated learning into the study mixtures of environmental chemicals. We demonstrate our method using biomarker data on chemical mixtures collected from a diverse mother-child cohort. METHODS: We obtained biomarker concentrations for 16 prevalent endocrine disrupting chemicals (EDCs) collected in the first-trimester from a large, ethnically/racially diverse cohort of healthy pregnant women (n = 604) during 2009-2012. This included 4 organochlorine pesticides (OCPs), 4 polybrominated diphenyl ethers (PBDEs), 4 polychlorinated biphenyls (PCBs), and 4 perfluoroalkyl substances (PFAS). We applied a two-stage exposure continuum mapping (ECM) approach to investigate the combined impact of the EDCs on birth weight. First, we analyzed our EDC data with SOM in order to reduce the dimensionality of our exposure matrix into a two-dimensional grid (i.e., map) where nodes depict the types of EDC mixture profiles observed within our data. We define this map as the 'exposure continuum map', as the gridded surface reflects a continuous sequence of exposure profiles where adjacent nodes are composed of similar mixtures and profiles at more distal nodes are more distinct. Lastly, we used GAM to estimate a joint-dose response based on the coordinates of our ECM in order to capture the relationship between participant location on the ECM and infant birth weight after adjusting for maternal age, race/ethnicity, pre-pregnancy body mass index (BMI), education, serum cotinine, total plasma lipids, and infant sex. Single chemical regression models were applied to facilitate comparison. RESULTS: We found that an ECM with 36 mixture profiles retained 70% of the total variation in the exposure data. Frequency analysis showed that the most common profiles included relatively low concentrations for most EDCs (~10%) and that profiles with relatively higher concentrations (for single or multiple EDCs) tended to be rarer (~1%) but more distinct. Estimation of a joint-dose response function revealed that lower birth weights mapped to locations where profile compositions were dominated by relatively high PBDEs and select OCPs. Higher birth weights mapped to locations where profiles consisted of higher PCBs. These findings agreed well with results from single chemical models. CONCLUSIONS: Findings from our study revealed a wide range of prenatal exposure scenarios and found that combinations exhibiting higher levels of PBDEs were associated with lower birth weight and combinations with higher levels of PCBs and PFAS were associated with increased birth weight. Our ECM approach provides a promising framework for supporting studies of other exposure mixtures.

Propensity score matching for multilevel spatial data: accounting for geographic confounding in health disparity studies.

BACKGROUND: Diabetes is a public health burden that disproportionately affects military veterans and racial minorities. Studies of racial disparities are inherently observational, and thus may require the use of methods such as Propensity Score Analysis (PSA). While traditional PSA accounts for patient-level factors, this may not be sufficient when patients are clustered at the geographic level and thus important confounders, whether observed or unobserved, vary by geographic location. METHODS: We employ a spatial propensity score matching method to account for "geographic confounding", which occurs when the confounding factors, whether observed or unobserved, vary by geographic region. We augment the propensity score and outcome models with spatial random effects, which are assigned scaled Besag-York-Mollié priors to address spatial clustering and improve inferences by borrowing information across neighboring geographic regions. We apply this approach to a study exploring racial disparities in diabetes specialty care between non-Hispanic black and non-Hispanic white veterans. We construct multiple global estimates of the risk difference in diabetes care: a crude unadjusted estimate, an estimate based solely on patient-level matching, and an estimate that incorporates both patient and spatial information. RESULTS: In simulation we show that in the presence of an unmeasured geographic confounder, ignoring spatial heterogeneity results in increased relative bias and mean squared error, whereas incorporating spatial random effects improves inferences. In our study of racial disparities in diabetes specialty care, the crude unadjusted estimate suggests that specialty care is more prevalent among non-Hispanic blacks, while patient-level matching indicates that it is less prevalent. Hierarchical spatial matching supports the latter conclusion, with a further increase in the magnitude of the disparity. CONCLUSIONS: These results highlight the importance of accounting for spatial heterogeneity in propensity score analysis, and suggest the need for clinical care and management strategies that are culturally sensitive and racially inclusive.

