Department of Epidemiology and Biostatistics Abstracts

Presenter(s): Arielle Coq

Showcase Advisor: Muntasir Masum

Abstract: Background: Alzhemier’s Disease and Dementia rates are increasing in New York State. Understanding the risk factors among NYS residents will allow NYDOH to tailor programs that will have the greatest impact for our residents.

Methods: The team will be analyzing the Behavioral Risk Factor Surveillance Survey (BRFSS) to explore the leading factors of Alzheimer's Disease among New York State residents. Rstudio will be used to conduct logistic regression of memory loss/confusion and eleven of fourteen identified risk factors of dementia.

Results: With a final sample size of 23,795 observations, 2,455 (10%) reported experiencing confusion/memory loss in the last 12 months. Respondents who had one of the risk factors identified were also more likely to experience memory loss or confusion in the last 12 months.

Conclusion: Based on the preliminary results, there may be an association between memory loss and the top risk factors of dementia among New York State Residents.

Presenter(s): Jack Farrell

Showcase Advisor: Edward Valachovic

Abstract: Alzheimer’s disease mortality rates are rising, necessitating reliable projections to support effective healthcare coordination. Analytical approaches capable of decomposing time series data into interpretable components to produce reliable forecasts are of paramount importance. Using monthly mortality data (1999-2023), the Kolmogorov-Zurbenko filter was applied to decompose the time series into trend, seasonal, and noise components. An ensemble forecasting approach blending recent and long-term trends yielded five-year projections with 80% and 95% confidence intervals, validated using a 90/10 training/testing split (MAPE = 8.7%). Results predict further increases over the forecast horizon—from 11,258 deaths per month in 2024 to 12,818 in 2028, a 13.9% five-year growth, with a total of 721,521 projected deaths. Log-transformed analysis computed trend (35.6%), seasonal (10.2%), and noise (5.1%) contributions to overall variability, consistent with mortality spikes each winter. This approach delivers interpretable forecasts to inform better public health policy and resource management for Alzheimer’s disease.

Presenter(s): Shreya Bohara Chand

Showcase Advisor: Jessica Kumar

Abstract: Introduction: The National Immunization Program of Nepal aimed to introduce a booster dose of the existing pentavalent (DPT-HepB-Hib)  vaccine and analyze the Full Immunization Declaration (FID) program to understand its association with the high immunization rate (95%) reported in the last fiscal year (2023-2024).

Method: The team proposed to conduct a literature review to be included in the grant proposal submitted to GAVI (Global Alliance for Vaccines and Immunization) and prepare a manuscript for the FID program. Manuscript was drafted through the analysis and evaluation of the FID guideline, published articles, annual health reports and interviewing focal immunization people.

Conclusion: The FID program is the most impactful immunization initiative in Nepal which has helped achieve notable key milestones. The literature review supports the introduction of a booster dose of the pentavalent vaccine in Nepal. Based on WHO's position paper recommendation, the dose should be administered at 12-23 months of age.

Presenter(s): Julia Francis

Showcase Advisor: Thomas O'Grady

Abstract: Financial resources shape access to healthcare, and lifestyle conditions can influence stroke risk among women. Using 1999-2018 NHANES data, we examined the association between income-to-poverty ratio (PIR) and self-reported history of stroke among female adults. Logistic regression models adjusted for age, race/ethnicity, and education   estimated odds ratios with age stratification. Among 28,608 females, 3.1% reported stroke, and 14.4% lived below the poverty line. Females with a PIR < 3.57 exhibited higher odds of stroke compared to females in the highest income category: Q1 (OR 2.80; 95% CI 2.21-3.54), Q2 (2.50; 1.97-3.18) and Q3 (1.70; 1.34-2.15). The PIR-stroke association differed significantly by age (P=0.0037), suggesting socioeconomic and life-course factors jointly shape cerebrovascular risk. These findings underscore the importance of addressing economic barriers as part of strategies to reduce the incidence of stroke burden among adult women.

Presenter(s): Marina Frimpong 

Showcase Advisor: Muntasir Masum

Abstract: Background:  The US faces persistent disparities in pediatric mental and behavioral health. Prior studies suggest children in language-minority households may fare better despite socioeconomic disadvantage, a pattern called the immigrant paradox. Few studies test whether this association persists after accounting for childhood adversity (ACE).

