22689 research outputs found
Sort by
Trends and Disparities in Self-Reported Health Among US Adults with Diabetes: NHANES 2001-2018
No full textSelf-reported health status is an indication of overall well-being beyond clinical indicators alone. It serves as a validated predictor of morbidity and mortality. Prior studies among individuals with diabetes in the United States have largely focused on diabetes treatment and control and rates of complications, while data on patient-reported outcomes, such as self-reported health, are lacking. The objective of this study was to characterize trends and disparities in self-reported health among American adults with diabetes using nationally representative data from 2001-2018. Additionally, an exploration of factors associated with high self-reported health was employed.
We performed serial cross-sectional analyses of adults aged ≥20 years with self-reported diabetes who participated in the National Health and Nutrition Examination Survey (NHANES) between 2001 to 2018. Self-reported health status was categorized as high (good/very good/excellent) or low (poor/fair). All analyses accounted for the complex NHANES survey design and survey weights were applied to obtain representative estimates.
Among 5,306 adults with diabetes (mean age, 60 years; 49% female; 63% Non-Hispanic White), the prevalence of high self-reported health remained stable from 2001-2004 (58.3%, 95% CI: 53.1-63.3%) to 2017-2018 (64.0%, 95% CI: 59.3-68.5%) (p-trend=0.096). Significant improvement was observed in adults aged ≥65 years (55.8% to 68.9%, p-trend=0.002).
We found that after more than a decade of progress in diabetes management and care, self-reported health has improved only among older adults, while persistent and widening disparities affect racial/ethnic minorities and socioeconomically disadvantaged populations
LEARNING FROM EXPERIENCE TO IMPROVE CLIMATE ADAPTATION UNDER THE UNFCCC: A CASE STUDY OF THE MEKONG DELTA
Get PDFNation states that are parties to the United Nations Framework Convention on Climate Change (UNFCCC) have established organs to finance climate adaptation measures, the largest of which is the Green Climate Fund (GCF). The UNFCCC also tasked lesser developed nations seeking to receive such funding to prepare and periodically update National Adaptation Plans (NAPs). The GCF has established procedures and criteria for post-project evaluations to promote “continuous learning” from experience to inform improvements in the design and execution of NAPs and adaptation projects. These criteria call for an examination of the extent to which the results achieved can be upscaled and replicated in other locations facing similar climate risks. This thesis examines the effectiveness of these evaluations for this intended purpose through a case study of a GCF project in the Mekong Delta in Vietnam.
This case was chosen because it is emblematic of highly productive and vulnerable deltaic systems in the global south and because the adaptation measures are appropriate to climate threats experienced in those settings. These measures are the regeneration of mangroves to buffer the landscape from extreme weather events and consequent floods and salinization; and the construction of resilient housing to withstand the same risks.
The research methods consisted of qualitative review of relevant documentation produced by the project and by various experts on climate adaptation in the Mekong Delta and by a series of structured interviews with the funder (GCF), the implementing agencies (the United Nations Development Programme) and counterpart agencies of the Vietnam Government at both the national and local levels, project beneficiaries in the local communities, and non-governmental organizations. The objectives and outcomes asserted in the funding proposal to justify the grant were compared to the results achieved. These interviewees revealed questionable efficacy of the mangrove regeneration sites and techniques and disqualification of the poorest residents from the housing benefit.
The author then examined the post-project evaluation submitted by UNDP and found it deficient in assessing how these discrepancies affect the potential for upscaling and replicating. Seven reforms in the evaluation process are proposed and explicated that incorporate but are more exacting and comprehensive than reforms recently instituted by the GCF itself to improve monitoring and evaluation by providing advisory services to project implementers. That program provided additional data that indicates that the flawed feedback process detected in the case study is representative of a widespread failure to derive lessons from the limitations or “unintended consequences” encountered in executing GCF-funded adaptation projects
Social Signals for Interactive, Error Aware Robotic Systems
Get PDFRobot errors during human-robot interaction are inescapable; they can occur during any task and do not necessarily fit human expectations. Left unmanaged, they harm task performance and user trust, resulting in user unwillingness to work with a robot. Previous error detection techniques have used task or error specific information for robust management and can lack the versatility to address robot errors across tasks and error types.