Participation in the Child and Adult Care Food Program is associated with fewer barriers to serving healthier foods in early care and education.

BACKGROUND: Early care and education (ECE) is an important setting for influencing young children's dietary intake. There are several factors associated with barriers to healthy eating in ECE, and recent evidence suggests that participation in the Child and Adult Care Food Program (CACFP), the primary national food assistance program in ECE, may be associated with fewer barriers to serving healthier foods. However, no prior studies have examined differences between CACFP participants and non-participants across a large, multi-state sample. This is the first study to examine the association between CACFP participation and barriers to serving healthier foods in ECE using a random sample of directors from two regions across the country. METHODS: We conducted a cross-sectional survey among a random sample of child care center directors from four states (Massachusetts, North Carolina, Rhode Island, and South Carolina). We conducted logistic and Poisson regression to calculate the odds and incidence rate ratios of reporting 1) no barriers, 2) specific barriers (e.g., cost), and 3) the total number of barriers, by CACFP status, adjusting for covariates of interest. RESULTS: We received 713 surveys (36% response rate). About half (55%) of centers participated in CACFP. The most prevalent reported barriers to serving healthier foods were cost (42%) and children's food preferences (19%). Directors from CACFP centers were twice as likely to report no barriers, compared to directors from non-CACFP centers (OR 2.03; 95% CI [1.36, 3.04]; p < 0.01). Directors from CACFP centers were less likely to report cost as a barrier (OR = 0.46; 95% [CI 0.31, 0.67]; p < 0.001), and reported fewer barriers overall (IRR = 0.77; 95% CI [0.64, 0.92]; p < 0.01), compared to directors from non-CACFP centers. CONCLUSIONS: CACFP directors reported fewer barriers to serving healthier foods in child care centers. Still, cost and children's food preferences are persistent barriers to serving healthier foods in ECE. Future research should evaluate characteristics of CACFP participation that may alleviate these barriers, and whether barriers emerge or persist following 2017 rule changes to CACFP nutrition standards.

A multivariate discrete failure time model for the analysis of infant motor development.

We develop a multivariate discrete failure time model for the analysis of infant motor development. We use the model to jointly evaluate the time (in months) to achievement of three well-established motor milestones: sitting up, crawling, and walking. The model includes a subject-specific latent factor that reflects underlying heterogeneity in the population and accounts for within-subject dependence across the milestones. The factor loadings and covariate effects are allowed to vary flexibly across milestones, and the milestones are permitted to have unique at-risk intervals corresponding to different developmental windows. We adopt a Bayesian inferential approach and develop a convenient data-augmented Gibbs sampler for posterior computation. We conduct simulation studies to illustrate key features of the model and use the model to analyze data from the Nurture study, a birth cohort examining infant health and development during the first year of life.

Effects of a New State Policy on Physical Activity Practices in Child Care Centers in South Carolina.

OBJECTIVES: We sought to determine the extent to which child care centers in South Carolina improved physical activity practices after a new policy took effect. METHODS: In 2012, South Carolina adopted new mandatory physical activity standards within its child care quality enhancement program. This quasi-experimental study used North Carolina, a state not making policy changes, as the comparison. Participants were 34 child care centers in South Carolina and 30 centers in North Carolina. Researchers used the Environment and Policy Assessment and Observation (EPAO) tool to conduct center observations before and after policy implementation and then conducted repeated-measures linear regression with interaction between state and time for the Physical Activity Environment Total Score and the 8 subscale scores. RESULTS: Compared with centers in North Carolina, EPAO subscale scores in South Carolina centers increased significantly for the Fixed Play Environment (P < .001) and Physical Activity Training and Education (P = .015). The state-by-time interaction of Physical Activity Environment Total Score approached statistical significance (P = .06). CONCLUSIONS: Adoption of new physical activity standards in South Carolina child care centers was associated with improvements in practices aimed at increasing children's physical activity.

The LZIP: A Bayesian latent factor model for correlated zero-inflated counts.

