Bayesian modeling of spatially differentiated multivariate enamel defects of the children's primary maxillary central incisor teeth

BACKGROUND

The analysis of dental caries has been a major focus of recent work on modeling dental defect data. While a dental caries focus is of major importance in dental research, the examination of developmental defects which could also contribute at an early stage of dental caries formation, is also of potential interest. This paper proposes a set of methods which address the appearance of different combinations of defects across different tooth regions. In our modeling we assess the linkages between tooth region development and both the type of defect and associations with etiological predictors of the defects which could be influential at different times during the tooth crown development.

METHODS

We develop different hierarchical model formulations under the Bayesian paradigm to assess exposures during primary central incisor (PMCI) tooth development and PMCI defects. We evaluate the Bayesian hierarchical models under various simulation scenarios to compare their performance with both simulated dental defect data and real data from a motivating application.

RESULTS

The proposed model provides inference on identifying a subset of etiological predictors of an individual defect accounting for the correlation between tooth regions and on identifying a subset of etiological predictors for the joint effect of defects. Furthermore, the model provides inference on the correlation between the regions of the teeth as well as between the joint effect of the developmental enamel defects and dental caries. Simulation results show that the proposed model consistently yields steady inferences in identifying etiological biomarkers associated with the outcome of localized developmental enamel defects and dental caries under varying simulation scenarios as deemed by small mean square error (MSE) when comparing the simulation results to real application results.

CONCLUSION

We evaluate the proposed model under varying simulation scenarios to develop a model for multivariate dental defects and dental caries assuming a flexible covariance structure that can handle regional and joint effects. The proposed model shed new light on methods for capturing inclusive predictors in different multivariate joint models under the same covariance structure and provides a natural extension to a nested hierarchical model.

Evaluation and comparison of spatial cluster detection methods for improved decision making of disease surveillance: a case study of national dengue surveillance in Thailand

BACKGROUND

Dengue is a mosquito-borne disease that causes over 300 million infections worldwide each year with no specific treatment available. Effective surveillance systems are needed for outbreak detection and resource allocation. Spatial cluster detection methods are commonly used, but no general guidance exists on the most appropriate method for dengue surveillance. Therefore, a comprehensive study is needed to assess different methods and provide guidance for dengue surveillance programs.

METHODS

To evaluate the effectiveness of different cluster detection methods for dengue surveillance, we selected and assessed commonly used methods: Getis Ord [Formula: see text], Local Moran, SaTScan, and Bayesian modeling. We conducted a simulation study to compare their performance in detecting clusters, and applied all methods to a case study of dengue surveillance in Thailand in 2019 to further evaluate their practical utility.

RESULTS

In the simulation study, Getis Ord [Formula: see text] and Local Moran had similar performance, with most misdetections occurring at cluster boundaries and isolated hotspots. SaTScan showed better precision but was less effective at detecting inner outliers, although it performed well on large outbreaks. Bayesian convolution modeling had the highest overall precision in the simulation study. In the dengue case study in Thailand, Getis Ord [Formula: see text] and Local Moran missed most disease clusters, while SaTScan was mostly able to detect a large cluster. Bayesian disease mapping seemed to be the most effective, with adaptive detection of irregularly shaped disease anomalies.

CONCLUSIONS

Bayesian modeling showed to be the most effective method, demonstrating the best accuracy in adaptively identifying irregularly shaped disease anomalies. In contrast, SaTScan excelled in detecting large outbreaks and regular forms. This study provides empirical evidence for the selection of appropriate tools for dengue surveillance in Thailand, with potential applicability to other disease control programs in similar settings.

Predictors of Developmental Defects of Enamel in Primary Maxillary Central Incisors Using Bayesian Model Selection

INTRODUCTION

Localized non-inheritable developmental defects of tooth enamel (DDE) are classified as enamel hypoplasia (EH), opacity (OP), and post-eruptive breakdown (PEB) using the enamel defects index. To better understand the etiology of DDE, we assessed the linkages amongst exposome variables for these defects during the specific time duration for enamel mineralization of the human primary maxillary central incisor enamel crowns. In general, these two teeth develop between 13 and 14 weeks in utero and 3-4 weeks' postpartum of a full-term delivery, followed by tooth eruption at about 1 year of age.

METHODS

We utilized existing datasets for mother-child dyads that encompassed 12 weeks' gestation through birth and early infancy, and child DDE outcomes from digital images of the erupted primary maxillary central incisor teeth. We applied a Bayesian modeling paradigm to assess the important predictors of EH, OP, and PEB.

RESULTS

The results of Gibbs variable selection showed a key set of predictors: mother's prepregnancy body mass index (BMI); maternal serum concentrations of calcium and phosphorus at gestational week 28; child's gestational age; and both mother's and child's functional vitamin D deficiency (FVDD). In this sample of healthy mothers and children, significant predictors for OP included the child having a gestational period >36 weeks and FVDD at birth, and for PEB included a mother's prepregnancy BMI <21.5 and higher serum phosphorus concentration at week 28.

CONCLUSION

In conclusion, our methodology and results provide a roadmap for assessing timely biomarker measures of exposures during specific tooth development to better understand the etiology of DDE for future prevention.

The Association between Mediated Deprivation and Ovarian Cancer Survival among African American Women

BACKGROUND

Deprivation indices are often used to adjust for socio-economic disparities in health studies. Their role has been partially evaluated for certain population-level cancer outcomes, but examination of their role in ovarian cancer is limited. In this study, we evaluated a range of well-recognized deprivation indices in relation to cancer survival in a cohort of self-identified Black women diagnosed with ovarian cancer. This study aimed to determine if clinical or diagnostic characteristics lie on a mediating pathway between socioeconomic status (SES) and deprivation and ovarian cancer survival in a minority population that experiences worse survival from ovarian cancer.

METHODS

We used mediation analysis to look at the direct and indirect causal effects of deprivation indices with main mediators of the SEER stage at diagnosis and residual disease. The analysis employed Bayesian structural equation models with variable selection. We applied a joint Bayesian structural model for the mediator, including a Weibull mixed model for the vital outcome with deprivation as exposure. We selected modifiers via a Monte Carlo model selection procedure.