Methods: We analyzed pooled 2020–2022 National Survey of Children’s Health data for children ages 2–17 years. The outcome was any impairing mental or behavioral condition. Household language was classified as English-only, bilingual, or non-English. Survey-weighted logistic regression estimated adjusted odds ratios controlling for demographics, socioeconomic factors, insurance, nativity, urbanicity, and ACE burden.

Results: One in five children had an impairing condition. Bilingual households had lower odds than English-only households (aOR=0.38), while non-English households showed no difference. Children with three or more ACEs had fourfold higher odds of impairment (aOR=4.66).

Conclusions: Bilingual household environments may protect child mental health, while ACE burden remains a driver of risk.

Presenter(s): Oluwabukunola Adebajo

Showcase Advisor: Alvaro Carrascal

Abstract: This internship with the Suicide Prevention Center of New York, housed within the New York State Office of Mental Health, supported suicide prevention efforts through data-driven research. The project examined suicide-related outcomes among youth using 2023 Youth Risk Behavior Survey data from New York State and New York City. Key activities included conducting a literature review, analyzing trends in suicidal ideation and attempts, and assessing risk and protective factors across demographic subgroups. Particular attention was given to disparities by sex, race and ethnicity, and other social determinants of health. Findings were summarized in a written report and presented internally to inform prevention strategies and planning. Early results highlight inequities in suicide-related experiences among youth and emphasize the need for targeted, culturally responsive interventions. This work demonstrates the value of epidemiologic analysis and surveillance data in guiding evidence-based public health policy and strengthening youth mental health initiatives statewide.

Presenter(s): Yineng Chen

Showcase Advisor: Edward Valachovic

Abstract: Seasonal patterns in ischemic heart disease (IHD) hospitalizations are well known, yet accurate quantification remains challenging due to temporal dependence, multiple overlapping periodicities, and noise. Existing periodic block bootstrap methods often produce relatively wide confidence intervals, limiting precision. We apply the Variable Multiple Bandpass Periodic Block Bootstrap (VMBPBB), a novel bootstrap approach for time series with multiple periodically correlated components, to daily IHD hospitalization rates in New York State from 2002 to 2023. VMBPBB first isolates periodic components using bandpass filtering and then resamples each component separately before recombination. This preserves periodic correlation structures while improving inferential precision. We identify statistically significant weekly and annual components in IHD hospitalizations and several annual harmonics. The weekly cycle peaks midweek and troughs on Sundays, while the annual cycle peaks in mid-summer and declines in late December. VMBPBB produces markedly narrower confidence bands, demonstrating its value for detecting complex periodic structure.

Presenter(s): Henry Scholl

Showcase Advisor: Deanna Ryan

Abstract: Introduction: Every few years, local health departments and hospitals conduct a Community Health Assessment (CHA) using opinions from county residents and data through sources at the local, state, and national level. 
Methods: Using a previously developed CHA survey as framework, meetings were held to revise survey questions using past data rates on demographics and disease rates 
Results: Survey questions were revised, focusing on community health, healthcare access, health conditions and behaviors, and children’s health. SurveyMonkey was chosen due to cost. The initial CHA draft was also started, including parts of the Executive Summary, Service Area, Demographics, and overall health. 
Conclusions: When developing surveys on community opinions, many parts of health must be considered, while ensuring it can be understood and properly distributing it throughout the community. Additionally, health statistics must be gathered and displayed in a way to communicate health rates over time and concerns to advocate for at-risk groups.

Presenter(s): Reilly Weinstein

Showcase Advisor: Thomas O'Grady

Abstract: Externalizing behavior (EB) in adolescence (e.g., aggression and rule-breaking) is linked to long-term adverse health and socioeconomic outcomes. Childhood material hardship—lacking basic needs such as food, housing, and medical care—may increase the risk of later EB through chronic stress pathways. Using a sample (n=2,775) from the Future of Families and Child Wellbeing Study, we examined the association between cumulative material hardship across ages 1–9 and high EB at age 15. In adjusted logistic regression models, each additional year of material hardship increased the odds of high EB (aOR=1.37, 95% CI [1.21, 1.47], p < .001). Compared to no material hardship, odds were higher for 3 and 4 years of exposure (aOR=2.17, 95% CI [1.51, 3.13]; aOR=3.14, 95% CI [2.18, 4.54], respectively). Findings demonstrate a dose-response pattern consistent with cumulative risk theory, underscoring the need for early and sustained interventions to mitigate persistent material hardship.