In this dissertation, I leverage natural human responses to robot errors in physical HRI for error detection across task, scenario, and error type for flexible robot error detection. My approach is two-fold: (1) understand and model social signals (facial action units, or AUs) in response to unexpected robot errors and (2) use these models for reliable, flexible automatic error detection. First, I explore how users respond to unexpected robot errors and if social signals can be effectively used to detect errors across individuals in Programming by Demonstration scenarios. I then expand my research to collaborative human-robot tasks by generating and analyzing a dataset derived from three distinct HRI studies to understand the prevalence of social signals across different errors, tasks, and scenarios. Next, I explore the impact of context on social signals for detection and investigate how people react to robot errors in-the-wild. From this set of four studies, I demonstrate that natural AUs can be used to detect and temporally localize errors with reasonable accuracy and timeliness across different tasks, error types, and people. Recognizing the limitations of relying solely on AUs, I move towards a multimodal approach.
I propose a conceptual framework for error awareness that takes a multimodal approach using social signals to provide error detection flexibility. Finally, I investigate the benefits of proactive error detection using social signals based upon my error-aware framework. Overall, results from my work show that in physical HRI, social signals exhibited in response to robot errors are good indicators and enable flexible automatic error detection. This dissertation contributes to our knowledge of how people implicitly react to robot errors and how behavioral signals can used for flexible robot error detection across error types and scenarios
RMC-7977: A PROMISING INHIBITOR FOR THE TREATMENT OF NF1-RELATED TUMORS
Get PDFNeurofibromatosis type 1 (NF1) is a common genetic disorder characterized by mutations in the NF1 gene, leading to the loss of neurofibromin, a RAS GTPase-activating protein, which results in the persistent activation of the RAS-MAPK signaling pathway, facilitating tumorigenesis, such as cutaneous and plexiform neurofibromas, malignant peripheral nerve sheath tumors (MPNSTs), and gliomas. Despite advances in our understanding of NF1-associated tumor biology, current treatment options remain limited, especially for MPNSTs and high-grade gliomas. RMC-7977, a novel tri-complex pan-RAS inhibitor, selectively targets the active GTP-bound form of RAS by forming a stable complex with cyclophilin A (CYPA) and RAS, offering a new therapeutic approach.
Here, we evaluated the efficacy of RMC-7977 in NF1 -mutated tumor models. IC50 assays were performed on multiple NF1 cell lines, identifying sensitive lines for in vivo investigation. Xenograft mouse models were then established via orthotopic implantation of luciferase-tagged cells into the sciatic nerve, brain, or subcutaneously of NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ (NSG) mice. We found that RMC-7977 significantly suppressed tumor progression in ST8814, S462, and JH-2-002 models, and improved survival in intracranial LN229 tumor-bearing mice, with delayed but statistically significant tumor growth inhibition and prolonged survival. Interestingly, despite its high molecular weight of 865.1 Da, RMC-7977 demonstrated the ability to penetrate the Blood-Brain Tumor Barrier (BBTB) and exert antitumor activity within the central nervous system. Overall, these findings provide foundational evidence for RMC-7977 as a promising therapeutic strategy for targeting RAS-driven NF1 malignancies
Characterizing Ambient, Personal, and Occupational Air Pollution and Lung Function Associations in South-Central Uganda
No full textBackground: Exposure to fine particulate matter (PM2.5) significantly impacts respiratory health, particularly in low- and middle-income countries, where ambient air pollution often exceed international health guidelines. In Sub-Saharan Africa, including Uganda, comprehensive ambient air quality data are limited due to financial, logistical, and infrastructure constraints. The lack of data hinders the ability to characterize exposure in different environments and to inform targeted public health interventions.
Methods: A total of 27 low-cost PM2.5 sensors were calibrated against reference-grade monitors (BAM-1022 in Kampala and E-Sampler in Rakai) and deployed across urban (Kampala, Masaka) and rural (Rakai Region) areas of South-Central Uganda. Ground sensor data were combined with satellite-derived aerosol optical depth, meteorological data, and land-use variables to develop a high-resolution (1 km²) ambient PM2.5 model. A total of 1,341 participants completed spirometry to assess lung function. A subsample of 103 participants completed time-activity diaries and GPS tracking to evaluate personal, microenvironmental, and occupational PM2.5 exposures.
