A Bayesian hierarchical sparse factor model for estimating simultaneous covariance matrices for gestational outcomes in consecutive pregnancies.

Covariance estimation for multiple groups is a key feature for drawing inference from a heterogeneous population. One should seek to share information about common features in the dependence structures across the various groups. In this paper, we introduce a novel approach for estimating the covariance matrices for multiple groups using a hierarchical latent factor model that shrinks the factor loadings across groups toward a global value. Using a sparse spike and slab model on these loading coefficients allows for a sparse formulation of our model. Parameter estimation is accomplished through a Markov chain Monte Carlo scheme, and a model selection approach is used to select the number of factors to use. We validate our model through extensive simulation studies. Finally, we apply our methodology to the NICHD Consecutive Pregnancies Study to estimate the correlations between birth weights and gestational ages of three consecutive birth within four different subgroups (underweight, normal, overweight, and obese) of women.

Decreasing alcohol use among young adults presenting for service in the U.S. Air Force: An epidemiological surveillance study.

U.S. surveys demonstrate recent decreases in the prevalence of alcohol use and binge drinking among young adults. The current study aims to determine whether similar trends are evident in a similarly aged cohort of service members in the US Air Force to inform ongoing prevention efforts. Participants were 103,240 Air Force personnel in entry-level training between 2016 and 2019. Participants anonymously completed the AUDIT (Alcohol Use Disorder Identification Test) regarding their pre-service drinking. Logistic regression analyses and the Cochran-Armitage test were conducted to measure population trends over the study duration with stratification by age (<21 vs. ≥21) and evaluation of specific alcohol behaviors. Between 2016 and 2019, the proportion of young service members endorsing any alcohol use significantly decreased for both the <21 group (i.e. from 38.9% to 32.6%) and the ≥21 group (i.e. from 80.6% to 77.5%). Among those who endorsed drinking, a decrease over time in binge use was also observed from 46.6% to 37.8% for the <21 group and from 34.2% to 27.5% for the ≥21 group. Responses to other specific alcohol risk items and total AUDIT scores also demonstrated decreases. Binge use and risky drinking remained disproportionately common among those under the legal drinking age. It is encouraging to observe a shift toward abstinence and decreased binge use among this population of young military recruits. However, given the risk for many adverse health and legal consequences in this population, more work is needed to prevent problematic drinking, especially among those under the legal drinking age.

Interobserver Agreement on the Interpretation of Programmed Death-ligand 1 (PD-L1) Combined Positive Score (CPS) Among Gynecologic Pathologists.

The anti-programmed cell death (PD-1) checkpoint inhibitor pembrolizumab is approved for the treatment of cervical carcinoma with a programmed cell death-ligand 1 (PD-L1) Combined Positive Score (CPS) of ≥1. We assessed interobserver agreement in cervical carcinoma PD-L1 CPS to identify whether it may affect patient selection for immunotherapeutic candidacy. Twenty-nine cervical carcinomas were stained for PD-L1 (Dako 22C3), and slides were interpreted by 5 subspecialty-trained gynecologic pathologists with experience reading PD-L1 immunohistochemistry. Expression was scored using CPS and read out as positive (≥1) or negative (<1); in positive cases, a final score was assigned (1 to 100). There was consensus agreement across all 5 pathologists for 90% (26/29) (Fleiss Kappa value for interobserver agreement: 0.799). The 3 cases with disagreement were composed of 2 squamous cell carcinomas and 1 small cell carcinoma. Of the 26 with unanimous agreement, 88% (23/26) were positive and 12% (3/26) were negative. All (16/16) pure squamous cell carcinomas with full consensus were interpreted as positive, whereas tumors with glandular components were commonly consensus negative (33%, 3/9); this difference was significant ( P =0.037). Disagreements were attributable to low CPS versus negative reads (2 cases) and difficulty discerning glandular involvement from pushing invasion (1 case). In summary, experienced gynecologic pathologists showed substantial interobserver agreement in the interpretation of PD-L1 CPS at the Food and Drug Administration-approved treatment threshold, with the majority of tumors being classified as positive. Pure squamous histology was strongly associated with a consensus-positive read, whereas a subset of tumors with glandular differentiation was negative by all readers. Disagreements occurred in tumors with low versus negative CPS values and in the setting of limited invasion.

An imputation approach for fitting two-part mixed effects models for longitudinal semi-continuous data.

Two-part mixed effects models are often used for analyzing longitudinal data with many zeros. Typically, these models are formulated with binary and continuous components separately with random effects that are correlated between the two components. Researchers have developed maximum-likelihood and Bayesian approaches for fitting these models that often require using particular software packages or very specialized software. We propose an imputation approach that will allow practitioners to separately use standard linear and generalized linear mixed models to estimate the fixed effects for two-part mixed effects models with complex random effects structures. An approximation to the conditional distribution of positive measurements given an individual's pattern of non-zero measurements is proposed that can be easily estimated and then imputed from. We show that for a wide range of parameter values, the imputation approach results in nearly unbiased estimation and can be implemented with standard software. We illustrate the proposed imputation approach for the analysis of longitudinal clinical trial data with many zeros.