Faculty Bibliography

Latent class models are widely used to identify unobserved subgroups (i.e., latent classes) based upon one or more manifest variables. The probability of belonging to each subgroup is typically modeled as a function of a set of measured covariates. In this paper, we extend existing latent class models to incorporate matrix covariates. This research is motivated by a randomized placebo-controlled depression clinical trial. One study goal is to identify a subgroup of subjects who experience symptoms improvement early on during antidepressant treatment, which is considered to be an indication of a placebo rather than a true pharmacological response. We want to relate the likelihood of belonging to this subgroup of early responders to baseline electroencephalography (EEG) measurement that takes the form of a matrix. The proposed method is built upon a low rank Candecomp/Parafac (CP) decomposition of the target coefficient matrix through low-dimensional latent variables, which effectively reduces the model dimensionality. We adopt a Bayesian hierarchical modeling approach to estimate the latent variables, which allows a flexible way to incorporate prior knowledge about covariate effect heterogeneity and offers a data-driven method of regularization. Simulation studies suggest that the proposed method is robust against potentially misspecified rank in the CP decomposition. With the motivating example we show how the proposed method can be applied to extract valuable information from baseline EEG measurements that explains the likelihood of belonging to the early responder subgroup, helping to identify placebo responders and suggesting new targets for the study of placebo response.

The dentate gyrus (DG) principal cells are glutamatergic granule cells (GCs), and they are located in a compact cell layer. However, GCs are also present in the adjacent hilar region, but have been described in only a few studies. Therefore, we used the transcription factor prospero homeobox 1 (Prox1) to quantify GCs at postnatal day (PND) 16, 30, and 60 in a common mouse strain, C57BL/6J mice. At PND16, there was a large population of Prox1-immunoreactive (ir) hilar cells, with more in the septal than temporal hippocampus. At PND30 and 60, the size of the hilar Prox1-ir cell population was reduced. Similar numbers of hilar Prox1-expressing cells were observed in PND30 and 60 Swiss Webster mice. Prox1 is usually considered to be a marker of postmitotic GCs. However, many Prox1-ir hilar cells, especially at PND16, were not double-labeled with NeuN, a marker typically found in mature neurons. Most hilar Prox1-positive cells at PND16 co-expressed doublecortin (DCX) and calretinin, markers of immature GCs. Double-labeling with a marker of actively dividing cells, Ki67, was not detected. These results suggest that, surprisingly, a large population of cells in the hilus at PND16 are immature GCs (Type 2b and Type 3 cells). We also asked whether hilar Prox1-ir cell numbers are modifiable. To examine this issue, we conditionally deleted the proapoptotic gene BAX in Nestin-expressing cells at a time when there are numerous immature GCs in the hilus, PND2-8. When these mice were examined at PND60, the numbers of Prox1-ir hilar cells were significantly increased compared to control mice. However, deletion of BAX did not appear to change the proportion that co-expressed NeuN, suggesting that the size of the hilar Prox1-expressing population is modifiable. However, deleting BAX, a major developmental disruption, does not appear to change the proportion that ultimately becomes neurons.

Gender inequalities shape the experience of food insecurity among women living with HIV (WLHIV). We systematically reviewed the impact of food insecurity on sexual risk behaviors and antiretroviral therapy (ART) adherence among WLHIV. We included qualitative or quantitative peer-reviewed articles, extracted data in duplicate, and assessed rigor. Seven studies, from sub-Saharan Africa, North America, and Europe, met inclusion criteria. Food insecurity was associated with increased sexual risk through transactional sex and inability to negotiate safer sex. Hunger and food insecurity were barriers to ART initiation/adherence. Multidimensional programming and policies should simultaneously address poverty, gender inequality, food insecurity, and HIV.

The negative consequences of untreated postpartum depression (PD) for both the woman and her infant are well established. The impact of maternal depression has led to recommendations on systematic perinatal depression screening. Unfortunately, large-scale initiatives on PD screening have found no benefit unless systems are in place to facilitate appropriate interventions for women who screen positive. Pediatric primary care has been a focus of efforts to support screening and management of PD because pediatric providers, unlike adult healthcare providers, have the most frequent contact with postpartum women through well-child visits. Well-child visits thus present an unparalleled opportunity to detect and intervene with PD. Literature reviews suggest that specific strategies are feasible within pediatric settings and could benefit both the woman and her child. In this article, we present a stepped care approach for screening and managing PD, integrating common elements found in existing pediatric-based models. A stepped care approach is ideal because PD is a heterogeneous condition, with a range of presentations and hence responsiveness to various interventions. This care pathway begins with systematic screening for depression symptoms, followed by a systematic risk assessment for women who screen positive and care management based on risk profiles and responsiveness. This approach allows pediatric providers to be optimally flexible and responsive in addressing the majority of women with PD within the context of the family-centered medical home to improve child well-being. Challenges to managing PD within pediatrics are discussed, including strategies for addressing them. Implications for research, policy, and practice are discussed.

