Predicting differential improvements in annual pollutant concentrations and exposures for regulatory policy assessment.

Over the past decade, researchers and policy-makers have become increasingly interested in regulatory and policy interventions to reduce air pollution concentrations and improve human health. Studies have typically relied on relatively sparse environmental monitoring data that lack the spatial resolution to assess small-area improvements in air quality and health. Few studies have integrated multiple types of measures of an air pollutant into one single modeling framework that combines spatially- and temporally-rich monitoring data. In this paper, we investigated the differential effects of California emissions reduction plan on reducing air pollution between those living in the goods movement corridors (GMC) that are within 500 m of major highways that serve as truck routes to those farther away or adjacent to routes that prohibit trucks. A mixed effects Deletion/Substitution/Addition (D/S/A) machine learning algorithm was developed to model annual pollutant concentrations of nitrogen dioxide (NO2) by taking repeated measures into consideration and by integrating multiple types of NO2 measurements, including those through government regulatory and research-oriented saturation monitoring into a single modeling framework. Difference-in-difference analysis was conducted to identify whether those living in GMC demonstrated statistically larger reductions in air pollution exposure. The mixed effects D/S/A machine learning modeling result indicated that GMC had 2 ppb greater reductions in NO2 concentrations from pre- to post-policy period than far away areas. The difference-in-difference analysis demonstrated that the subjects living in GMC experienced statistically significant greater reductions in NO2 exposure than those living in the far away areas. This study contributes to scientific knowledge by providing empirical evidence that improvements in air quality via the emissions reductions plan policies impacted traffic-related air pollutant concentrations and associated exposures most among low-income Californians with chronic conditions living in GMC. The identified differences in pollutant reductions across different location domains may be applicable to other states or other countries if similar policies are enacted.

Part 3. Modeling of Multipollutant Profiles and Spatially Varying Health Effects with Applications to Indicators of Adverse Birth Outcomes.

The highly intercorrelated nature of air pollutants makes it difficult to examine their combined effects on health. As such, epidemiological studies have traditionally focused on single-pollutant models that use regression-based techniques to examine the marginal association between a pollutant and a health outcome. These relatively simple, additive models are useful for discerning the effect of a single pollutant on a health outcome with all other pollutants held to fixed values. However, pollutants occur in complex mixtures consisting of highly correlated combinations of individual exposures. For example, evidence for synergy among pollutants in causing health effects has been recently reviewed by Mauderly and Samet (2009). Also, studies cited in the Ozone Criteria Document (U.S. Environmental Protection Agency [U.S. EPA*] 2006) confirmed that synergisms between ozone and other pollutants have been demonstrated in laboratory studies involving humans and animals. Thus, the highly correlated nature of air pollution exposures makes marginal, single-pollutant models inadequate. This issue was raised in a report by the National Research Council (NRC 2004), which called for a multipollutant approach to air quality management. Here we present and apply a series of statistical approaches that treat patterns of covariates as a whole unit, stochastically grouping pollutant patterns into clusters and then using these cluster assignments as random effects in a regression model. Using this approach, the effect of a multipollutant pattern, or profile, is determined in a manner that takes into account the uncertainty in the clustering process. The models are set in a Bayesian framework, and in general, Markov chain Monte Carlo (MCMC) techniques (Gilks et al. 1998). For interpretation purposes, a best clustering is derived, and the uncertainty related to this best clustering is determined by utilizing model averaging techniques, in a manner such that consistent clustering obtained by the estimation process generally yields smaller standard errors while inconsistent clustering is generally associated with larger errors. These multivariate methods are applied to a range of different problems related to air pollution exposures, namely an association of multipollutant profiles with indicators of poverty and to an assessment of the association between measures of various air pollutants, patterns of socioeconomic status (SES), and birth outcomes. All of these studies involve an examination of regional-level exposures, at the census tract (CT) and census block group (CBG) levels, and individual-level outcomes throughout Los Angeles (LA) County. Results indicate that effects of pollutants vary spatially and vary in a complex interconnected manner that cannot be discerned using standard additive line ar models. Results obtaine d from these studies can be used to efficiently use limited resources to inform policies in targeting are as where air pollution reductions result in maximum health benefits.

