Non-linear regression models for Approximate Bayesian Computation

Approximate Bayesian inference on the basis of summary statistics is well-suited to complex problems for which the likelihood is either mathematically or computationally intractable. However the methods that use rejection suffer from the curse of dimensionality when the number of summary statistics is increased. Here we propose a machine-learning approach to the estimation of the posterior density by introducing two innovations. The new method fits a nonlinear conditional heteroscedastic regression of the parameter on the summary statistics, and then adaptively improves estimation using importance sampling. The new algorithm is compared to the state-of-the-art approximate Bayesian methods, and achieves considerable reduction of the computational burden in two examples of inference in statistical genetics and in a queueing model.     

A computer model for unconscious spread of anxiety-linked inhibition in cognitive networks

Unconscious inhibitory processes, triggered by a potential anxiety reaction, are reviewed in the context of an emerging rapprochement between psychodynamic and cognitive approaches in experimental psychology. Conditions underlying spread of inhibitory action to other cognitive networks are first explored in three tachistoscopic experiments utilizing words posthypnotically tied to a potential anxiety, pleasure, or neutral reaction. Response times of subjects, instructed to ignore those words while naming pictures or solving anagrams as quickly as possible, reveal a highly differentiated pattern of circumstances governing likelihood of inhibitory spread from anxiety-linked words to target stimuli. Next a computer model is constructed to simulate cognitive processes from onset of display to eventual response, and the model is then tested for its fit to the empirical data. Finally, an illustrative study shows that a subset of computer-generated predictions for spread of inhibitory action is verifiable experimentally.     

To adjust or not to adjust for baseline when analyzing repeated binary responses? The case of complete data when treatment comparison at study end is of interest

The benefits of adjusting for baseline covariates are not as straightforward with repeated binary responses as with continuous response variables. Therefore, in this study, we compared different methods for analyzing repeated binary data through simulations when the outcome at the study endpoint is of interest. Methods compared included chi‐square, Fisher's exact test, covariate adjusted/unadjusted logistic regression (Adj.logit/Unadj.logit), covariate adjusted/unadjusted generalized estimating equations (Adj.GEE/Unadj.GEE), covariate adjusted/unadjusted generalized linear mixed model (Adj.GLMM/Unadj.GLMM). All these methods preserved the type I error close to the nominal level. Covariate adjusted methods improved power compared with the unadjusted methods because of the increased treatment effect estimates, especially when the correlation between the baseline and outcome was strong, even though there was an apparent increase in standard errors. Results of the Chi‐squared test were identical to those for the unadjusted logistic regression. Fisher's exact test was the most conservative test regarding the type I error rate and also with the lowest power. Without missing data, there was no gain in using a repeated measures approach over a simple logistic regression at the final time point. Analysis of results from five phase III diabetes trials of the same compound was consistent with the simulation findings. Therefore, covariate adjusted analysis is recommended for repeated binary data when the study endpoint is of interest. Copyright © 2015 John Wiley & Sons, Ltd.     

History of sexual abuse and parental alcohol misuse: Risk,outcomes and protective factors in adolescents

This study examined the school performance, suicidal involvement, disordered eating behaviors, pregnancy risk and chemical use of teenagers with a history of sexual abuse or parental alcohol misuse and found that they had higher rates of these adverse outcomes than among a comparison group of teenagers without such background risk factors. Adolescents with dual-risk background reported higher levels of suicide risk, disordered eating, sexual behaviors and chemical abuse than teenagers with only one background risk factor. Among index group members, protective factors against these adverse outcomes included a high degree of religiosity and the ability to discuss problems with family or friends. Risk factors that increased the likelihood of adverse outcomes included depression, perceived substance use in school and worries about family financial security.The first author was supported in the preparation of this article by a grant from the Lois and Samuel Silberman Fund of New York.     

Multiple imputation compared with restricted pseudo‐likelihood and generalized estimating equations for analysis of binary repeated measures in clinical studies

Non‐likelihood‐based methods for repeated measures analysis of binary data in clinical trials can result in biased estimates of treatment effects and associated standard errors when the dropout process is not completely at random. We tested the utility of a multiple imputation approach in reducing these biases. Simulations were used to compare performance of multiple imputation with generalized estimating equations and restricted pseudo‐likelihood in five representative clinical trial profiles for estimating (a) overall treatment effects and (b) treatment differences at the last scheduled visit. In clinical trials with moderate to high (40–60%) dropout rates with dropouts missing at random, multiple imputation led to less biased and more precise estimates of treatment differences for binary outcomes based on underlying continuous scores. Copyright © 2005 John Wiley & Sons, Ltd.     

Unmet service needs of families of young children with chronic illnesses and disabilities

Data about health and educational services were collected from 187 families with young children (under two years) who have chronic illnesses and disabilities. The purpose was to assess parents' satisfaction with services their families received and to identify families' unmet service needs as well as the reasons for these unmet needs. Although families' ratings of the adequacy of services are relatively high, 28% of the families report unmet service needs. The unmet needs most often reported are occupational, physical, or speech therapy; respite care/child care; and special education. Lack of funding is the most commonly given reason for unmet needs. Children with multiple impairments are at highest risk for unmet needs. Only 22% of unmet needs reported at the initial interview have been met by the time of the six-month follow-up. Increased coordination and monitoring of services across service sectors are indicated. Her research interests include the effects of disability on family identity and families' responses to minority stress. She received her Ph.D. from the University of Minnesota. She gratefully acknowledges Shelley Blazis and Nadav Casuto for statistical consulting. Her research interests focus on the impact of chronic illness and disability on families. She received her Ph.D. from the University of Minnesota. Her research interests include families and health and chronic illness and disability in children. She received her Ph.D. in Family Social Science at the University of Minnesota. His areas of interest include neurodevelopmental outcomes of low birth weight infants and training in developmental pediatrics. He was graduated from the University of Minnesota Medical School. His areas of interest include adolescents with chronic illness and disabilities, adolescent sexual decision making, and international adolescent health care issues. He was graduated from Howard University College of Medicine was awarded his earned doctorate in Health Policy from the University of Minnesota. Preparation of this article was supported by the National Institute on Disability and Rehabilitation Research Grant #H133890012.     

A Frailty Model in Crossover Studies with Time-to-Event Response Variable

In this article, we develop a model to study treatment, period, carryover, and other applicable effects in a crossover design with a time-to-event response variable. Because time-to-event outcomes on different treatment regimens within the crossover design are correlated for an individual, we adopt a proportional hazards frailty model. If the frailty is assumed to have a gamma distribution, and the hazard rates are piecewise constant, then the likelihood function can be determined via closed-form expressions. We illustrate the methodology via an application to a data set from an asthma clinical trial and run simulations that investigate sensitivity of the model to data generated from different distributions.