Presenter(s): Zoya Hayes

Showcase Advisor: Muntasir Masum

Abstract: Cervical cancer screening surveillance has constantly demonstrated disparities leading to increased incidence and mortality in ethnic minority groups. The current project explores this relationship between race/ethnicity and cervical cancer screening in 50,000 adult women aged 25 to 64, adjusting for sociodemographic and healthcare access factors through the 2024 Behavioral Risk Factor Surveillance System and logistic regression analysis. Preliminary findings suggest that 90% of the sample have had a cervical cancer screening; however, non-Hispanic Asian or Pacific Islander identifying women have lower rates of screenings compared to other ethnic groups. Analysis of covariates is expected to demonstrate the countless factors that contribute to the variation in cervical cancer screening. This research is critical for public health because cervical cancer and limited access to screening can disproportionately affect specific sub-populations, and begin to uncover ethnic disparities in reproductive health that could enlighten targeted implementation strategies, expanded screening programs, and reduced structural barriers.

Presenter(s): Jabin Tasnim Jahan

Showcase Advisor: Redd Driver

Abstract: This poster presents findings from an internship project with the New York State Department of Health AIDS Institute, evaluating client experiences in a pilot program providing affirming, low-barrier, integrated care for sex workers. An anonymous consumer experience survey assessed clinic environment, interpersonal interactions, substance use, and social determinants of health (SDOH). The final analytic sample included 26 respondents, and descriptive statistics were used to summarize responses. Most participants reported an affirming clinic environment and high levels of identity respect, along with positive ratings for communication, privacy, and shared decision-making. Reports of disrespect or negative service experiences were uncommon. Substance use was common, and for some participants, it interfered with care engagement. Many respondents also reported unmet SDOH needs, including food insecurity, housing instability, financial strain, and loneliness. Findings highlight the importance of affirming care models while identifying opportunities to strengthen trauma-informed, culturally responsive care and expand partnerships to address SDOH.

Presenter(s): Nour El.Huda Ahmed

Showcase Advisor: Amy Robbins

Abstract: The New York State Department of Health Bureau of Communicable Disease Control uses SaTScan to detect potential clusters of enteric diseases. This project evaluated alerts for five enteric diseases to determine the impact of international travel on alert generation and whether alerts corresponded to true epidemiological clusters. 
StaTScan alerts reported between January 1, 2023, and October 31, 2025, were reviewed. Unique alert IDs were created for analysis. Microsoft Excel was used to calculate case counts, the number of travel-associated cases per alert, and to classify type of cluster.

If cases with international travel were excluded from the analysis, 85.6% of alerts would generate a SaTScan signal. Most alerts didn’t correspond to identified clusters, representing 87.8% of campylobacteriosis, 87.5% of cryptosporidiosis, 90% of giardiasis, 69.2% of salmonellosis, and 78.3% of shigellosis alerts.

Most alerts did not represent true outbreaks, highlighting the need to adjust SaTScan parameters to prioritize investigations and resources.

Presenter(s): Lauren Martin

Showcase Advisor: Muntasir Masum

Abstract: Housing instability is an important social determinant of health that may contribute to poor mental health outcomes in the United States. This study examines the association between housing instability and frequent mental distress across different racial/ethnic groups. Data were drawn from the 2024 Behavioral Risk Factor Surveillance System (BRFSS), a nationally representative cross-sectional survey. The analytic sample included 162,928 respondents after applying complete-case analysis and dropping missing data. Logistic regression models will be used to examine the association between housing instability and frequent mental distress among different racial/ethnic groups. Preliminary unadjusted logistic regression results suggest that respondents experiencing housing instability have significantly higher odds of reporting frequent mental distress compared to those with stable housing. These findings highlight the importance of housing stability as a key social determinant of mental health and may help inform targeted public health interventions aimed at reducing mental health disparities.

Presenter(s): Erin Shaut

Showcase Advisor: Ryan Stryker

Abstract: Introduction:  Lead is a naturally occurring metal that, when ingested/inhaled, causes cognitive decline and developmental delays.  Steuben County Public Health develops an annual public-facing map containing clustered locations of reported children with elevated blood-lead levels (>5μg/dL).

