A Bayesian model for sparse functional data

Summary.   We propose a method for analyzing data which consist of curves on multiple individuals, i.e., longitudinal or functional data. We use a Bayesian model where curves are expressed as linear combinations of B-splines with random coefficients. The curves are estimated as posterior means obtained via Markov chain Monte Carlo (MCMC) methods, which automatically select the local level of smoothing. The method is applicable to situations where curves are sampled sparsely and/or at irregular time points. We construct posterior credible intervals for the mean curve and for the individual curves. This methodology provides unified, efficient, and flexible means for smoothing functional data.     

Bayesian latent multi‐state modeling for nonequidistant longitudinal electronic health records

Large amounts of longitudinal health records are now available for dynamic monitoring of the underlying processes governing the observations. However, the health status progression across time is not typically observed directly: records are observed only when a subject interacts with the system, yielding irregular and often sparse observations. This suggests that the observed trajectories should be modeled via a latent continuous‐time process potentially as a function of time‐varying covariates. We develop a continuous‐time hidden Markov model to analyze longitudinal data accounting for irregular visits and different types of observations. By employing a specific missing data likelihood formulation, we can construct an efficient computational algorithm. We focus on Bayesian inference for the model: this is facilitated by an expectation‐maximization algorithm and Markov chain Monte Carlo methods. Simulation studies demonstrate that these approaches can be implemented efficiently for large data sets in a fully Bayesian setting. We apply this model to a real cohort where patients suffer from chronic obstructive pulmonary disease with the outcome being the number of drugs taken, using health care utilization indicators and patient characteristics as covariates.     

A high‐resolution approach for the spatiotemporal analysis of forest canopy space using terrestrial laser scanning data

Forest canopies and tree crown structures are of high ecological importance. Measuring canopies and crowns by direct inventory methods is time‐consuming and of limited accuracy. High‐resolution inventory tools, in particular terrestrial laser scanning (TLS), is able to overcome these limitations and obtain three‐dimensional (3D) structural information about the canopy with a very high level of detail. The main objective of this study was to introduce a novel method to analyze spatiotemporal dynamics in canopy occupancy at the individual tree and local neighborhood level using high‐resolution 3D TLS data. For the analyses, a voxel grid approach was applied. The tree crowns were modeled through the combination of two approaches: the encasement of all crown points with a 3D α‐shape, which was then converted into a voxel grid, and the direct voxelization of the crown points. We show that canopy occupancy at individual tree level can be quantified as the crown volume occupied only by the respective tree or shared with neighboring trees. At the local neighborhood level, our method enables the precise determination of the extent of canopy space filling, the identification of tree–tree interactions, and the analysis of complementary space use. Using multitemporal TLS data recordings, this method allows the precise detection and quantification of changes in canopy occupancy through time. The method is applicable to a wide range of investigations in forest ecology research, including the study of tree diversity effects on forest productivity or growing space analyses for optimal tree growth. Due to the high accuracy of this novel method, it facilitates the precise analyses even of highly plastic individual tree crowns and, thus, the realistic representation of forest canopies. Moreover, our voxel grid framework is flexible enough to allow for the inclusion of further biotic and abiotic variables relevant to complex analyses of forest canopy dynamics.     

Spatial Modeling of Air Pollution in Studies of Its Short‐Term Health Effects

Summary In studies that estimate the short‐term effects of air pollution on health, daily measurements of pollution concentrations are often available from a number of monitoring locations within the study area. However, the health data are typically only available in the form of daily counts for the entire area, meaning that a corresponding single daily measure of pollution is required. The standard approach is to average the observed measurements at the monitoring locations, and use this in a log‐linear health model. However, as the pollution surface is spatially variable this simple summary is unlikely to be an accurate estimate of the average pollution concentration across the region, which may lead to bias in the resulting health effects. In this article, we propose an alternative approach that jointly models the pollution concentrations and their relationship with the health data using a Bayesian spatio‐temporal model. We compare this approach with the simple spatial average using a simulation study, by investigating the impact of spatial variation, monitor placement, and measurement error in the pollution data. An epidemiological study from Greater London is then presented, which estimates the relationship between respiratory mortality and four different pollutants.     

