This dissertation aims to explore the relationship between social support and social and psychological wellbeing in the adult population of sexual and gender minorities, or Lesbian, Gay, Bisexual, Transgender (LGBT) adults. We apply Antonucci’s (1987) Convoy Model of Social Relations as a lens through which to conceptualize social support across four dimensions: structure, function, quality, and closeness. Chapter One contains an introduction to the LGBT adult population; a description of the Convoy Model of Social Relations and of technology used to exchange social support; and an outline of the specific aims addressed in this dissertation. Chapter Two is an integrative review of literature about social support networks of LGBT older adults age 50 years and older. Chapter Three is a study designed to investigate relationships between LGBT identity, social support characteristics, the use of online social communication, and psychological and social wellbeing. Chapter Four is a qualitative descriptive study that explores LGBT older adults’ attitudes, perceptions, and uses of technology for social connectedness and support during COVID-19. Chapter Five contains a synthesis of all findings in this dissertation; a discussion of the results as they relate to the Convoy Model of Social Relations; and implications for clinical interventions and future research.

The theme of this dissertation focuses on developing statistical models to learn progression dynamics and mechanisms of neuropsychiatric disorders using data from various domains. Due to limited knowledge about the underlying pathological processes in neurological disorders, it remains a challenge to establish reliable diagnostic criteria and predict disease prognosis in the presence of substantial phenotypic heterogeneity. As a result, current diagnosis and treatment of neurological disorders often rely on late-stage clinical symptoms, which poses barriers for developing effective interventions at the premanifest stage. It is crucial to characterize the temporal disease progression course and study the underlying mechanisms using clinical assessments, blood biomarkers, and neuroimaging biomarkers to evaluate disease stages, identify markers that are useful for early clinical diagnosis, compare or monitor treatment effects and accelerate drug discovery. We propose three projects to tackle challenges in leveraging multi-domain biomarkers and clinical symptoms to learn disease dynamics and progression of neurological disorders: (1) A nonlinear mixture model with subject-specific random inflection points to jointly fit multiple longitudinal markers and estimate marker progression trajectories in a single modality; (2) A multi-layer exponential family factor model integrating multi-domain data to learn lower-dimensional latent space of disease impairment and fully map disease risk and progression; (3) A latent state space model that jointly analyzes multi-channel EEG signals and learns dynamics of different sources corresponding to brain cortical activities. In addition, motivated by the ongoing COVID-19 pandemic, we propose a parsimonious survival-convolution model to predict daily new cases and estimate the time-varying reproduction numbers to evaluate effects of mitigation strategies. In the first project, we propose a nonlinear mixture model with random time shifts to jointly estimate long-term progression trajectories using multivariate discrete longitudinal outcomes. The model can identify early disease markers, their orders of occurrence, and the rates of impairment. Specifically, a latent binary variable representing disease susceptibility status incorporates subject covariates (e.g., biological measures) in the mixture model to capture between-subject heterogeneity. Measures of disease impairment for susceptible patients are modeled jointly under the exponential family framework. Our model allows for subject-specific and marker-specific inflection points associated with patients' characteristics (e.g., genetic mutation) to indicate a critical time when the fastest degeneration occurs. Furthermore, it uses subject-specific latent scores shared among markers to improve efficiency. The model is estimated using an EM algorithm. Extensive simulation studies are conducted to demonstrate validity of the proposed method and algorithm. Lastly, we apply our method to the Parkinson's Progression Markers Initiative (PPMI), and show utility to identify early disease signs