Biomedical subjects

Minje Sung

Publications and source records attributed to Minje Sung.

3 recordsLinked to original sources

PubMed2007

Bayesian analysis of non-homogeneous Markov chains: application to mental health data.

In this paper we present a formal treatment of non-homogeneous Markov chains by introducing a hierarchical Bayesian framework. Our work is motivated by the analysis of correlated categorical data which arise in assessment of psychiatric treatment programs. In our development, we introduce a Markovian structure to describe the non-homogeneity of transition patterns. In doing so, we introduce a logistic regression set-up for Markov chains and incorporate covariates in our model. We present a Bayesian model using Markov chain Monte Carlo methods and develop inference procedures to address issues encountered in the analyses of data from psychiatric treatment programs. Our model and inference procedures are implemented to some real data from a psychiatric treatment study.

Adolescent↗

PubMed2006

Bayesian semi-parametric ROC analysis.

This paper describes a semi-parametric Bayesian approach for estimating receiver operating characteristic (ROC) curves based on mixtures of Dirichlet process priors (MDP). We address difficulties in modelling the underlying distribution of screening scores due to non-normality that may lead to incorrect choices of diagnostic cut-offs and unreliable estimates of prevalence of the disease. MDP is a robust tool for modelling non-standard diagnostic distributions associated with imperfect classification of an underlying diseased population, for example, when a diagnostic test is not a gold standard. For posterior computations, we propose an efficient Gibbs sampling framework based on a finite-dimensional approximation to MDP. We show, using both simulated and real data sets, that MDP modelling for ROC curve estimation closely parallels the frequentist kernel density estimation (KDE) approach.

Adolescent↗

PubMed2004

Effects of age at first substance use and psychiatric comorbidity on the development of substance use disorders.

In this paper, we examine the effects of age at first substance use, and history of psychiatric disorders, on the development of substance use disorder (SUD) by age 16. We use a prospective, longitudinal design to disaggregate the effects of age at first use and time since first use on the development of adolescent SUD. Second, we test the hypothesis that adolescent SUD is an unlikely progression from early substance use unless children also show other early conduct problems. A population sample of 1,420 children from the Great Smoky Mountains Study (GSMS) was assessed annually between ages 9 and 16. Logistic regression models were applied within the hierarchical Bayesian framework, where the covariate effects were described by time-varying parameters having a first-order auto-regressive prior distribution. Posterior analyses based on a Gibbs sampling approach revealed that, controlling for years of exposure, the risk of transition to SUD increased with age at onset for onsets before age 13, but began to fall for onset at 14. Among users, use alone, without early conduct problems, led to a 11% prevalence of SUD by age 16. Past conduct disorder (CD) had a strong additive effect at ages 13-15, but at age 16, when substance use and abuse became more normative, the excess risk from prior CD decreased. Boys, but not girls, with a history of depression were at increased risk of SUD. Anxiety increased the risk of SUD in girls at age 16, but not before that. Results only partially support the study hypothesis; early use was a major predictor of adolescent SUD even in the absence of CD.

Adolescent↗