Dynamic clinical prediction models for discrete time‐to‐event data with competing risks—A case study on the OUTCOMEREA database

The development of clinical prediction models requires the selection of suitable predictor variables. Techniques to perform objective Bayesian variable selection in the linear model are well developed and have been extended to the generalized linear model setting as well as to the Cox proportional hazards model. Here, we consider discrete time‐to‐event data with competing risks and propose methodology to develop a clinical prediction model for the daily risk of acquiring a ventilator‐associated pneumonia (VAP) attributed to P. aeruginosa (PA) in intensive care units. The competing events for a PA VAP are extubation, death, and VAP due to other bacteria. Baseline variables are potentially important to predict the outcome at the start of ventilation, but may lose some of their predictive power after a certain time. Therefore, we use a landmark approach for dynamic Bayesian variable selection where the set of relevant predictors depends on the time already spent at risk. We finally determine the direct impact of a variable on each competing event through cause‐specific variable selection.     

BMA‐Mod: A Bayesian model averaging strategy for determining dose‐response relationships in the presence of model uncertainty

Successful pharmaceutical drug development requires finding correct doses. The issues that conventional dose‐response analyses consider, namely whether responses are related to doses, which doses have responses differing from a control dose response, the functional form of a dose‐response relationship, and the dose(s) to carry forward, do not need to be addressed simultaneously. Determining if a dose‐response relationship exists, regardless of its functional form, and then identifying a range of doses to study further may be a more efficient strategy. This article describes a novel estimation‐focused Bayesian approach (BMA‐Mod) for carrying out the analyses when the actual dose‐response function is unknown. Realizations from Bayesian analyses of linear, generalized linear, and nonlinear regression models that may include random effects and covariates other than dose are optimally combined to produce distributions of important secondary quantities, including test‐control differences, predictive distributions of possible outcomes from future trials, and ranges of doses corresponding to target outcomes. The objective is similar to the objective of the hypothesis‐testing based MCP‐Mod approach, but provides more model and distributional flexibility and does not require testing hypotheses or adjusting for multiple comparisons. A number of examples illustrate the application of the method.     

Infinite hidden Markov models for multiple multivariate time series with missing data

Exposure to air pollution is associated with increased morbidity and mortality. Recent technological advancements permit the collection of time-resolved personal exposure data. Such data are often incomplete with missing observations and exposures below the limit of detection, which limit their use in health effects studies. In this paper, we develop an infinite hidden Markov model for multiple asynchronous multivariate time series with missing data. Our model is designed to include covariates that can inform transitions among hidden states. We implement beam sampling, a combination of slice sampling and dynamic programming, to sample the hidden states, and a Bayesian multiple imputation algorithm to impute missing data. In simulation studies, our model excels in estimating hidden states and state-specific means and imputing observations that are missing at random or below the limit of detection. We validate our imputation approach on data from the Fort Collins Commuter Study. We show that the estimated hidden states improve imputations for data that are missing at random compared to existing approaches. In a case study of the Fort Collins Commuter Study, we describe the inferential gains obtained from our model including improved imputation of missing data and the ability to identify shared patterns in activity and exposure among repeated sampling days for individuals and among distinct individuals.     

A hierarchical model for analyzing multisite individual-level disease surveillance data from multiple systems

Passive surveillance systems are widely used to monitor diseases occurrence over wide spatial areas due to their cost-effectiveness and integration into broadly distributed healthcare systems. However, such systems are generally associated with imperfect ascertainment of disease cases and with heterogeneous capture probabilities arising from factors such as differential access to care. Augmenting passive surveillance systems with other surveillance efforts provides a way to estimate the true number of incident cases. We develop a hierarchical modeling framework for analyzing data from multiple surveillance systems that allows for individual-level covariate-dependent heterogeneous capture probabilities, and borrows information across surveillance sites to improve estimation of the true number of incident cases. Inference is carried out via a two-stage Bayesian procedure. Simulation studies illustrated superior performance of the proposed approach with respect to bias, root mean square error, and coverage compared to a model that does not borrow information across sites. We applied the proposed model to data from three surveillance systems reporting pulmonary tuberculosis (PTB) cases in a major center of ongoing transmission in China. The analysis yielded bias-corrected estimates of PTB cases from the passive system and led to the identification of risk factors associated with PTB rates, as well as factors influencing the operating characteristics of the implemented surveillance systems.     