Motivated by a study of molecular differences among breast cancer patients, we develop a Bayesian latent factor zero-inflated Poisson (LZIP) model for the analysis of correlated zero-inflated counts. The responses are modeled as independent zero-inflated Poisson distributions conditional on a set of subject-specific latent factors. For each outcome, we express the LZIP model as a function of two discrete random variables: the first captures the propensity to be in an underlying "at-risk" state, while the second represents the count response conditional on being at risk. The latent factors and loadings are assigned conditionally conjugate gamma priors that accommodate overdispersion and dependence among the outcomes. For posterior computation, we propose an efficient data-augmentation algorithm that relies primarily on easily sampled Gibbs steps. We conduct simulation studies to investigate both the inferential properties of the model and the computational capabilities of the proposed sampling algorithm. We apply the method to an analysis of breast cancer genomics data from The Cancer Genome Atlas.

Control beliefs and risk for 4-year mortality in older adults: a prospective cohort study.

BACKGROUND: Control beliefs are important psychological factors that likely contribute to heterogeneity in health outcomes for older adults. We evaluated whether control beliefs are associated with risk for 4-year mortality, after accounting for established "classic" biomedical risk factors. We also determined if an enhanced risk model with control beliefs improved identification of individuals with low vs. high mortality risk. METHODS: We used nationally representative data from the Health and Retirement Study (2006-2012) for adults 50 years or older in 2006 (n = 7313) or 2008 (n = 6301). We assessed baseline perceived global control (measured as 2 dimensions-"constraints" and "mastery"), and health-specific control. We also obtained baseline data for 12 established biomedical risk factors of 4-year mortality: age, sex, 4 medical conditions (diabetes mellitus, cancer, lung disease and heart failure), body mass index less than 25 kg/m2, smoking, and 4 functional difficulties (with bathing, managing finances, walking several blocks and pushing or pulling heavy objects). Deaths within 4 years of follow-up were determined through interviews with respondents' family and the National Death Index. RESULTS: After accounting for classic biomedical risk factors, perceived constraints were significantly associated with higher mortality risk (third quartile scores odds ratio [OR] 1.37, 95% CI 1.03-1.81; fourth quartile scores OR 1.45, 95% CI, 1.09-1.92), while health-specific control was significantly associated with lower risk (OR 0.69-0.78 for scores above first quartile). Higher perceived mastery scores were not consistently associated with decreased risk. The enhanced model with control beliefs found an additional 3.5% of participants (n = 222) with low predicted risk of 4-year mortality (i.e., 4% or less); observed mortality for these individuals was 1.8% during follow-up. Compared with participants predicted to have low mortality risk only by the classic biomedical model, individuals identified by only the enhanced model were older, had higher educational status, higher income, and higher prevalence of diabetes mellitus and cancer. CONCLUSION: Control beliefs were significantly associated with risk for 4-year mortality; accounting for these factors improved identification of low-risk individuals. More work is needed to determine how assessment of control beliefs could enable targeting of clinical interventions to support at-risk older adults.

Bayesian two-part spatial models for semicontinuous data with application to emergency department expenditures.

In health services research, it is common to encounter semicontinuous data characterized by a point mass at zero and a continuous distribution of positive values. Examples include medical expenditures, in which the zeros represent patients who do not use health services, while the continuous distribution describes the level of expenditures among users. Semicontinuous data are customarily analyzed using two-part mixture models. In the spatial analysis of semicontinuous data, two-part models are especially appealing because they provide a joint picture of how health services utilization and associated expenditures vary across geographic regions. However, when applying these models, careful attention must be paid to distributional choices, as model misspecification can lead to biased and imprecise inferences. This paper introduces a broad class of Bayesian two-part models for the spatial analysis of semicontinuous data. Specific models considered include two-part lognormal, log skew-elliptical, and Bayesian non-parametric models. Multivariate conditionally autoregressive priors are used to link model components and provide spatial smoothing across neighboring regions, resulting in a joint spatial modeling framework for health utilization and expenditures. We develop a fully conjugate Gibbs sampling scheme, leading to efficient posterior computation. We illustrate the approach using data from a recent study of emergency department expenditures.