Discrepancy in perception of bullying experiences may lead to later internalizing or externalizing behavior in adolescents. A 1,663 South Korean 7th and 8th graders (mean age: 13.1 and 14.1 years old), were seen for a follow-up study to examine the relationships between the discrepancy in perception of their bullying experiences (defined as discrepancy between self- and peer-reports of bullying experiences) and internalizing or externalizing behavior at follow-up. Bullying was assessed by self- and peer-report. The discrepancy in perception of bullying experiences was defined by the concordance or discordance between self- and peer-reports. Internalizing and externalizing behavior was evaluated using the Youth Self Report and Child Behavior Checklist, at baseline and follow-up. Two by two ANCOVA was performed with a factorial design, categorizing discrepancy in perception of bullying experiences based on the agreement between self-report and peer-report. Internalizing/externalizing behavior-at-follow-up was used as an outcome, adjusting for other known risk factors for internalizing/externalizing behavior, including baseline internalizing/externalizing behavior, and bullying experiences. Adolescents with perceptions of bullying experiences discrepant from peer-reports showed increased internalizing/externalizing behavior at follow-up. Bullying also stands out as an independent risk factor for the development of future externalizing behavior even among adolescents with accurate perceptions of bullying experiences. These specific groups of youth warrant more focused assessment and intervention.

The youth health indicator (YHI) program was developed to improve health and reduce risk behaviors for youth treated in clinic and day treatment psychiatric settings. This study examined implementation of the YHI program and factors associated with BMI % change for youth participating in the program. Outpatient facilities which implemented the YHI program were surveyed (N = 10) and lessons learned were summarized. Mixed random effects repeated measures analysis was used to examine BMI % trajectory for youth admitted during 2010-2014 with BMI % measured (N = 6403). Treatment settings indicated a variety of strategies to undertake and sustain the YHI program. Factors associated with BMI % change over time were identified. The YHI program resulted in the development of a wide variety of programmatic innovations targeted at improving youth health. Continued work needs to be done to improve the health outcomes for youth in these treatment settings.

Caregiver report is the most common measure of change in pediatric psychiatry. Yet, placebo response rates pose significant challenges to reliably detect a treatment response. The present study simulated an eight-week clinical trial protocol for Autism Spectrum Disorder (ASD) for the purpose of testing the feasibility and validity of several outcome measures. Twenty caregivers answered questions about their child's behavior on their smartphone each week and completed a battery of paper questionnaires during weeks one and eight. No treatment was administered. Caregivers reported a significant decrease in problem behaviors on the Aberrant Behavior Checklist (ABC) (29% decrease) and general ASD behaviors on the Social Responsiveness Scale (SRS) (7% decrease). There was also a trend of behavior improvement from smartphone questions but no significant changes in clinical ratings of core diagnostic features of ASD. Participation in a comprehensive protocol in the absence of a particular treatment significantly influenced how caregivers perceived the severity of their children's problem behaviors. These placebo-like effects represent substantial challenges for randomized controlled trials (RCTs) that use treatment as usual and have implications for future behavioral and pharmacological treatment trial designs. Autism Res 2017, 10: 1567-1572. © 2017 International Society for Autism Research, Wiley Periodicals, Inc.

INTRODUCTION AND AIMS:There is limited research regarding the effects of alcohol consumption by breastfeeding mothers on infant development. This study examined the frequency, correlates and outcomes of alcohol use during lactation. DESIGN AND METHODS:Data were from an Australian cohort study. Maternal demographics and substance use were assessed during pregnancy and at 8 weeks and 12 months postpartum. Breastfeeding duration, infant feeding, sleeping and development (Ages and Stages Questionnaire) were also assessed postpartum. Logistic regression and general linear model analyses examined characteristics of women who drank during breastfeeding, and the association between alcohol use during breastfeeding and infant outcomes. RESULTS:Alcohol use was reported by 60.7% and 69.6% of breastfeeding women at 8 weeks and 12 months postpartum, respectively. Breastfeeding women who consumed alcohol were more likely to be born in Australia or another English-speaking country, be tertiary educated and have higher household incomes. Most drank at low levels (≤14 standard drinks per week, <3 per occasion) and employed strategies (e.g. timing of alcohol use) to minimise alcohol passed onto infants via breastmilk. Alcohol consumption was unrelated to breastfeeding duration, infant feeding and sleeping behaviour at 8 weeks, and most infant developmental outcomes at 8 weeks or 12 months, after adjusting for confounders. The only significant association showed that infants whose mothers drank at 8 weeks postpartum had more favourable results for personal-social development at 12 months compared with those whose mothers abstained. DISCUSSION AND CONCLUSIONS:Low level drinking during breastfeeding is not linked with shorter breastfeeding duration or adverse outcomes in infants up to 12 months of age. [Wilson J, Tay RY, McCormack C, Allsop S, Najman J, Burns L, Olsson CA, Elliott E, Jacobs S, Mattick RP, Hutchinson D. Alcohol consumption by breastfeeding mothers: Frequency, correlates and infant outcomes. Drug Alcohol Rev 2017;00:000-000].