Summary: Ambient PM2.5 concentrations showed strong seasonal and diurnal patterns, with higher urban concentrations compared to rural. Reconstructed personal PM2.5 closely aligned with measured exposures among rural participants but overestimated measured exposures among urban participants. Increases in ambient PM2.5 concentrations at short-term exposure windows (Lag Days 1 and 2) were associated with significant lower FEV₁ and FVC z-scores among vulnerable subgroups, including females, participants under 25 years, and those with prior COVID-19 infection. Microenvironmental apportionment showed that rural participants experienced the highest PM2.5 concentrations in the Home and Other microenvironments, while urban participants had the highest exposures in Work and Transit.
Conclusion: This research demonstrates the value of combining validated low-cost sensor data, satellite-derived information, and environmental covariates to characterize ambient PM2.5 concentrations in regions with limited ambient air quality monitoring. Differences between reconstructed and actual personal exposures highlight the challenge of capturing localized sources. Moreover, associations between short-term PM2.5 concentrations and lower lung function highlight the potential for immediate respiratory effects following acute exposure. Identifying specific microenvironments and occupational activities contributing to elevated exposures provides actionable insights, underscoring the need for targeted interventions addressing both ambient and localized PM2.5 sources to effectively protect respiratory health
EXPLORING THE COGNITIVE IMPACT OF POLYPHARMACY: A CLUSTERING APPROACH TO MEDICATION PATTERNS
Get PDFPolypharmacy, the concurrent use of multiple medications, is prevalent among older adults and individuals with chronic conditions, including those with HIV. While necessary for managing comorbidities, polypharmacy has been associated with cognitive decline, yet the specific interactions between medication usage patterns and cognitive performance remain unclear. In this study, we investigate whether participants from an HIV Neurobehavioral Research Program dataset—many of whom have diverse comorbidities—naturally cluster into distinct communities based on their medication profiles and whether these communities exhibit differences in cognitive performance.
We construct a binary matrix of participants by medications and focus on individuals taking five or more medications (clinical polypharmacy threshold). Using this matrix, we generate similarity measures and apply multiple clustering techniques, including spectral clustering and Louvain community detection, to identify medication-based communities. We then assess whether these communities differ in cognitive performance across tasks such as speed of information processing, executive function, and working memory. Finally, we examine whether certain medication classes are overrepresented within specific communities, potentially revealing hidden drug interactions or protective effects.
A critical aspect of this analysis is determining the most appropriate similarity metric and clustering method for this data. We explore the stability of clusters across different methods and assess quality using measures such as the eigengap heuristic, modularity scores, and community consistency metrics. Additionally, we consider potential confounding effects of comorbidities and propose regression-based approaches to account for these factors.
This work may have important implications for understanding how medication patterns relate to cognitive function. By identifying medication clusters associated with cognitive impairment or resilience, we hope to inform future research on polypharmacy management and medication safety, ultimately guiding clinical decision-making to optimize cognitive health in vulnerable populations
Development of Biomaterial-Based Delivery Vehicles for Vaccine Applications
No full textVaccination remains one of the most effective public health strategies, preventing millions of deaths annually and offering a critical defense against infectious diseases. While traditional vaccines were developed empirically, modern approaches aim to rationally design platforms that elicit durable and robust immune responses. Despite the success of subunit vaccines in safety and manufacturing, they often lack the potency and long-term immunity provided by live attenuated vaccines. Advances in immunology, bioinformatics, and biomaterials offer new opportunities to address these limitations. Subunit vaccines rely on the presentation of antigens alongside adjuvants to activate both humoral and cellular immune pathways, with dendritic cells playing a central role in priming T and B cells. Adjuvants such as alum and MF59 have shown immunogenic potential, but their precise mechanisms remain under investigation. Emerging adjuvants targeting Toll-like receptors offer the ability to fine-tune immune responses, especially toward T-cell-mediated immunity.