RESULTS

The results suggest that high SES-related indices, such as Yost, Kolak urbanicity (URB), mobility (MOB) and SES dimensions, and concentrated disadvantage index (CDI), all have a significant impact on improved survival. In contrast, area deprivation index (ADI)/Singh, and area level poverty (POV) did not have a major impact. In some cases, the indirect effects have very wide credible intervals, so the total effect is not well estimated despite the estimation of the direct effect.

CONCLUSIONS

First, it is clear that commonly used indices such as Yost, or CDI both significantly impact the survival experience of Black women diagnosed with epithelial ovarian cancer. In addition, the Kolak dimension indices (URB, MOB, mixed immigrant: MICA and SES) also demonstrate a significant association, depending on the mediator. Mediation effects differ according to the mediator chosen.

Deprivation and segregation in ovarian cancer survival among African American women: a mediation analysis

PURPOSE

Deprivation and segregation indices are often examined as possible explanations for observed health disparities in population-based studies. In this study, we assessed the role of recognized deprivation and segregation indices specifically as they affect survival in a cohort of self-identified Black women diagnosed with ovarian cancer who enrolled in the African American Cancer Epidemiology Study.

METHODS

Mediation analysis was used to examine the direct and indirect effects between deprivation or segregation and overall survival via a Bayesian structural equation model with Gibbs variable selection.

RESULTS

The results suggest that high socioeconomic status-related indices have an association with increased survival, ranging from 25% to 56%. In contrast, index of concentration at the extremes-race does not have a significant impact on overall survival. In many cases, the indirect effects have very wide credible intervals; consequently, the total effect is not well estimated despite the estimation of the direct effect.

CONCLUSIONS

Our results show that Black women living in higher socioeconomic status neighborhoods are associated with increased survival with ovarian cancer using area-level economic indices such as Yost or index of concentration at the extremes-income. In addition, the Kolak urbanization index has a similar impact and highlights the importance of area-level deprivation and segregation as potentially modifiable social factors in ovarian cancer survival.

Association of inflammation-related exposures and ovarian cancer survival in a multi-site cohort study of Black women

BACKGROUND

An association was observed between an inflammation-related risk score (IRRS) and worse overall survival (OS) among a cohort of mostly White women with invasive epithelial ovarian cancer (EOC). Herein, we evaluated the association between the IRRS and OS among Black women with EOC, a population with higher frequencies of pro-inflammatory exposures and worse survival.

METHODS

The analysis included 592 Black women diagnosed with EOC from the African American Cancer Epidemiology Study (AACES). Cox proportional hazards models were used to compute hazard ratios (HRs) and 95% confidence intervals (CIs) for the association of the IRRS and OS, adjusting for relevant covariates. Additional inflammation-related exposures, including the energy-adjusted Dietary Inflammatory Index (E-DIITM), were evaluated.

RESULTS

A dose-response trend was observed showing higher IRRS was associated with worse OS (per quartile HR: 1.11, 95% CI: 1.01-1.22). Adding the E-DII to the model attenuated the association of IRRS with OS, and increasing E-DII, indicating a more pro-inflammatory diet, was associated with shorter OS (per quartile HR: 1.12, 95% CI: 1.02-1.24). Scoring high on both indices was associated with shorter OS (HR: 1.54, 95% CI: 1.16-2.06).

CONCLUSION

Higher levels of inflammation-related exposures were associated with decreased EOC OS among Black women.

Evaluation of predictive capability of Bayesian spatio-temporal models for Covid-19 spread

BACKGROUND

Bayesian models have been applied throughout the Covid-19 pandemic especially to model time series of case counts or deaths. Fewer examples exist of spatio-temporal modeling, even though the spatial spread of disease is a crucial factor in public health monitoring. The predictive capabilities of infectious disease models is also important.

METHODS

In this study, the ability of Bayesian hierarchical models to recover different parts of the variation in disease counts is the focus. It is clear that different measures provide different views of behavior when models are fitted prospectively. Over a series of time horizons one step predictions have been generated and compared for different models (for case counts and death counts). These Bayesian SIR models were fitted using MCMC at 28 time horizons to mimic prospective prediction. A range of goodness of prediction measures were analyzed across the different time horizons.

RESULTS

A particularly important result is that the peak intensity of case load is often under-estimated, while random spikes in case load can be mimicked using time dependent random effects. It is also clear that during the early wave of the pandemic simpler model forms are favored, but subsequently lagged spatial dependence models for cases are favored, even if the sophisticated models perform better overall.

DISCUSSION

The models fitted mimic the situation where at a given time the history of the process is known but the future must be predicted based on the current evolution which has been observed. Using an overall 'best' model for prediction based on retrospective fitting of the complete pandemic waves is an assumption. However it is also clear that this case count model is well favored over other forms. During the first wave a simpler time series model predicts case counts better for counties than a spatially dependent one. The picture is more varied for morality.

CONCLUSIONS

From a predictive point of view it is clear that spatio-temporal models applied to county level Covid-19 data within the US vary in how well they fit over time and also how well they predict future events. At different times, SIR case count models and also mortality models with cumulative counts perform better in terms of prediction. A fundamental result is that predictive capability of models varies over time and using the same model could lead to poor predictive performance. In addition it is clear that models addressing the spatial context for case counts (i.e. with lagged neighborhood terms) and cumulative case counts for mortality data are clearly better at modeling spatio-temporal data which is commonly available for the Covid-19 pandemic in different areas of the globe.

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.

Spatiotemporal reproduction number with Bayesian model selection for evaluation of emerging infectious disease transmissibility: an application to COVID-19 national surveillance data

BACKGROUND

To control emerging diseases, governments often have to make decisions based on limited evidence. The effective or temporal reproductive number is used to estimate the expected number of new cases caused by an infectious person in a partially susceptible population. While the temporal dynamic is captured in the temporal reproduction number, the dominant approach is currently based on modeling that implicitly treats people within a population as geographically well mixed.

METHODS

In this study we aimed to develop a generic and robust methodology for estimating spatiotemporal dynamic measures that can be instantaneously computed for each location and time within a Bayesian model selection and averaging framework. A simulation study was conducted to demonstrate robustness of the method. A case study was provided of a real-world application to COVID-19 national surveillance data in Thailand.