Methods:  Multivariate analysis and logistic regression was performed using data collected from a county-wide survey.  The 2024 map incorporates an interactive online map using Esri GIS software.

Results:  Lower annual income and SES contributed to higher concern for community members; statistical analysis showed no significant associations. More populated areas with historical buildings have higher concentrations of children with EBLLs, and regions of lower SES in Steuben County reported more cases.

Conclusions:  Although there were no significant correlations, survey results imply gaps in knowledge about treatment and prevention of childhood lead poisoning.  Geospatial analysis also suggests a link between socioeconomic status and lead exposure which should be explored further in a larger sample.

Presenter(s): Oghenekaro Ekor

Showcase Advisor: Adam Rowe

Abstract: This project investigated respiratory-related hospitalizations in New York State’s Capital Region, analyzing geospatial and demographic trends for COVID-19, influenza, and RSV during the 2023–2024 and 2024–2025 seasons using RESP-NET surveillance data. Through analyses in R and ArcGIS, incorporating REDCap and U.S. Census ACS data, we assessed variations by age, race/ethnicity, gender, ZIP code, and hospital utilization. A notable 26% decline in overall hospitalizations was observed across the two seasons. However, this decline was less significant for Hispanic and Black populations. Adults aged 65 and older consistently experienced the highest hospitalization rates. Geospatial analysis identified clusters in urban counties, specifically Albany and Schenectady, with Albany Medical Center identified as the primary regional healthcare hub. The study concludes by emphasizing structural and geographic inequities in care access and recommending targeted vaccination outreach, improved surveillance, and equitable resource allocation for the Capital Region.

Presenter(s): Suruchi Shahi

Showcase Advisor: Muntasir Masum

Abstract: Preterm birth (PTB), defined as delivery before 37 completed weeks of gestation, affects 10.4% of U.S. births and remains a leading cause of neonatal morbidity and mortality. Breastfeeding provides important immune and developmental benefits, particularly for preterm infants, yet complications often hinder breastfeeding for them. We conducted a secondary analysis of 1,618 singleton births from the National Survey of Family Growth (NSFG) 2022–2023 to examine the association between gestational age (preterm, term, post-term) and breastfeeding initiation. Multivariable logistic regression estimated adjusted odds ratios (aORs) controlling for several covariates using survey weights. Overall, 81% of mothers initiated breastfeeding. Post-term birth was associated with higher odds of breastfeeding initiation (aOR=1.80, 95% CI: 1.14–2.94). Socioeconomic and racial disparities persisted, with non-Hispanic Black mothers having lower odds of breastfeeding initiation and mothers with higher income and education showing greater odds independent of gestational age. Targeted lactation support for mothers of preterm infants is necessary.

Presenter(s): Morgan Scarzafava (Muller)

Showcase Advisor: Muntasir Masum

Abstract: Discriminatory housing policies from the 1930s have been linked to health disparities across several outcomes. This study investigated whether exposure to historical redlining predicts binge drinking in adulthood, mediated by the alcohol environment.  
This study used Add Health data from Waves I-V. The exposure was residence in a redlined neighborhood. The outcome was binge drinking in early-midlife. We estimated survey-weighted logistic regression models, as well mediation and moderation effect.

Residing in a redlined neighborhood was associated with 17% higher odds of regular binge drinking. Alcohol outlet density mediated 14.4% of this association. The relationship was concentrated among non-Hispanic White respondents. Redlining did not predict binge drinking intensity.

Historical redlining is associated with elevated binge drinking prevalence in early midlife, operating partly through the alcohol environment. This effect observed under racial/ethnic stratification depicts racial patterning of structural determinants. Findings highlight the enduring behavioral health consequences of discriminatory housing policy.

Presenter(s): Izzy D'Ambro

Showcase Advisor: Wendy Patterson

Abstract: Background: Postpartum hemorrhage is a leading cause of preventable severe maternal morbidity in the United States. To strengthen maternal health surveillance at the New York State Department of Health (NYSDOH), we developed an analysis to identify hemorrhage events and risk patterns.

Methods: Delivery hospitalizations from 2021-2024 were identified from hospital inpatient data using International Classification of Diseases, 10th Revision (ICD-10) codes. Multivariable logistic regression assessed demographic and clinical predictors of hemorrhage.