Biomaterial-based strategies, particularly using nanoparticles and microgel systems, enhance vaccine efficacy by targeting lymphoid tissues, sustaining antigen release, and creating localized inflammatory environments. Polymeric systems like Poly-lactic-co-glycolic acid (PLGA) microparticles are well established and approved by the FDA. However, challenges such as degradation mediated protein denaturation, and variable encapsulation efficiency must be addressed. In this study, we propose the use of PEG-based microgels crosslinked via bioorthogonal click chemistry as a biocompatible depot formulation for sustained vaccine delivery. Additionally, we explore polycatechin nanoparticles, derived from green tea flavonoids, for their immunomodulatory potential
and aim to investigate their in vivo biodistribution by labeling them with far-red fluorescence dyes for detection via Flow cytometry analysis of cells isolated from lymph nodes and spleen cells. Together, these approaches aim to contribute towards enhancing the quality, durability, and precision of vaccine-induced immunity
Breast Cancer Diagnosis Via Swarm-Optimized Adaptive Neuro-Fuzzy Inference System
Get PDFThis thesis discusses a method for classification of breast cancer imaging data through the application of an adaptive neuro-fuzzy inference system (ANFIS) and particle swarm optimization (PSO) for hyperparameter optimization of the ANFIS system. A robust parameter tuning method is used to select the most optimal configuration for the ANFIS and PSO components. Using these methods, high classification accuracies can be achieved for both the original and diagnostic versions of the Wisconsin Breast Cancer Dataset
The Immunomodulatory Potential of Cannabinoids Across Species
No full textCannabis, also known as marijuana, is a widely recognized plant with many uses, from hemp fiber used to make cloth and other home goods to the combination of flowers, leaves, seeds, and their extracts used for recreational purposes or their perceived medicinal benefits. Though cannabis is composed of hundreds of chemical constituents, the most well-known are delta-9-tetrahydrocannabinol (delta-9-THC),and cannabidiol (CBD). While delta-9-THC is psychoactive and responsible for the “high” feeling after ingesting cannabis, CBD is nonintoxicating and known for its putative anti-inflammatory properties. Despite the ubiquity and widespread use of medical and recreational marijuana, as well as delta-9-THC- and CBD-infused commercial products, much remains unknown about cannabis, its constituents, and their interactions with and effects on the body. Therefore, the studies comprising this thesis have sought to increase our knowledge with regard to the effects of delta-9-THC and CBD on endogenous processes and in the context of HIV associate inflammation. The data contained herein has elucidated that cannabinoid compounds exhibit immunomodulatory potential in several ways, including impacting homeostatic immune processes, modulating immune dysfunction and uncontrolled inflammation in the context of HIV/SIV, and influencing endocannabinoid receptor expression. The importance of species selection for model systems in cannabinoid research is highlighted in the following chapters as well, particularly as it applies to canonical and extended endocannabinoid receptors and translation of research from bench to bedside. Furthermore, the emerging significance of extended endocannabinoid receptors and propensity of cannabinoids to modulate these receptors is becoming increasingly clear and necessitates expansion of future studies to include them. Altogether, this thesis demonstrates that cannabinoids have profound effects across species, including influence on homeostasis and modulation of HIV and HIV associated inflammation
Algorithmic Decision-Making by Health Insurers: Implications for Patients and Providers
Get PDFThis dissertation examines what happens when insurers use algorithms to decide what care to cover and what care to deny. I focus on the post-acute care context, where patients receive rehabilitation and skilled nursing services after a hospitalization. I study Medicare Advantage, the private insurance market in Medicare that now covers the majority of beneficiaries. I examine (1) how the use of algorithms affects patients and (2) how the use of algorithms and other managed care techniques affects post-acute care providers.
In paper 1, I examine the causal effect of algorithm-aided insurer decision-making on health care use and patient outcomes. Using a difference-in-differences design and administrative data, I study the partnership of a large Medicare Advantage insurer with a firm that uses an algorithm to aid post-acute care coverage decisions. I find that the introduction of algorithm-aided decision-making led to an immediate and substantial decline in the number of days patients spend at skilled nursing facilities. However, I find that there were no observable changes in health outcomes.
In paper 2, I use similar methods and data to examine the causal effect of algorithmic decision-making when implemented by two smaller Medicare Advantage insurers. I find similar effects as in paper 1, suggesting that the results are more broadly generalizable beyond the experience of a single insurer.
In paper 3, I examine whether the rise of Medicare Advantage enrollment has affected the finances of skilled nursing facilities, through its use of utilization management strategies (including the use of algorithm-aided prior authorization) and its negotiation of prices. I exploit the varied growth of Medicare Advantage across counties between 2012 and 2019 to examine this question. I find that rising local Medicare Advantage enrollment was associated with lower profit margins at skilled nursing facilities