RESULTS

Overall, the proposed method allowed for estimation of different scenarios of reproduction numbers in the simulation study. The model selection chose the true serial interval when included in our study whereas model averaging yielded the weighted outcome which could be less accurate than model selection. In the case study of COVID-19 in Thailand, the best model based on model selection and averaging criteria had a similar trend to real data and was consistent with previously published findings in the country.

CONCLUSIONS

The method yielded robust estimation in several simulated scenarios of force of transmission with computing flexibility and practical benefits. Thus, this development can be suitable and practically useful for surveillance applications especially for newly emerging diseases. As new outbreak waves continue to develop and the risk changes on both local and global scales, our work can facilitate policymaking for timely disease control.

Survival of epithelial ovarian cancer in Black women: a society to cell approach in the African American cancer epidemiology study (AACES)

PURPOSE

The causes for the survival disparity among Black women with epithelial ovarian cancer (EOC) are likely multi-factorial. Here we describe the African American Cancer Epidemiology Study (AACES), the largest cohort of Black women with EOC.

METHODS

AACES phase 2 (enrolled 2020 onward) is a multi-site, population-based study focused on overall survival (OS) of EOC. Rapid case ascertainment is used in ongoing patient recruitment in eight U.S. states, both northern and southern. Data collection is composed of a survey, biospecimens, and medical record abstraction. Results characterizing the survival experience of the phase 1 study population (enrolled 2010-2015) are presented.

RESULTS

Thus far, ~ 650 patients with EOC have been enrolled in the AACES. The five-year OS of AACES participants approximates those of Black women in the Surveillance Epidemiology and End Results (SEER) registry who survive at least 10-month past diagnosis and is worse compared to white women in SEER, 49 vs. 60%, respectively. A high proportion of women in AACES have low levels of household income (45% < $25,000 annually), education (51% ≤ high school education), and insurance coverage (32% uninsured or Medicaid). Those followed annually differ from those without follow-up with higher levels of localized disease (28 vs 24%) and higher levels of optimal debulking status (73 vs 67%).

CONCLUSION

AACES is well positioned to evaluate the contribution of social determinants of health to the poor survival of Black women with EOC and advance understanding of the multi-factorial causes of the ovarian cancer survival disparity in Black women.

Bayesian space-time SIR modeling of Covid-19 in two US states during the 2020-2021 pandemic

This paper describes the Bayesian SIR modeling of the 3 waves of Covid-19 in two contrasting US states during 2020-2021. A variety of models are evaluated at the county level for goodness-of-fit and an assessment of confounding predictors is also made. It is found that models with three deprivation predictors and neighborhood effects are important. In addition, the work index from Google mobility was also found to provide an increased explanation of the transmission dynamics.

Issues in Bayesian prospective surveillance of spatial health data

In this paper I review some of the major issues that arise when geo-referenced health data are to be the subject of prospective surveillance. The review focusses on modelbased approaches to this activity, and proposes the Bayesian paradigm as a convenient vehicle for modeling. Various posterior functional measures are discussed including the SCPO and SKL and a number of extensions to these are considered. Overall the value of Bayesian Hierarchical Modeling (BHM) with surveillance functionals is stressed in its relevance to early warning of adverse risk scenarios.

Using Bayesian time-stratified case-crossover models to examine associations between air pollution and "asthma seasons" in a low air pollution environment

Many areas of the United States have air pollution levels typically below Environmental Protection Agency (EPA) regulatory limits. Most health effects studies of air pollution use meteorological (e.g., warm/cool) or astronomical (e.g., solstice/equinox) definitions of seasons despite evidence suggesting temporally-misaligned intra-annual periods of relative asthma burden (i.e., “asthma seasons”). We introduce asthma seasons to elucidate whether air pollutants are associated with seasonal differences in asthma emergency department (ED) visits in a low air pollution environment. Within a Bayesian time-stratified case-crossover framework, we quantify seasonal associations between highly resolved estimates of six criteria air pollutants, two weather variables, and asthma ED visits among 66,092 children ages 5–19 living in South Carolina (SC) census tracts from 2005 to 2014. Results show that coarse particulates (particulate matter <10 μm and >2.5 μm: PM_10-2.5) and nitrogen oxides (NO_x) may contribute to asthma ED visits across years, but are particularly implicated in the highest-burden fall asthma season. Fine particulate matter (<2.5 μm: PM_2.5) is only associated in the lowest-burden summer asthma season. Relatively cool and dry conditions in the summer asthma season and increased temperatures in the spring and fall asthma seasons are associated with increased ED visit odds. Few significant associations in the medium-burden winter and medium-high-burden spring asthma seasons suggest other ED visit drivers (e.g., viral infections) for each, respectively. Across rural and urban areas characterized by generally low air pollution levels, there are acute health effects associated with particulate matter, but only in the summer and fall asthma seasons and differing by PM size.

Association between Neighborhood Social Deprivation and Stage at Diagnosis among Breast Cancer Patients in South Carolina

The purpose of this study was to examine the association between neighborhood social deprivation and individual-level characteristics on breast cancer staging in African American and white breast cancer patients. We established a retrospective cohort of patients with breast cancer diagnosed from 1996 to 2015 using the South Carolina Central Cancer Registry. We abstracted sociodemographic and clinical variables from the registry and linked these data to a county-level composite that captured neighborhood social conditions-the social deprivation index (SDI). Data were analyzed using chi-square tests, Student's t-test, and multivariable ordinal regression analysis to evaluate associations. The study sample included 52,803 female patients with breast cancer. Results from the multivariable ordinal regression model demonstrate that higher SDI (OR = 1.06, 95% CI: 1.02-1.10), African American race (OR = 1.35, 95% CI: 1.29-1.41), and being unmarried (OR = 1.17, 95% CI: 1.13-1.22) were associated with a distant stage at diagnosis. Higher tumor grade, younger age, and more recent year of diagnosis were also associated with distant-stage diagnosis. As a proxy for neighborhood context, the SDI can be used by cancer registries and related population-based studies to identify geographic areas that could be prioritized for cancer prevention and control efforts.