Informal confidential voting interviewing in a sexual risk assessment of men who have sex with men (MSM) and transgenders (hijra) in Bangalore, India.

OBJECTIVE: The accuracy of self-reporting sensitive sexual risk behaviours is highly susceptible to misreporting. Informal confidential voting interviews (ICVIs) may minimise social desirability bias by increasing the privacy of the interview setting. The objective was to investigate determinants of risky behaviour among men who have sex with men (MSM) and 'hijra' (transgenders) reported through two interviewing tools: ICVIs and face-to-face interviews (FTFIs). METHODS: Cluster random sampling was used to recruit MSM in 85 cruising sites in Bangalore, including eight hammams (bath houses) and 77 public locations where MSM and hijra cruise for sex. Individuals were randomly allocated to one of the data collection methods(5:2 FTFI : ICVI). Data were analysed using standard regression and a profile regression approach that associates clusters of behaviours with our outcome (FTFI vs ICVI). RESULTS: A total of 372 MSM and hijra were interviewed for the FTFIs and 153 respondents completed ICVIs. Participants were more likely to report injecting drug use (4% vs 1%; p=0.008) and paying to have sex with a female sex worker (FSW) in the last year (28% vs 8%; p=0.001) in the ICVIs. There were no differences to questions on sociodemographics, sexual debut with another male, non-condom use (12% vs 14%), ever selling sex to men (58% vs 56%), current female partner (26% vs 20%) and non-condom use with a main female partner (17% vs 19%). CONCLUSIONS: The significant differences between interview modes for certain outcomes, such as intravenous drug use and sex with a FSW, demonstrate how certain behaviour is stigmatised among the MSM community. Nevertheless, the lack of effect of the interviewing tool in other outcomes may indicate either less reporting bias in reporting this behaviour or environmental factors such as the interviewers not adequately screening themselves from the respondent or a potential disadvantage of using other MSM as interviewers.

Invited commentary: Lost in estimation--searching for alternatives to markov chains to fit complex Bayesian models.

Bayesian methods have seen an increase in popularity in a wide variety of scientific fields, including epidemiology. One of the main reasons for their widespread application is the power of the Markov chain Monte Carlo (MCMC) techniques generally used to fit these models. As a result, researchers often implicitly associate Bayesian models with MCMC estimation procedures. However, Bayesian models do not always require Markov-chain-based methods for parameter estimation. This is important, as MCMC estimation methods, while generally quite powerful, are complex and computationally expensive and suffer from convergence problems related to the manner in which they generate correlated samples used to estimate probability distributions for parameters of interest. In this issue of the Journal, Cole et al. (Am J Epidemiol. 2012;175(5):368-375) present an interesting paper that discusses non-Markov-chain-based approaches to fitting Bayesian models. These methods, though limited, can overcome some of the problems associated with MCMC techniques and promise to provide simpler approaches to fitting Bayesian models. Applied researchers will find these estimation approaches intuitively appealing and will gain a deeper understanding of Bayesian models through their use. However, readers should be aware that other non-Markov-chain-based methods are currently in active development and have been widely published in other fields.

Bayesian profile regression with an application to the National Survey of Children's Health.

Standard regression analyses are often plagued with problems encountered when one tries to make inference going beyond main effects using data sets that contain dozens of variables that are potentially correlated. This situation arises, for example, in epidemiology where surveys or study questionnaires consisting of a large number of questions yield a potentially unwieldy set of interrelated data from which teasing out the effect of multiple covariates is difficult. We propose a method that addresses these problems for categorical covariates by using, as its basic unit of inference, a profile formed from a sequence of covariate values. These covariate profiles are clustered into groups and associated via a regression model to a relevant outcome. The Bayesian clustering aspect of the proposed modeling framework has a number of advantages over traditional clustering approaches in that it allows the number of groups to vary, uncovers subgroups and examines their association with an outcome of interest, and fits the model as a unit, allowing an individual's outcome potentially to influence cluster membership. The method is demonstrated with an analysis of survey data obtained from the National Survey of Children's Health. The approach has been implemented using the standard Bayesian modeling software, WinBUGS, with code provided in the supplementary material available at Biostatistics online. Further, interpretation of partitions of the data is helped by a number of postprocessing tools that we have developed.