and compare clinical symptomatology for the genetic form of Parkinson's Disease (PD) and idiopathic PD. In the second project, we tackle challenges to leverage multi-domain markers to learn early disease progression of neurological disorders. We propose to integrate heterogeneous types of measures from multiple domains (e.g., discrete clinical symptoms, ordinal cognitive markers, continuous neuroimaging and blood biomarkers) using a hierarchical Multi-layer Exponential Family Factor (MEFF) model, where the observations follow exponential family distributions with lower-dimensional latent factors. The latent factors are decomposed into shared factors across multiple domains and domain-specific factors, where the shared factors provide robust information to perform behavioral phenotyping and partition patients into clinically meaningful and biologically homogeneous subgroups. Domain-specific factors capture the remaining unique variations for each domain. The MEFF model also captures the nonlinear trajectory of disease progression and order critical events of neurodegeneration measured by each marker. To overcome computational challenges, we fit our model by approximate inference techniques for large-scale data. We apply the developed method to Parkinson's Progression Markers Initiative (PPMI) data to integrate biological, clinical and cognitive markers arising from heterogeneous distributions. The model learns lower-dimensional representations of Parkinson's disease and the temporal ordering of the neurodegeneration of PD. In the third project, we propose methods that can be used to analyze multi-channel electroencephalogram (EEG) signals intensively measured at a high temporal resolution. Modern neuroimaging technologies have substantially advanced the measurement of brain activities. EEG as a non-invasive neuroimaging technique measures changes in electrical voltage on the scalp induced by cortical activities. With its high temporal resolution, EEG has emerged as an increasingly useful tool to study brain connectivity. Challenges with modeling EEG signals of complex brain activities include interactions among unknown sources, low signal-to-noise ratio and substantial between-subject heterogeneity. In this work, we propose a state space model that jointly analyzes multi-channel EEG signals and learns dynamics of different sources corresponding to brain cortical activities. Our model borrows strength from spatially correlated measurements and uses low-dimensional latent sources to explain all observed channels. The model can account for patient heterogeneity and quantify the effect of a subject's covariates on the latent space. The EM algorithm, Kalman filtering, and bootstrap resampling are used to fit the state space model and provide comparisons between patient diagnostic groups. We apply the developed approach to a case-control study of alcoholism and reveal significant attenuation of brain activities in response to visual stimuli in alcoholic subjects compared to healthy controls. Lastly, motivated by the ongoing COVID-19 pandemic, we propose a robust and parsimonious survival-convolution model aiming to predict COVID-19 disease course and compare effectiveness of mitigation measures across countries to inform policy decision making. We account for transmission during a pre-symptomatic incubation period and use a time-varying effective reproduction number to reflect the temporal trend of transmission and change in response to a public health intervention. We estimate the intervention effect on reducing the infection rate using a natural experiment design and quantify uncertainty by permutation. In China and South Korea, we predicted the entire disease epidemic using only early phase data (two to three weeks after the outbreak). A fast rate of decline in reproduction number was observed and adopting mitigation strategies early in the epidemic was effective in reducing the infection rate in these two countries. The nationwide lockdown in Italy did not accelerate the speed at which the infection rate decreases. In the United States, the reproduction number significantly decreased during a 2-week period after the declaration of national emergency, but declines at a much slower rate afterwards. If the trend continues after May 1, COVID-19 may be controlled by late July. However, a loss of temporal effect (e.g., due to relaxing mitigation measures after May 1) could lead to a long delay in controlling the epidemic.