Joint space-time Bayesian disease mapping via quantification of disease risk association

Alzheimer's disease is an increasingly prevalent neurological disorder with no effective therapies. Thus, there is a need to characterize the progression of Alzheimer's disease risk in order to preclude its inception in patients. Characterizing Alzheimer's disease risk can be accomplished at the population-level by the space-time modeling of Alzheimer's disease incidence data. In this paper, we develop flexible Bayesian hierarchical models which can borrow risk information from conditions antecedent to Alzheimer's disease, such as mild cognitive impairment, in an effort to better characterize Alzheimer's disease risk over space and time. From an application of these models to real-world Alzheimer's disease and mild cognitive impairment spatiotemporal incidence data, we found that our novel models provided improved model goodness of fit, and via a simulation study, we demonstrated the importance of diagnosing the label-switching problem for our models as well as the importance of model specification in order to best capture the contribution of time in modeling Alzheimer's disease risk.

Modelling and predicting the spatio-temporal spread of COVID-19, associated deaths and impact of key risk factors in England

COVID-19 caseloads in England have passed through a first peak, and at the time of this analysis appeared to be gradually increasing, potentially signalling the emergence of a second wave. To ensure continued response to the epidemic is most effective, it is imperative to better understand both retrospectively and prospectively the geographical evolution of COVID-19 caseloads and deaths at small-area resolution, identify localised areas in space-time at significantly higher risk, quantify the impact of changes in localised population mobility (or movement) on caseloads, identify localised risk factors for increased mortality and project the likely course of the epidemic at high spatial resolution in coming weeks. We applied a Bayesian hierarchical space-time SEIR model to assess the spatiotemporal variability of COVID-19 caseloads (transmission) and deaths at small-area scale in England [Middle Layer Super Output Area (MSOA), 6791 units] and by week (using observed data from week 5 to 34 of 2020), including key determinants, the modelled transmission dynamics and spatial-temporal random effects. We also estimate the number of cases and deaths at small-area resolution with uncertainty projected forward in time by MSOA (up to week 51 of 2020), the impact mobility reductions (and subsequent easing) have had on COVID-19 caseloads and quantify the impact of key socio-demographic risk factors on COVID-19 related mortality risk by MSOA. Reductions in population mobility during the course of the first lockdown had a significant impact on the reduction of COVID-19 caseloads across England, however local authorities have had a varied rate of reduction in population movement which our model suggest has substantially impacted the geographic heterogeneity in caseloads at small-area scale. The steady gain in population mobility, observed from late April, appears to have contributed to a slowdown in caseload reductions towards late June and subsequent start of the second wave. MSOA with higher proportions of elderly (70+ years of age) and elderly living in deprivation, both with very distinct geographic distributions, have a significantly elevated COVID-19 mortality rates. While non-pharmaceutical interventions (that is, reductions in population mobility and social distancing) had a profound impact on the trajectory of the first wave of the COVID-19 outbreak in England, increased population mobility appears to have significantly contributed to the second wave. A number of contiguous small-areas appear to be at a significant elevated risk of high COVID-19 transmission, many of which are also at increased risk for higher mortality rates. A geographically staggered re-introduction of intensified social distancing measures is advised and limited cross MSOA movement if the magnitude and geographic extent of the second wave is to be reduced.

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.

Abstract PO-167: Area-level social deprivation and stage at diagnosis among breast cancer patients in South Carolina

Background: Cancer incidence, staging and mortality rates differ across geographic areas; however, there is a need for a better understanding of how neighborhood level socioeconomic and access to care factors impact cancer burden to tailor cancer control interventions in ways that appropriately target geographic determinants of cancer health disparities. The purpose of this study was to characterize the distribution of neighborhood deprivation in a cohort of breast cancer patients and examine the effect of social deprivation, healthcare professional shortage Area (HPSA) designation, and individual-level characteristics on breast cancer staging.

Methods: We established a retrospective cohort of breast cancer patients diagnosed from 1996 to 2015 using the South Carolina Central Cancer Registry.

Sociodemographic (e.g., race, age) and clinical variables were abstracted from the registry. We linked registry data to county-level variables to determine levels of social deprivation and residence in a health care professional shortage area using the Robert Graham Center’s Social Deprivation Index (SDI) and the Health Resources and Services Administration primary care HPSA designation. Bivariate analyses and multivariate regression analyses were used to examine associations. Results: The sample included 54,501 female breast cancer patients. Overall, the mean for SDI was 54.2 (+18.1) and the range was 76 (19-95). Approximately 44.4% of women lived in areas with high levels of social deprivation (e.g., SDI score of 52 to 95). In the logistic regression model, living in a geographic area with high social deprivation was significantly associated with African American race (OR=2.3, 95% C.I. 2.2-2.4), being unmarried (OR=1.2, 95% C.I. 1.1-1.3), and HPSA designation (OR=14.0, 95% C.I. 13.5- 14.6). Higher tumor grade (OR=1.2; 95% CI.1.2-1.3) and later stage (OR=1.1, 95% C.I. 1.1-1.2) were also significantly associated with neighborhood deprivation. Conclusion: This study shows that SDI differs by race and clinical characteristics among breast cancer patients. The SDI could be integrated into tumor registries and cancer research to understand the effects of neighborhood level variables on cancer health disparities to improve the precision of cancer control interventions that are developed to address geographic determinants.

Citation Format: Andrew B. Lawson, Chanita Hughes-Halbert, Oluwole A. Babatunde, Whitney E. Zahnd, Jan M. Eberth. Area-level social deprivation and stage at diagnosis among breast cancer patients in South Carolina [abstract]. In: Proceedings of the AACR Virtual Conference: Thirteenth AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2020 Oct 2-4. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2020;29(12 Suppl):Abstract nr PO-167.

A Bayesian spatio-temporal analysis of neighborhood pediatric asthma emergency department visit disparities

Asthma disparities have complex, neighborhood-level drivers that are not well understood. Consequently, identifying particular contextual factors that contribute to disparities is a public health goal. We study pediatric asthma emergency department (ED) visit disparities and neighborhood factors associated with them in South Carolina (SC) census tracts from 1999 to 2015. Leveraging a Bayesian framework, we identify risk clusters, spatially-varying relationships, and risk percentile-specific associations. Clusters of high risk occur in both rural and urban census tracts with high probability, with neighborhood-specific associations suggesting unique risk factors for each locale. Bayesian methods can help clarify the neighborhood drivers of health disparities.