Results: Among 770,946 deliveries in New York State, 1,754 (0.23%) involved hemorrhage. Higher odds of hemorrhage were predicted among Black or African American, Asian, and Hispanic/Latino patients and among those aged 35-44 years. Cesarean delivery and multiple gestation increased risk, while commercial insurance was protective when compared to Medicaid.

Conclusions: This analysis revealed persistent and intersecting disparities in maternal hemorrhage and provides a framework for equity-focused monitoring and prevention in New York State.

Presenter(s): Grace Jir

Showcase Advisor: Perry Smith

Abstract: During Summer 2025, I completed a public health internship with Refugee and Immigrant Support Services of Emmaus (RISSE) in Albany, New York, focusing on improving healthcare navigation and health literacy among refugee and immigrant communities. Many newcomers face barriers such as language differences, limited health literacy, and unfamiliarity with the U.S. healthcare system, which can lead to delayed care and poorer health outcomes. I collaborated with RISSE’s Family Services and Community Outreach Units to develop a health literacy and self-efficacy survey assessing confidence in navigating hospitals, scheduling appointments, communicating with providers, understanding post-visit instructions, and health insurance literacy. Survey findings showed that although many participants felt comfortable navigating healthcare facilities, significant gaps existed in understanding insurance and follow-up care, alongside common emotional health challenges. Based on these findings, multilingual resource guides, educational materials, and recommendations were developed to improve healthcare access and support culturally responsive community outreach initiatives.

Presenter(s): Lia Thompson

Showcase Advisor: Tomoko Udo

Abstract: Some research suggests that nearly one million non-fatal overdoses may occur annually. In recent years, increasing attention has been paid to the possible health complications of non-fatal overdoses, especially because many overdose cases do not result in 911 calls or emergency room visits. This study explores the factors associated with calling 911 and/or receiving services at the emergency room after experiencing an overdose among 289 PWID who completed the New York State Incidence Survey for Infectious Disease Elimination (InSIDE) from January 2023 to September 2025. Ethnicity and sharing injection equipment were associated with all three outcomes, and insurance was associated with ER visits. Race and homelessness were associated with calling 911 or doing at least one of the two. The results of this study demonstrate the importance of understanding the impact of the drug overdose epidemic beyond mortality and highlights the need for future research to examine community-level factors.

Presenter(s): Megan DiMaio

Showcase Advisor: Edward Valachovic

Abstract: We propose a data-driven framework for identifying and validating periodic principal components in time series, improving methods for analysis of complex cyclical phenomena such as environmental data. Building on the Variable Band Pass Block Bootstrap (VBPBB) proposed by Valachovic, which addresses dependence in correlated data, our approach replaces theoretically assumed component selection with an optimization-based maxima method. By smoothing periodograms, locating local maxima, and constructing adaptive analysis windows, candidate frequencies are extracted directly from observed data. Each potential component is tested for statistical significance using VBPBB filtering and bootstrapped confidence intervals. Simulation studies across low, medium, and high frequency signals with varying signal-to-noise ratios demonstrate strong performance, particularly in distinguishing harmonics and multiple concurrent periodicities. Challenges arise when components are closely spaced at very low frequencies, but alternative strategies improve detection. This work contributes a practical methodology for uncovering hidden temporal structure and improving inference in dependent time series data.

Presenter(s): Asees Dhaliwal

Showcase Advisor: Brooke Turcotte

Abstract: The County Health Indicators by Race and Ethnicity (CHIRE) Dashboard is vital for local health assessments and planning, providing 54 sociodemographic and health indicators by race and ethnicity at state, regional, and county levels. It informs strategic actions, resource allocation, and intervention planning to address health disparities, support the Prevention Agenda 2025-2030, and promote health equity. This project updated the CHIRE with the latest data available, including Native American data, expanding available regions, and enhancing the static HTML format with an interactive Tableau platform. The modernized dashboard now features dynamic multi-year trend analyses, interactive geographic comparisons, and exportable data tables, significantly enhancing accessibility, transparency, and usability. The improved platform empowers users with enhanced surveillance and decision-making capabilities to effectively advance health equity across the state.