The human body is exposed to a multitude of prevalent viruses, requiring ongoing surveillance and protection by the immune system. Maintenance of human anti-viral adaptive immunity in diverse tissue sites is determined by a multitude of factors and critical for long-term protection against repeat exposure to viral infection. Yet, studies of anti-viral immunity have primarily been limited to animal studies and studies of peripheral blood in humans. Studies in mice have demonstrated that memory T cells in tissues provide superior protection against viral infection compared to circulating T cells, particularly tissue-resident memory T cells (TRM), which remain in tissues long-term without re-entering circulation. However, much remains to be understood about how anti-viral immune responses are maintained in human tissues and how adaptive immune cells in various tissues sites function upon re-exposure to viral antigens. We have established a human tissue resource through a collaboration with LiveOnNY, a local organ procurement organization, to obtain blood and multiple lymphoid and mucosal sites from donors of all ages. Using this tissue resource, we employed comprehensive cellular and molecular analysis to investigate tissue immunity to three prevalent but distinct viruses—influenza A, CMV, and SARS-CoV-2. We compared CD8+ T cells recognizing ubiquitous and longstanding viruses influenza A and CMV across multiple tissue sites of 58 organ donors ages 1-78 years in order to elucidate how covariates of virus, tissue, age, and sex impact the anti-viral immune response. Using flow cytometry, T cell receptor repertoire sequencing, functional assays, and single-cell transcriptional profiling, we showed that virus specificity and tissue localization are the primary drivers of anti-viral T cell immune responses in the human body, with age and sex further influencing T cell subset differentiation. Specifically, virus specificity correlated with virus-specific T cell distribution, memory subset differentiation, and clonal repertoire, while tissue localization determined overall subset distribution and functional responses. We further investigated the tissue-localized immune response to emergent SARS-CoV-2. By examining multiple tissues of organ donors who had recovered from natural infection by SARS-CoV-2, we showed that adaptive memory immune responses persisted months after infection, with memory T and B cells preferentially localized in the lung and lung-associated lymph node. Persisting memory cell populations included tissue-resident T and B cells, particularly in the lung, as well as germinal center B cells in the lung-associated lymph node along with follicular helper T cells, indicating ongoing generation of humoral immunity. Together, these findings highlight the importance of tissue-localized anti-viral immunity and help to define characteristics of site-specific protective immunity that may be leveraged for the development of more effective treatment and prevention strategies.

Latinx communities in the United States made up 18% of the total population, yet accounted for 33% of COVID-19 morbidity and mortality. This supported the study aim to increase Latinx COVID-19 knowledge and self-efficacy for performing COVID-19 risk reduction mitigation behaviors via dissemination of the new online e-health intervention of the “Our COVID-19 Knowledge Test.” The study recruited online a largely female Latinx adult sample (N=118) with 68.6% born in the U.S. that was well-educated, given a mean education level of a bachelor’s degree; and, a mean annual household income of $50,000 to $99,000. During the pandemic year of 2020, 46.5% of the survey participants experienced moderate to maximum/extreme cultural stress, and moderately high COVID-19 related stress—while 66.9% reported depression, 78.8% anxiety, and 45.2% trauma. Their high rates of COVID-19 depression and anxiety were more than double those rates reported across samples identified globally during the pandemic. They experienced significant declines in their self-rated mental health status and physical health status from pre-pandemic to during the pandemic, high social support, and closest to a good quality of life. Supporting the value of the new “Our COVID-19 Knowledge Test” as a brief online e-health intervention, paired t-tests showed statistically significant increases in self-ratings for both COVID-19 knowledge and self-efficacy for COVID-19 risk reduction behaviors after taking the True-False test. Participants endorsed the dissemination of the new True-False “Our COVID-19 Knowledge Test” with all True answers as a brief online e-health intervention they would recommend to others as a way to learn about COVID-19. Meanwhile, on this True-False test, the sample evidenced very high knowledge of COVID-19. The sample also had a high intention to vaccinate or already vaccinated at 87%. Findings from independent t-tests, Pearson correlations, and regression analyses collectively affirmed the critical importance of having both high knowledge and high self-efficacy for performing preventive behaviors for reducing the risk of COVID-19 transmission, implementing mitigation strategies, and reducing mortality. Implications and recommendations focused on the value of the genre of a True-False test, with all True answers, for disseminating evidence-based information, and countering misinformation during pandemics and public health crises. Finally, the short tools used in this study were recommended for application in future research and as screening tools.