Multivariate Air Pollution Prediction Modeling with partial Missingness

Missing observations from air pollution monitoring networks have posed a longstanding problem for health investigators of air pollution. Growing interest in mixtures of air pollutants has further complicated this problem, as many new challenges have arisen that require development of novel methods. The objective of this study is to develop a methodology for multivariate prediction of air pollution. We focus specifically on tackling different forms of missing data, such as: spatial (sparse sites), outcome (pollutants not measured at some sites), and temporal (varieties of interrupted time series). To address these challenges, we develop a novel multivariate fusion framework, which leverages the observed inter-pollutant correlation structure to reduce error in the simultaneous prediction of multiple air pollutants. Our joint fusion model employs predictions from the Environmental Protection Agency's Community Multiscale Air Quality (CMAQ) model along with spatio-temporal error terms. We have implemented our models on both simulated data and a case study in South Carolina for 8 pollutants over a 28-day period in June 2006. We found that our model, which uses a multivariate correlated error in a Bayesian framework, showed promising predictive accuracy particularly for gaseous pollutants.

Toward Preventing Enamel Hypoplasia: Modeling Maternal and Neonatal Biomarkers of Human Calcium Homeostasis

AIM

The aim of this study was to assess biomarkers of calcium homeostasis and tooth development, in mothers during pregnancy and their children at birth, for enamel hypoplasia (EH) in the primary maxillary central incisor teeth.

METHODS

Bayesian methodology was used for secondary data analyses from a randomized, controlled trial of prenatal vitamin D3 supplementation in healthy mothers (N = 350) and a follow-up study of a subset of the children. The biomarkers were serum calcium (Ca), phosphorus (P), intact parathyroid hormone (iPTH), total circulating 25-dihydroxyvitamin D (25(OH)D), and 1,25-dihydroxyvitamin D (1,25(OH)2D). The maternal biomarkers were assayed monthly during pregnancy, and the child's biomarkers were derived from cord blood. Digital images of the child's 2 teeth were scored for EH using Enamel Defects Index criteria for each of the incisal, middle, and cervical regions for an EH extent score.

RESULTS

The child EH prevalence was 41% (60/145), with most defects present in the incisal and middle tooth regions. Cord blood iPTH and 1,25(OH)2D levels were significantly associated with EH extent after controlling for maternal factors. For every 1 pg/mL increase in cord blood iPTH, the EH extent decreased by approximately 6%. For every 10 pg/mL increase in cord blood 1,25(OH)2D, the EH extent increased by almost 30% (holding all other terms constant and adjusting for subject-level heterogeneity). The relationship between maternal 25(OH)D and maternal mean iPTH varied significantly by EH extent.

CONCLUSION

The results suggest possible modifiable relationships of maternal and neonatal factors of calcium homeostasis during pregnancy and at birth for EH, contributing to the frontier of knowledge regarding sound tooth development for dental caries prevention.

Spatiotemporal multi-disease transmission dynamic measure for emerging diseases: an application to dengue and zika integrated surveillance in Thailand

Background

New emerging diseases are public health concerns in which policy makers have to make decisions in the presence of enormous uncertainty. This is an important challenge in terms of emergency preparation requiring the operation of effective surveillance systems. A key concept to investigate the dynamic of infectious diseases is the basic reproduction number. However it is difficult to be applicable in real situations due to the underlying theoretical assumptions.

Methods

In this paper we propose a robust and flexible methodology for estimating disease strength varying in space and time using an alternative measure of disease transmission within the hierarchical modeling framework. The proposed measure is also extended to allow for incorporating knowledge from related diseases to enhance performance of surveillance system.

Results

A simulation was conducted to examine robustness of the proposed methodology and the simulation results demonstrate that the proposed method allows robust estimation of the disease strength across simulation scenarios. A real data example is provided of an integrative application of Dengue and Zika surveillance in Thailand. The real data example also shows that combining both diseases in an integrated analysis essentially decreases variability of model fitting.

Conclusions

The proposed methodology is robust in several simulated scenarios of spatiotemporal transmission force with computing flexibility and practical benefits. This development has potential for broad applicability as an alternative tool for integrated surveillance of emerging diseases such as Zika.

Temporally dependent accelerated failure time model for capturing the impact of events that alter survival in disease mapping

The introduction of spatial and temporal frailty parameters in survival models furnishes a way to represent unmeasured confounding in the outcome of interest. Using a Bayesian accelerated failure time model, we are able to flexibly explore a wide range of spatial and temporal options for structuring frailties as well as examine the benefits of using these different structures in certain settings. A setting of particular interest for this work involved using temporal frailties to capture the impact of events of interest on breast cancer survival. Our results suggest that it is important to include these temporal frailties when there is a true temporal structure to the outcome and including them when a true temporal structure is absent does not sacrifice model fit. Additionally, the frailties are able to correctly recover the truth imposed on simulated data without affecting the fixed effect estimates. In the case study involving Louisiana breast cancer-specific mortality, the temporal frailty played an important role in representing the unmeasured confounding related to improvements in knowledge, education, and disease screenings as well as the impacts of Hurricane Katrina and the passing of the Affordable Care Act. In conclusion, the incorporation of temporal, in addition to spatial, frailties in survival analysis can lead to better fitting models and improved inference by representing both spatially and temporally varying unmeasured risk factors and confounding that could impact survival. Specifically, we successfully estimated changes in survival around the time of events of interest.

Perceived discrimination, trust in physicians, and prolonged symptom duration before ovarian cancer diagnosis in the African American Cancer Epidemiology Study

BACKGROUND

Discrimination and trust are known barriers to accessing health care. Despite well-documented racial disparities in the ovarian cancer care continuum, the role of these barriers has not been examined. This study evaluated the association of everyday discrimination and trust in physicians with a prolonged interval between symptom onset and ovarian cancer diagnosis (hereafter referred to as prolonged symptom duration).

METHODS

Subjects included cases enrolled in the African American Cancer Epidemiology Study, a multisite case-control study of epithelial ovarian cancer among black women. Logistic regression was used to calculate odds ratios (ORs) and 95% confidence intervals (CIs) for associations of everyday discrimination and trust in physicians with a prolonged symptom duration (1 or more symptoms lasting longer than the median symptom-specific duration), and it controlled for access-to-care covariates and potential confounders.