NDAMDA: Network distance analysis for MiRNA‐disease association prediction

In recent years, microRNAs (miRNAs) are attracting an increasing amount of researchers’ attention, as accumulating studies show that miRNAs play important roles in various basic biological processes and that dysregulation of miRNAs is connected with diverse human diseases, particularly cancers. However, the experimental methods to identify associations between miRNAs and diseases remain costly and laborious. In this study, we developed a computational method named Network Distance Analysis for MiRNA‐Disease Association prediction (NDAMDA) which could effectively predict potential miRNA‐disease associations. The highlight of this method was the use of not only the direct network distance between 2 miRNAs (diseases) but also their respective mean network distances to all other miRNAs (diseases) in the network. The model's reliable performance was certified by the AUC of 0.8920 in global leave‐one‐out cross‐validation (LOOCV), 0.8062 in local LOOCV and the average AUCs of 0.8935 ± 0.0009 in fivefold cross‐validation. Moreover, we applied NDAMDA to 3 different case studies to predict potential miRNAs related to breast neoplasms, lymphoma, oesophageal neoplasms, prostate neoplasms and hepatocellular carcinoma. Results showed that 86%, 72%, 86%, 86% and 84% of the top 50 predicted miRNAs were supported by experimental association evidence. Therefore, NDAMDA is a reliable method for predicting disease‐related miRNAs.     

Modeling seasonality in space‐time infectious disease surveillance data

Infectious disease data from surveillance systems are typically available as multivariate times series of disease counts in specific administrative geographical regions. Such databases are useful resources to infer temporal and spatiotemporal transmission parameters to better understand and predict disease spread. However, seasonal variation in disease notification is a common feature of surveillance data and needs to be taken into account appropriately. In this paper, we extend a time series model for spatiotemporal surveillance counts to incorporate seasonal variation in three distinct components. A simulation study confirms that the different types of seasonality are identifiable and that a predictive approach suggested for model selection performs well. Application to surveillance data on influenza in Southern Germany reveals a better model fit and improved one‐step‐ahead predictions if all three components allow for seasonal variation.     

Analyzing time‐ordered event data with missed observations

A common problem with observational datasets is that not all events of interest may be detected. For example, observing animals in the wild can difficult when animals move, hide, or cannot be closely approached. We consider time series of events recorded in conditions where events are occasionally missed by observers or observational devices. These time series are not restricted to behavioral protocols, but can be any cyclic or recurring process where discrete outcomes are observed. Undetected events cause biased inferences on the process of interest, and statistical analyses are needed that can identify and correct the compromised detection processes. Missed observations in time series lead to observed time intervals between events at multiples of the true inter‐event time, which conveys information on their detection probability. We derive the theoretical probability density function for observed intervals between events that includes a probability of missed detection. Methodology and software tools are provided for analysis of event data with potential observation bias and its removal. The methodology was applied to simulation data and a case study of defecation rate estimation in geese, which is commonly used to estimate their digestive throughput and energetic uptake, or to calculate goose usage of a feeding site from dropping density. Simulations indicate that at a moderate chance to miss arrival events (p = 0.3), uncorrected arrival intervals were biased upward by up to a factor 3, while parameter values corrected for missed observations were within 1% of their true simulated value. A field case study shows that not accounting for missed observations leads to substantial underestimates of the true defecation rate in geese, and spurious rate differences between sites, which are introduced by differences in observational conditions. These results show that the derived methodology can be used to effectively remove observational biases in time‐ordered event data.     

A new mixed‐effects regression model for the analysis of zero‐modified hierarchical count data

Count data sets are traditionally analyzed using the ordinary Poisson distribution. However, such a model has its applicability limited as it can be somewhat restrictive to handle specific data structures. In this case, it arises the need for obtaining alternative models that accommodate, for example, (a) zero‐modification (inflation or deflation at the frequency of zeros), (b) overdispersion, and (c) individual heterogeneity arising from clustering or repeated (correlated) measurements made on the same subject. Cases (a)–(b) and (b)–(c) are often treated together in the statistical literature with several practical applications, but models supporting all at once are less common. Hence, this paper's primary goal was to jointly address these issues by deriving a mixed‐effects regression model based on the hurdle version of the Poisson–Lindley distribution. In this framework, the zero‐modification is incorporated by assuming that a binary probability model determines which outcomes are zero‐valued, and a zero‐truncated process is responsible for generating positive observations. Approximate posterior inferences for the model parameters were obtained from a fully Bayesian approach based on the Adaptive Metropolis algorithm. Intensive Monte Carlo simulation studies were performed to assess the empirical properties of the Bayesian estimators. The proposed model was considered for the analysis of a real data set, and its competitiveness regarding some well‐established mixed‐effects models for count data was evaluated. A sensitivity analysis to detect observations that may impact parameter estimates was performed based on standard divergence measures. The Bayesian ‐value and the randomized quantile residuals were considered for model diagnostics.