Simulation is a powerful technique to study complex problems and systems. This thesis explores two different problems. Part 1 (Chapters 2 and 3) focuses on the theory and practice of the problem of simulating graphs with a prescribed degree sequence. Part 2 (Chapter 4) focuses on how simulation can be useful to assess policy changes in emergency medical services (EMS) systems. In particular, and partially motivated by the COVID-19 pandemic, we build a simulation model based on New York City’s EMS system and use it to assess a change in its hospital transport policy. In Chapter 2, we study the problem of sampling uniformly from discrete or continuous product sets subject to linear constraints. This family of problems includes sampling weighted bipartite, directed, and undirected graphs with given degree sequences. We analyze two candidate distributions for sampling from the target set. The first one maximizes entropy subject to satisfying the constraints in expectation. The second one is the distribution from an exponential family that maximizes the minimum probability over the target set. Our main result gives a condition under which the maximum entropy and the max-min distributions coincide. For the discrete case, we also develop a sequential procedure that updates the maximum entropy distribution after some components have been sampled. This procedure sacrifices the uniformity of the samples in exchange for always sampling a valid point in the target set. We show that all points in the target set are sampled with positive probability, and we find a lower bound for that probability. To address the loss of uniformity, we use importance sampling weights. The quality of these weights is affected by the order in which the components are simulated. We propose an adaptive rule for this order to reduce the skewness of the weights of the sequential algorithm. We also present a monotonicity property of the max-min probability. In Chapter 3, we leverage the general results obtained in the previous chapter and apply them to the particular case of simulating bipartite or directed graphs with given degree sequences. This problem is also equivalent to the one of sampling 0–1 matrices with fixed row and column sums. In particular, the structure of the graph problem allows for a simple iterative algorithm to find the maximum entropy distribution. The sequential algorithm described previously also simplifies in this setting, and we use it in an example of an inter-bank network. In additional numerical examples, we confirm that the adaptive rule, proposed in the previous chapter, does improve the importance sampling weights of the sequential algorithm. Finally, in Chapter 4, we build and test an emergency medical services (EMS) simulation model, tailored for New York City’s EMS system. In most EMS systems, patients are transported by ambulance to the closest most appropriate hospital. However, in extreme cases, such as the COVID-19 pandemic, this policy may lead to hospital overloading, which can have detrimental effects on patients. To address this concern, we propose an optimization-based, data-driven hospital load balancing approach. The approach finds a trade-off between short transport times for patients that are not high acuity while avoiding hospital overloading. To test the new rule, we run the simulation model and use historical EMS incident data from the worst weeks of the pandemic as a model input. Our simulation indicates that 911 patient load balancing is beneficial to hospital occupancy rates and is a reasonable rule for non-critical 911 patient transports. The load balancing rule has been recently implemented in New York City’s EMS system. This work is part of a broader collaboration between Columbia University and New York City’s Fire Department.

Preterm birth is a key indicator of maternal and child health, affecting 1 in 10 deliveries in the United States (US) and contributing to long-term morbidity and healthcare costs. The COVID-19 pandemic and policies to mitigate the spread of infection may have indirectly impacted preterm birth, but the results of early epidemiological studies were mixed and declines were largely concentrated in high-income countries and populations. Moreover, while most studies focused on stress-related pathways associated with lockdown policies, healthcare disruptions may have also played a role. The goal of this dissertation was to investigate changes in preterm birth and healthcare disruptions related to the COVID-19 lockdown in a low-income population in the US. In the first aim, I conducted a systematic review of the literature on the pandemic and preterm birth, with a focus on studies that examine heterogeneity by income. In the second aim, New York State (NYS) Medicaid claims were used to examine changes in preterm birth rates during the state’s lockdown policy (NYS on PAUSE) using difference-in-difference methods. In the third aim, changes in preterm were further stratified into those that were spontaneous or medically induced, which may reflect a healthcare pathway. Weekly rates of healthcare utilization, antenatal surveillance, and maternal complications were also assessed using interrupted time series models to characterize healthcare disruptions over the course of the lockdown and across the state. Results from the systematic review documented the rapid growth in research on this topic since the beginning of pandemic. Among the 67 articles included, most reported some decline in preterm birth rates; however, there was large variation by country, methods of exposure assessment, and onset of delivery. Only seven studies focused on differences by individual income (or income proxies) and those that did were inconsistent. Results from Aim 2 suggested that NYS on PAUSE was associated with nearly a percentage point decline in preterm birth rates in the Medicaid-insured population, without a concomitant increase in stillbirth. Aim 3 demonstrated that the change in preterm was largely driven by declines in medically induced preterm. Interrupted time series models showed substantial, but time-limited, declines in pregnancy-related healthcare utilization at the beginning of NYS on PAUSE. Overall, the findings in this dissertation suggest there were modest declines in preterm birth during the COVID-19 lockdown among low-income women in NYS, particularly in medically induced preterm. Healthcare disruptions were common for Medicaid-insured women and may partially explain the reduction in preterm birth in this population. Future research is needed to determine whether this change was positive for some and negative for others, and what that might mean for efforts to improve pregnancy outcomes in the future.