RESULTS

Among the 486 cases in this analysis, 302 women had prolonged symptom duration. In the fully adjusted model, a 1-unit increase in the frequency of everyday discrimination increased the odds of prolonged symptom duration 74% (OR, 1.74; 95% CI, 1.22-2.49), but trust in physicians was not associated with prolonged symptom duration (OR, 0.86; 95% CI, 0.66-1.11).

CONCLUSIONS

Perceived everyday discrimination was associated with prolonged symptom duration, whereas more commonly evaluated determinants of access to care and trust in physicians were not. These results suggest that more research on the effects of interpersonal barriers affecting ovarian cancer care is warranted.

Analysis of racial differences in hospital stays in the presence of geographic confounding

Using recent methods for spatial propensity score modeling, we examine differences in hospital stays between non-Hispanic black and non-Hispanic white veterans with type 2 diabetes. We augment a traditional patient-level propensity score model with a spatial random effect to create a matched sample based on the estimated propensity score. We then use a spatial negative binomial hurdle model to estimate differences in both hospital admissions and inpatient days. We demonstrate that in the presence of unmeasured geographic confounding, spatial propensity score matching in addition to the spatial negative binomial hurdle outcome model yields improved performance compared to the outcome model alone. In the motivating application, we construct three estimates of racial differences in hospitalizations: the risk difference in admission, the mean difference in number of inpatient days among those hospitalized, and the mean difference in number of inpatient days across all patients (hospitalized and non-hospitalized). Results indicate that non-Hispanic black veterans with type 2 diabetes have a lower risk of hospital admission and a greater number of inpatient days on average. The latter result is especially important considering that we observed much smaller effect sizes in analyses that did not incorporate spatial matching. These results emphasize the need to address geographic confounding in health disparity studies.

Interval Estimation for Age-Adjusted Rate Ratios Using Bayesian Convolution Model

Spatial correlation raises challenges in estimating confidence intervals for region specific event rates and rate ratios between geographic units that are nested. Methods have been proposed to incorporate spatial correlation by assuming various distributions for the structure of autocorrelation patterns. However, the derivation of these statistics based on approximation may have to condition on the distributional assumption underlying the data generating process, which may not hold for certain situations. This paper explores the feasibility of utilizing a Bayesian convolution model (BCM), which includes an uncorrelated heterogeneity (UH) and a conditional autoregression (CAR) component to accommodate both uncorrelated and correlated spatial heterogeneity, to estimate the 95% confidence intervals for age-adjusted rate ratios among geographic regions with existing spatial correlations. A simulation study is conducted and a BCM method is applied to two cancer incidence datasets to calculate age-adjusted rate/ratio for the counties in the State of Kentucky relative to the entire state. In comparison to three existing methods, without and with spatial correlation, the Bayesian convolution model-based estimation provides moderate shrinkage effect for the point estimates based on the neighbor structure across regions and produces a wider interval due to the inclusion of uncertainty in the spatial autocorrelation parameters. The overall spatial pattern of region incidence rate from BCM approach appears to be like the direct estimates and other methods for both datasets, even though "smoothing" occurs in some local regions. The Bayesian Convolution Model allows flexibility in the specification of risk components and can improve the accuracy of interval estimates of age-adjusted rate ratios among geographical regions as it considers spatial correlation.

A data-driven approach for estimating the change-points and impact of major events on disease risk

Considering the impact of events on disease risk is important. Here, a Bayesian spatio-temporal accelerated failure time model furnished an ideal situation for modeling events that could impact survival experience via spatial and temporal frailty estimates. Through a hierarchical structure, this model allowed the data to detect the change-point(s) in addition to generating the event-related estimates. Both a real data case study and a simulation study were employed for testing these methods. The results suggested that meaningful and accurate change-points could be detected. Further, accurate event-related estimates for individuals in relation to those change-points could be obtained. By allowing the data to drive the change-point choices, the models were better fitting and the inference was more accurate.

Temporal Trends in Cardiovascular Hospital Discharges Following a Mass Chlorine Exposure Event in Graniteville, South Carolina

Background: On January 6, 2005, a train derailed in Graniteville, South Carolina, releasing nearly 60,000 kg of toxic chlorine gas. The disaster left nine people dead and was responsible for hundreds of hospitalizations and outpatient visits in the subsequent weeks. While chlorine gas primarily affects the respiratory tract, a growing body of evidence suggests that acute exposure may also cause vascular injury and cardiac toxicity. Here, we describe the incidence of cardiovascular hospitalizations among residents of the zip codes most affected by the chlorine gas plume, and compare the incidence of cardiovascular discharges in the years leading up to the event (2000–2004) to the incidence in the years following the event (2005–2014).

Methods: De-identified hospital discharge information was collected from the South Carolina Revenue and Fiscal Affairs Office for individuals residing in the selected zip codes for the years 2000 to 2014. A quasi-experimental study design was utilized with a population-level interrupted time series model to examine hospital discharge rates for Graniteville-area residents for three cardiovascular diagnoses: hypertension (HTN), acute myocardial infarction (AMI), and coronary heart disease (CHD). We used linear regression with autoregressive error correction to compare slopes for pre- and post-spill time periods. Data from the 2000 and 2010 censuses were used to calculate rates and to provide information on potential demographic shifts over the course of the study.

Results: A significant increase in hypertension-related hospital discharge rates was observed for the years following the Graniteville chlorine spill (slope 8.2, p < 0.001). Concurrent changes to CHD and AMI hospital discharge rates were in the opposite direction (slopes −3.2 and −0.3, p < 0.01 and 0.14, respectively). Importantly, the observed trend cannot be attributed to an aging population.

Conclusions: An unusual increase in hypertension-related hospital discharge rates in the area affected by the Graniteville chlorine spill contrasts with national and state-level trends. A number of factors related to the spill may be contributing the observation: disaster-induced hypertension, healthcare services access issues, and, possibly, chlorine-induced susceptibility to vascular pathologies. Due to the limitations of our data, we cannot determine whether the individuals who visited the hospital were the ones exposed to chlorine gas, however, the finding warrants additional research. Future studies are needed to determine the etiology of the increase and whether individuals exposed to chlorine are at a heightened risk for hypertensive heart disease.