Assessing uncertainty in spatial exposure models for air pollution health effects assessment.

BACKGROUND: Although numerous epidemiologic studies now use models of intraurban exposure, there has been little systematic evaluation of the performance of different models. OBJECTIVES: In this present article we proposed a modeling framework for assessing exposure model performance and the role of spatial autocorrelation in the estimation of health effects. METHODS: We obtained data from an exposure measurement substudy of subjects from the Southern California Children's Health Study. We examined how the addition of spatial correlations to a previously described unified exposure and health outcome modeling framework affects estimates of exposure-response relationships using the substudy data. The methods proposed build upon the previous work, which developed measurement-error techniques to estimate long-term nitrogen dioxide exposure and its effect on lung function in children. In this present article, we further develop these methods by introducing between- and within-community spatial autocorrelation error terms to evaluate effects of air pollution on forced vital capacity. The analytical methods developed are set in a Bayesian framework where multistage models are fitted jointly, properly incorporating parameter estimation uncertainty at all levels of the modeling process. RESULTS: Results suggest that the inclusion of residual spatial error terms improves the prediction of adverse health effects. These findings also demonstrate how residual spatial error may be used as a diagnostic for comparing exposure model performance.

A survey of current Bayesian gene mapping methods.

Recently, there has been much interest in the use of Bayesian statistical methods for performing genetic analyses. Many of the computational difficulties previously associated with Bayesian analysis, such as multidimensional integration, can now be easily overcome using modern high-speed computers and Markov chain Monte Carlo (MCMC) methods. Much of this new technology has been used to perform gene mapping, especially through the use of multi-locus linkage disequilibrium techniques. This review attempts to summarise some of the currently available methods and the software available to implement these methods.

Markov chain Monte Carlo without likelihoods.

Many stochastic simulation approaches for generating observations from a posterior distribution depend on knowing a likelihood function. However, for many complex probability models, such likelihoods are either impossible or computationally prohibitive to obtain. Here we present a Markov chain Monte Carlo method for generating observations from a posterior distribution without the use of likelihoods. It can also be used in frequentist applications, in particular for maximum-likelihood estimation. The approach is illustrated by an example of ancestral inference in population genetics. A number of open problems are highlighted in the discussion.

Fine-scale mapping of disease genes with multiple mutations via spatial clustering techniques.

We present a method to perform fine mapping by placing haplotypes into clusters on the basis of risk. Each cluster has a haplotype "center." Cluster allocation is defined according to haplotype centers, with each haplotype assigned to the cluster with the "closest" center. The closeness of two haplotypes is determined by a similarity metric that measures the length of the shared segment around the location of a putative functional mutation for the particular cluster. Our method allows for missing marker information but still estimates the risks of complete haplotypes without resorting to a one-marker-at-a-time analysis. The dimensionality issues that can occur in haplotype analyses are removed by sampling over the haplotype space, allowing for estimation of haplotype risks without explicitly assigning a parameter to each haplotype to be estimated. In this way, we are able to handle haplotypes of arbitrary size. Furthermore, our clustering approach has the potential to allow us to detect the presence of multiple functional mutations.

Application of Bayesian spatial statistical methods to analysis of haplotypes effects and gene mapping.

We propose a method to analyze haplotype effects using ideas derived from Bayesian spatial statistics. We assume that two haplotypes that are similar to one another in structure are likely to have similar risks, and define a distance metric to specify the appropriate level of closeness between the two haplotypes. Through the choice of distance metric, varying levels of population genetics theory can be incorporated into the modeling process, including some that allow estimation of the location of the disease causing mutation(s). This location can be estimated, along with the other parameters of the model, using Markov chain Monte Carlo (MCMC) estimation methods. We demonstrate the effectiveness of the model on two real datasets, a well-known dataset used to fine-map the gene for cystic fibrosis, and one used to localize the gene for Friedreich's ataxia.