My dissertation examines the experiences of Teachers College faculty that transitioned to online teaching for the first time during the forced circumstances of COVID-19. More specifically, I explore: 1) the relationship between feeling prepared, supported, and connected with professional development; 2) the experiences of faculty making the transition to online teaching; and 3) how faculty described re-evaluating, as Boud describes it, their teaching experiences. To better understand this, I used the case-selection variant of the explanatory sequential, mixed-methods design (quan → QUAL). I surveyed 85 participants (Phase 1) that engaged in professional development opportunities provided by the institution to better understand their experiences preparing for this transition and then interviewed 10-participants (Phase 2) to better understand their unique experiences. I found that most participants that made this transition grew both in their ability to use technology and comfort with teaching online. Participants described the experience as a challenging transition, especially as there was little time to prepare; however, participants also learned (through consultations, intensive programs, colleagues, and students) from this experience. In Phase 2, 7 of 10 participants indicated that they will take their learnings from teaching online and integrate them into their face-to-face teaching.

Worldwide, there were 463 million adults (20-79 years) with diabetes in 2019. These figures are expected to increase to 700 million by 2045. Additionally, approximately 4.2 million deaths worldwide were attributable to diabetes, and global health care expenditures on individuals with diabetes were estimated to be 760 billion U.S. dollars. One of the most effective ways to control this debilitating disease is to prevent it before it happens, which, based on evidence from the Diabetes Prevention Program (DPP) randomized controlled trial and other studies in different countries, is feasible with a change in weight, dietary habits, and physical activity levels. While many studies have shown the benefits of plant-based diets in diabetes prevention, no DPP studies have been found that have incorporated a plant-based diet for their dietary component. The purpose of this pilot study was to explore the feasibility, acceptability, and preliminary efficacy of implementing a worksite plant-based diabetes prevention program to inform larger randomized trials to be conducted in the future. This was a mixed-methods pilot study using a one group pretest-posttest design. The study was delivered to Teachers College, Columbia University employees and staff and was designed to use a modified version of the CDC’s national curriculum, one that emphasizes plant-based eating patterns. The sessions ran during lunch hour where a healthy lunch that supported the behavior change goal of the session was provided. The Health Action Process Approach (HAPA) framework was used for both curriculum design and program evaluation. The principles of facilitated group discussion in a safe environment were used to deliver the sessions. The program was conducted in the spring semester of 2020, once a week, for a series of 13 weeks plus two voluntary booster sessions held 1 and 2 months after the program was completed. Midway through the semester, the program went virtual using synchronous video-conferencing technology due to the COVID-19 pandemic. Data on consumption of healthy and unhealthy plant- and animal-based foods, physical activity, and psychosocial variables were measured pre and post program using validated questionnaires; blood glucose values were measured as HbA1c using A1CNow® SELFCHECK; and weights were measured weekly using Tanita SC-331S scales when classes were in-person and home scales when classes went virtual. Evaluations of participants’ acceptability and satisfaction were assessed at the end of the program both quantitatively and through interviews. Finally, fidelity to the plant-based curriculum and evaluation of educational plan completion and engagement were done weekly. Forty-one individuals expressed interest in the study, but only 18 met the eligibility criteria, of whom 14 were finally enrolled, constituting 78%. The participants were ethnically/racially diverse. Attrition was very low with only one dropout, and this did not change when the program went virtual. Program delivery was in fact feasible and all 13 lessons and booster sessions were completed. The plant-based DPP was received with a high degree of acceptability and satisfaction by the participants. Participants described the safe environment created and the facilitated dialogue approach in the sessions, along with peer support as instrumental for their behavior changes. There were some significant improvements in the physiological, behavioral, and psychosocial outcome measures explored in the study which included: weight, diet quality in terms of plant-based and animal-based foods, physical activity levels, blood glucose levels, and behavioral and psychosocial determinants of behavior change of the Health Action Process Approach (HAPA) theoretical framework. After study completion and analysis of results, it is clear that conducting a worksite plant-based diabetes prevention program is in fact feasible and acceptable, and may be efficacious at eliciting positive changes in physiological, behavioral, and psychosocial variables that can potentially attenuate risk of developing diabetes. The findings will be useful for designing larger controlled studies.