Bayesian cure-rate survival model with spatially structured censoring

We propose a Bayesian spatial model for time-to-event data in which we allow the censoring mechanism to depend on covariates and have a spatial structure. The survival model incorporates a cure rate fraction and assumes that the time-to-event follows a Weibull distribution, with covariates such as race, stage, grade, marital status and age at diagnosis being linked to its scale parameter. With right censoring being a primary concern, we consider a joint logistic regression model for the death versus censoring indicator, allowing dependence on covariates and including a spatial structure via the use of random effects. We apply the models to examine prostate cancer data from the Surveillance, Epidemiology, and End Results (SEER) registry, which displays marked spatial variation.

A Bayesian Quantile Modeling for Spatiotemporal Relative Risk: An Application to Adverse Risk Detection of Respiratory Diseases in South Carolina, USA

Quantile modeling has been seen as an alternative and useful complement to ordinary regression mainly focusing on the mean. To directly apply quantile modeling to areal data the discrete conditional quantile function of the data can be an issue. Although jittering by adding a small number from a uniform distribution to impose pseudo-continuity has been proposed, the approach can have a great influence on responses with small values. Thus we proposed an alternative to model the quantiles of relative risk for spatiotemporal areal health data within a Bayesian framework using the log-Laplace distribution. A simulation study was conducted to assess the performance of the proposed method and examine whether the model could robustly estimate quantiles of spatiotemporal count data. To perform a test with a real data example, we evaluated the potential application of clustering under the proposed log-Laplace and mean regression. The data were obtained from the total number of emergency room discharges for respiratory conditions, both infectious and non-infectious diseases, in the U.S. state of South Carolina in 2009. From both simulation and case studies, the proposed quantile modeling demonstrated potential for broad applicability in various areas of spatial health studies including anomaly detection.

Multiscale measurement error models for aggregated small area health data

Spatial data are often aggregated from a finer (smaller) to a coarser (larger) geographical level. The process of data aggregation induces a scaling effect which smoothes the variation in the data. To address the scaling problem, multiscale models that link the convolution models at different scale levels via the shared random effect have been proposed. One of the main goals in aggregated health data is to investigate the relationship between predictors and an outcome at different geographical levels. In this paper, we extend multiscale models to examine whether a predictor effect at a finer level hold true at a coarser level. To adjust for predictor uncertainty due to aggregation, we applied measurement error models in the framework of multiscale approach. To assess the benefit of using multiscale measurement error models, we compare the performance of multiscale models with and without measurement error in both real and simulated data. We found that ignoring the measurement error in multiscale models underestimates the regression coefficient, while it overestimates the variance of the spatially structured random effect. On the other hand, accounting for the measurement error in multiscale models provides a better model fit and unbiased parameter estimates.

A Bayesian two-stage spatially dependent variable selection model for space-time health data

In space-time epidemiological modeling, most studies have considered the overall variations in relative risk to better estimate the effects of risk factors on health outcomes. However, the associations between risk factors and health outcomes may vary across space and time. Especially, the temporal patterns of the covariate effects may depend on space. Thus, we propose a Bayesian two-stage spatially dependent variable selection approach for space-time health data to determine the spatially varying subsets of regression coefficients with common temporal dependence. The two-stage structure allows reduction of the spatial confounding bias in the estimates of the regression coefficients. A simulation study is conducted to examine the performance of the proposed two-stage model. We apply the proposed model to the number of inpatients with lung cancer in 159 counties of Georgia, USA.

Zero-inflated multiscale models for aggregated small area health data

It is our primary focus to study the spatial distribution of disease incidence at different geographical levels. Often, spatial data are available in the form of aggregation at multiple scale levels such as census tract, county, state, and so on. When data are aggregated from a fine (e.g. county) to a coarse (e.g. state) geographical level, there will be loss of information. The problem is more challenging when excessive zeros are available at the fine level. After data aggregation, the excessive zeros at the fine level will be reduced at the coarse level. If we ignore the zero inflation and the aggregation effect, we could get inconsistent risk estimates at the fine and coarse levels. Hence, in this paper, we address those problems using zero inflated multiscale models that jointly describe the risk variations at different geographical levels. For the excessive zeros at the fine level, we use a zero inflated convolution model, whereas we consider a regular convolution model for the smoothed data at the coarse level. These methods provide a consistent risk estimate at the fine and coarse levels when high percentages of structural zeros are present in the data.

Spatiotemporal multivariate mixture models for Bayesian model selection in disease mapping

It is often the case that researchers wish to simultaneously explore the behavior of and estimate overall risk for multiple, related diseases with varying rarity while accounting for potential spatial and/or temporal correlation. In this paper, we propose a flexible class of multivariate spatio-temporal mixture models to fill this role. Further, these models offer flexibility with the potential for model selection as well as the ability to accommodate lifestyle, socio-economic, and physical environmental variables with spatial, temporal, or both structures. Here, we explore the capability of this approach via a large scale simulation study and examine a motivating data example involving three cancers in South Carolina. The results which are focused on four model variants suggest that all models possess the ability to recover simulation ground truth and display improved model fit over two baseline Knorr-Held spatio-temporal interaction model variants in a real data application.

Addressing geographic confounding through spatial propensity scores: a study of racial disparities in diabetes

Motivated by a study exploring differences in glycemic control between non-Hispanic black and non-Hispanic white veterans with type 2 diabetes, we aim to address a type of confounding that arises in spatially referenced observational studies. Specifically, we develop a spatial doubly robust propensity score estimator to reduce bias associated with geographic confounding, which occurs when measured or unmeasured confounding factors vary by geographic location, leading to imbalanced group comparisons. We augment the doubly robust estimator with spatial random effects, which are assigned conditionally autoregressive priors to improve inferences by borrowing information across neighboring geographic regions. Through a series of simulations, we show that ignoring spatial variation results in increased absolute bias and mean squared error, while the spatial doubly robust estimator performs well under various levels of spatial heterogeneity and moderate sample sizes. In the motivating application, we construct three global estimates of the risk difference between race groups: an unadjusted estimate, a doubly robust estimate that adjusts only for patient-level information, and a hierarchical spatial doubly robust estimate. Results indicate a gradual reduction in the risk difference at each stage, with the inclusion of spatial random effects providing a 20% reduction compared to an estimate that ignores spatial heterogeneity. Smoothed maps indicate poor glycemic control across Alabama and southern Georgia, areas comprising the so-called "stroke belt." These results suggest the need for community-specific interventions to target diabetes in geographic areas of greatest need.

