Bayesian generalized linear models for meta-analysis of diagnostic tests

With the recognition of the importance of evidence-based medicine, there is an emerging need for methods to systematically synthesize available data. Specifically, methods to provide accurate estimates of test characteristics for diagnostic tests are needed to help physicians make better clinical decisions. To provide more flexible approaches for meta-analysis of diagnostic tests, we developed three Bayesian generalized linear models. Two of these models, a bivariate normal and a binomial model, analyzed pairs of sensitivity and specificity values while incorporating the correlation between these two outcome variables. Noninformative independent uniform priors were used for the variance of sensitivity, specificity and correlation. We also applied an inverse Wishart prior to check the sensitivity of the results. The third model was a multinomial model where the test results were modeled as multinomial random variables. All three models can include specific imaging techniques as covariates in order to compare performance. Vague normal priors were assigned to the coefficients of the covariates. The computations were carried out using the 'Bayesian inference using Gibbs sampling' implementation of Markov chain Monte Carlo techniques. We investigated the properties of the three proposed models through extensive simulation studies. We also applied these models to a previously published meta-analysis dataset on cervical cancer as well as to an unpublished melanoma dataset. In general, our findings show that the point estimates of sensitivity and specificity were consistent among Bayesian and frequentist bivariate normal and binomial models. However, in the simulation studies, the estimates of the correlation coefficient from Bayesian bivariate models are not as good as those obtained from frequentist estimation regardless of which prior distribution was used for the covariance matrix. The Bayesian multinomial model consistently underestimated the sensitivity and specificity regardless of the sample size and correlation coefficient. In conclusion, the Bayesian bivariate binomial model provides the most flexible framework for future applications because of its following strengths: (1) it facilitates direct comparison between different tests; (2) it captures the variability in both sensitivity and specificity simultaneously as well as the intercorrelation between the two; and (3) it can be directly applied to sparse data without ad hoc correction. ^

RNA-SEQUENCING APPLICATIONS: GENE EXPRESSION QUANTIFICATION AND METHYLATOR PHENOTYPE IDENTIFICATION

My dissertation focuses on two aspects of RNA sequencing technology. The first is the methodology for modeling the overdispersion inherent in RNA-seq data for differential expression analysis. This aspect is addressed in three sections. The second aspect is the application of RNA-seq data to identify the CpG island methylator phenotype (CIMP) by integrating datasets of mRNA expression level and DNA methylation status. Section 1: The cost of DNA sequencing has reduced dramatically in the past decade. Consequently, genomic research increasingly depends on sequencing technology. However it remains elusive how the sequencing capacity influences the accuracy of mRNA expression measurement. We observe that accuracy improves along with the increasing sequencing depth. To model the overdispersion, we use the beta-binomial distribution with a new parameter indicating the dependency between overdispersion and sequencing depth. Our modified beta-binomial model performs better than the binomial or the pure beta-binomial model with a lower false discovery rate. Section 2: Although a number of methods have been proposed in order to accurately analyze differential RNA expression on the gene level, modeling on the base pair level is required. Here, we find that the overdispersion rate decreases as the sequencing depth increases on the base pair level. Also, we propose four models and compare them with each other. As expected, our beta binomial model with a dynamic overdispersion rate is shown to be superior. Section 3: We investigate biases in RNA-seq by exploring the measurement of the external control, spike-in RNA. This study is based on two datasets with spike-in controls obtained from a recent study. We observe an undiscovered bias in the measurement of the spike-in transcripts that arises from the influence of the sample transcripts in RNA-seq. Also, we find that this influence is related to the local sequence of the random hexamer that is used in priming. We suggest a model of the inequality between samples and to correct this type of bias. Section 4: The expression of a gene can be turned off when its promoter is highly methylated. Several studies have reported that a clear threshold effect exists in gene silencing that is mediated by DNA methylation. It is reasonable to assume the thresholds are specific for each gene. It is also intriguing to investigate genes that are largely controlled by DNA methylation. These genes are called “L-shaped” genes. We develop a method to determine the DNA methylation threshold and identify a new CIMP of BRCA. In conclusion, we provide a detailed understanding of the relationship between the overdispersion rate and sequencing depth. And we reveal a new bias in RNA-seq and provide a detailed understanding of the relationship between this new bias and the local sequence. Also we develop a powerful method to dichotomize methylation status and consequently we identify a new CIMP of breast cancer with a distinct classification of molecular characteristics and clinical features.

An empirical model to estimate the demand for primary care in urban settings

Objective. To measure the demand for primary care and its associated factors by building and estimating a demand model of primary care in urban settings.^ Data source. Secondary data from 2005 California Health Interview Survey (CHIS 2005), a population-based random-digit dial telephone survey, conducted by the UCLA Center for Health Policy Research in collaboration with the California Department of Health Services, and the Public Health Institute between July 2005 and April 2006.^ Study design. A literature review was done to specify the demand model by identifying relevant predictors and indicators. CHIS 2005 data was utilized for demand estimation.^ Analytical methods. The probit regression was used to estimate the use/non-use equation and the negative binomial regression was applied to the utilization equation with the non-negative integer dependent variable.^ Results. The model included two equations in which the use/non-use equation explained the probability of making a doctor visit in the past twelve months, and the utilization equation estimated the demand for primary conditional on at least one visit. Among independent variables, wage rate and income did not affect the primary care demand whereas age had a negative effect on demand. People with college and graduate educational level were associated with 1.03 (p < 0.05) and 1.58 (p < 0.01) more visits, respectively, compared to those with no formal education. Insurance was significantly and positively related to the demand for primary care (p < 0.01). Need for care variables exhibited positive effects on demand (p < 0.01). Existence of chronic disease was associated with 0.63 more visits, disability status was associated with 1.05 more visits, and people with poor health status had 4.24 more visits than those with excellent health status. ^ Conclusions. The average probability of visiting doctors in the past twelve months was 85% and the average number of visits was 3.45. The study emphasized the importance of need variables in explaining healthcare utilization, as well as the impact of insurance, employment and education on demand. The two-equation model of decision-making, and the probit and negative binomial regression methods, was a useful approach to demand estimation for primary care in urban settings.^