This study seeks to expand notions of research, what it can be and how it can be conducted, through focusing on children’s approaches to exploring their worlds. The purpose of this study was to examine how children employ literacies of research across spaces. Through this framework, I conceptualize children’s literacies of research to include the social practices children engage in when investigating issues that matter to them. Previous participatory studies with young people have focused on apprenticing youth and children into traditional research practices in order to then conduct studies with them that are relevant to their lives. This study builds on this work but begins by exploring the notion of research itself, seeking to understand children’s perspectives on how they examine topics of interest. Framed by critical and transformative theoretical frameworks, specifically critical childhoods, sociocultural approaches to literacy, and youth participatory action research (YPAR), this study engaged a small group of nine- and ten-year-old children, representing a range of racial, cultural, and linguistic backgrounds, as co-researchers. The following research questions shaped the study: How do nine- and ten-year-old children in a participatory research group engage with opportunities to follow their own lines of inquiry?; What themes do they investigate and how?; What literacy and research practices do they draw on, resist, remix, and/or transform and how?; and How do adults interact with children around child-led research? The findings suggest the playful, relational, dynamic, intertextual, and resistant natures of children’s literacies of research. This study was interrupted by the first wave of the COVID-19 pandemic and the research group transitioned to a virtual space. The findings also indicate the innovative ways children resisted the isolating circumstances of the COVID-19 pandemic through creating and repurposing digital platforms to sustain friendships and connect with classmates. Children’s literacies of research have implications for how research is conceptualized and taught in literacy classrooms and in the academy as well as how researchers engage with children in studies.

The online convenience sample of 249 nurses all treated COVID-19 patients in the past year—with 45.0% in the emergency department and 36.9% in intensive care. Nurses were 68.7% female with a mean age of 32.17 years, as well as mostly white (69.1%). Some 28.5% had COVID-19, with 16.1% testing positive more than once in the past year. Using paired t-tests comparing scores for before versus during the pandemic, their physical health status and mental/emotional status were each significantly worse during the pandemic, their level of self-efficacy for performing nursing tasks was significantly worse during the pandemic, and their fear level was significantly higher during the pandemic. Nurses negotiated the pandemic with just moderate social support, while having moderate work setting concerns (e.g., safety), and rating the work climate as “to some extent” less favorable than before the pandemic. Nurses suffered moderate burnout using the Oldenburg Burnout Inventory—while females suffered higher burnout than males (p = .000) and non-whites higher burnout than whites. Past month mean Perceived Stress Scale scores were moderate. Nurses used alcohol/drugs closest to 30% of the time to cope with stress, while 35.7% increased use during the pandemic. They reported moderate mental distress over the past year, while 61.0% reported insomnia, 57.4% anxiety, 39.0% depression, 35.7% trauma, and 27.3% received counseling. Nurses reported moderate well-being over the past two weeks, and moderately high satisfaction with life. Backward stepwise regression found higher burnout significantly predicted by: fewer years working in nursing; higher Body Mass Index; more concerns at work (e.g., safety); higher past month perceived stress; higher past year mental distress; and, lower past two weeks’ well-being—with 52.2% of the variance predicted. Qualitative data reinforce important recommendations.