Assessment of spatial variation in breast cancer-specific mortality using Louisiana SEER data

BACKGROUND

Previous studies suggest spatial differences in mortality for many types of cancer, including breast cancer. Identifying explanations for these spatial differences results in a better understanding of what leads to longer survival time.

METHODS

We used a Bayesian accelerated failure time model with spatial frailty terms to investigate potential spatial differences in breast cancer mortality following breast cancer diagnosis using 2000-2013 Louisiana SEER data.

RESULTS

There are meaningful spatial differences in breast cancer mortality across the parishes of Louisiana, even after adjusting for known demographic and clinical risk factors. For example, the average survival time of a woman diagnosed in Orleans parish was 1.51 times longer than that of a woman diagnosed in Terrebonne parish. Additionally, there is evidence to suggest shorter survival times in lower income parishes along the Red and Mississippi Rivers, as well as parishes with lower socioeconomic status, less access to care and fresh food, worse quality of care, and more workers in certain industries.

CONCLUSION

The addition of spatial frailties to account for an individual's geographic location is useful when analyzing breast cancer mortality data. Our findings suggest that survival following breast cancer diagnosis could potentially be improved if socioeconomic status differences were addressed, healthcare improved in quality and became more accessible, and certain industrial situations were improved for individuals diagnosed in parishes identified as having shorter average survival times.

Comparing multilevel and multiscale convolution models for small area aggregated health data

In spatial epidemiology, data are often arrayed hierarchically. The classification of individuals into smaller units, which in turn are grouped into larger units, can induce contextual effects. On the other hand, a scaling effect can occur due to the aggregation of data from smaller units into larger units. In this paper, we propose a shared multilevel model to address the contextual effects. In addition, we consider a shared multiscale model to adjust for both scale and contextual effects simultaneously. We also study convolution and independent multiscale models, which are special cases of shared multilevel and shared multiscale models, respectively. We compare the performance of the models by applying them to real and simulated data sets. We found that the shared multiscale model was the best model across a range of simulated and real scenarios as measured by the deviance information criterion (DIC) and the Watanabe Akaike information criterion (WAIC).

Space-time variation of respiratory cancers in South Carolina: a flexible multivariate mixture modeling approach to risk estimation

PURPOSE

Many types of cancer have an underlying spatiotemporal distribution. Spatiotemporal mixture modeling can offer a flexible approach to risk estimation via the inclusion of latent variables.

METHODS

In this article, we examine the application and benefits of using four different spatiotemporal mixture modeling methods in the modeling of cancer of the lung and bronchus as well as "other" respiratory cancer incidences in the state of South Carolina.

RESULTS

Of the methods tested, no single method outperforms the other methods; which method is best depends on the cancer under consideration. The lung and bronchus cancer incidence outcome is best described by the univariate modeling formulation, whereas the "other" respiratory cancer incidence outcome is best described by the multivariate modeling formulation.

CONCLUSIONS

Spatiotemporal multivariate mixture methods can aid in the modeling of cancers with small and sparse incidences when including information from a related, more common type of cancer.

Spatially-dependent Bayesian model selection for disease mapping

In disease mapping where predictor effects are to be modeled, it is often the case that sets of predictors are fixed, and the aim is to choose between fixed model sets. Model selection methods, both Bayesian model selection and Bayesian model averaging, are approaches within the Bayesian paradigm for achieving this aim. In the spatial context, model selection could have a spatial component in the sense that some models may be more appropriate for certain areas of a study region than others. In this work, we examine the use of spatially referenced Bayesian model averaging and Bayesian model selection via a large-scale simulation study accompanied by a small-scale case study. Our results suggest that BMS performs well when a strong regression signature is found.

Bayesian latent structure modeling of walking behavior in a physical activity intervention

The analysis of walking behavior in a physical activity intervention is considered. A Bayesian latent structure modeling approach is proposed whereby the ability and willingness of participants is modeled via latent effects. The dropout process is jointly modeled via a linked survival model. Computational issues are addressed via posterior sampling and a simulated evaluation of the longitudinal model's ability to recover latent structure and predictor effects is considered. We evaluate the effect of a variety of socio-psychological and spatial neighborhood predictors on the propensity to walk and the estimation of latent ability and willingness in the full study.

Mind the Scales: Harnessing Spatial Big Data for Infectious Disease Surveillance and Inference

Spatial big data have the velocity, volume, and variety of big data sources and contain additional geographic information. Digital data sources, such as medical claims, mobile phone call data records, and geographically tagged tweets, have entered infectious diseases epidemiology as novel sources of data to complement traditional infectious disease surveillance. In this work, we provide examples of how spatial big data have been used thus far in epidemiological analyses and describe opportunities for these sources to improve disease-mitigation strategies and public health coordination. In addition, we consider the technical, practical, and ethical challenges with the use of spatial big data in infectious disease surveillance and inference. Finally, we discuss the implications of the rising use of spatial big data in epidemiology to health risk communication, and public health policy recommendations and coordination across scales.

Spatio-temporal Bayesian model selection for disease mapping

Spatio-temporal analysis of small area health data often involves choosing a fixed set of predictors prior to the final model fit. In this paper, we propose a spatio-temporal approach of Bayesian model selection to implement model selection for certain areas of the study region as well as certain years in the study time line. Here, we examine the usefulness of this approach by way of a large-scale simulation study accompanied by a case study. Our results suggest that a special case of the model selection methods, a mixture model allowing a weight parameter to indicate if the appropriate linear predictor is spatial, spatio-temporal, or a mixture of the two, offers the best option to fitting these spatio-temporal models. In addition, the case study illustrates the effectiveness of this mixture model within the model selection setting by easily accommodating lifestyle, socio-economic, and physical environmental variables to select a predominantly spatio-temporal linear predictor.