Presenter(s): Allison Klein

Showcase Advisor: Brooke Turcotte

Abstract: The Perinatal ZIP Code Data Profiles supports local health departments and community partners by providing key perinatal indicators, including birth outcomes, maternal characteristics, and infant health measures at state, county, and subcounty levels. It informs local assessment, planning, and targeted intervention efforts to improve maternal and child health outcomes. This project modernized the previous static HTML tables and transitioned them into a fully interactive Tableau dashboard, incorporating updated demographic data and enhancing usability. Datasets were restructured and validated to support dynamic visualizations, with the redesign now featuring multi-year trend analyses, interactive geographic comparisons, ZIP code-level mapping, and exportable data tables. These enhancements significantly improve accessibility and transparency. The modernized dashboard strengthens public health surveillance and provides stakeholders with data-driven insights to identify high-risk populations, monitor disparities, and guide evidence-based decision-making across New York State.

Presenter(s): Patrico Tyrell

Showcase Advisor: Thomas O'Grady

Abstract: Background: Opioid–stimulant co-use is increasing and may complicate treatment decisions. Prior research suggests stimulant co-use is associated with lower medication for opioid use disorder (MOUD) initiation and poorer retention. We examined whether co-use affects treatment allocation at admission in national data.

Methods: We analyzed 2023 Treatment Episode Data Set–Admissions (TEDS-A) data, including 391,697 first-time opioid-related admissions (aged ≥12). Co-use was defined as reporting at least one opioid and one stimulant. Outcomes were MOUD receipt and placement in 24-hour treatment (detoxification, inpatient, or residential). Logistic regression estimated adjusted odds ratios (aORs).

Results: Overall, 8.8% reported co-use. Compared to opioid use alone, co-use was associated with higher odds of MOUD receipt (aOR=3.97) and 24-hour placement (aOR=1.82).

Conclusions: Co-use is associated with more intensive treatment at admission, suggesting recognition of greater clinical severity. Whether this translates into improved long-term outcomes requires further study.

Presenter(s): Abigail Ulofoshio

Showcase Advisor: Allison Appleton

Abstract: The postpartum period is a critical time physically and mentally for mothers. With rates of maternal mortality due to mental illness increasing, engaging mothers in routine, timely postpartum care is imperative. This project explores postpartum care attendance of Medicaid members with mental and behavioral health needs in New York State in 2023. Data from the Psychiatric Services and Clinical Knowledge Enhancement System (PSYCKES) were analyzed to investigate whether women had a postpartum visit on or between 7 and 84 days after delivery. Approximately 65% of our sample (n = 6,725/10,232) attended an outpatient postpartum care visit. Members with suicide-related diagnoses or prescribed mental health medications had lower odds of receiving timely postpartum care. Efforts toward improving postpartum care access and quality are imperative to prevent adverse outcomes and promote maternal and infant health in this population.

Presenter(s): Natalia Small

Showcase Advisor: Muntasir Masum

Abstract: This study investigates predictors of HIV/AIDS mortality in major U.S. cities to inform the development of new epidemiological tools for RNA pathogens. HIV, a rapidly mutating RNA retrovirus, causes severe immune suppression by targeting CD4+ T-lymphocytes. Despite the introduction of Highly Active Antiretroviral Therapy (HAART) in 1996, which significantly improved patient outcomes, research gaps remain regarding long-term transmission and mortality trends. After a comprehensive literature review was conducted, CDCWonder Surveillance data (1981-2002) employed assessment of descriptive statistics and multiple logistic regression in R to analyze variables like “yeardiagnosed”, “casedefinition”, “hivexposure”, and “sexandsexualorientation”. Results highlight significant associations between mortality and injection drug use, which is linked to natural CD4 cell declines. The study concludes that adjusting for demographic and scientific shifts is essential for refining surveillance. These findings establish a statistical foundation for future tools to better assess RNA pathogen pathogenicity and potentially discover ways to advance toward eliminating HIV mortality.

Presenter(s): Emmanuel Nana Arko

Showcase Advisor: Muntasir Masum

Abstract: Maternal mortality remains a critical public health challenge in the United States, with persistent racial and ethnic disparities. This project proposal outlines a secondary data analysis examining the extent and trajectory of these disparities between 2018 and 2023, using data from the National Center for Health Statistics National Vital Statistics System (Hoyert, 2025). Descriptive analyses include mortality trend analysis, race/ethnicity rate comparisons, age-stratified rates, and disparity ratios. Statistical methods include linear regression, Pearson correlation, and public health disparity measures such as rate ratios, absolute rate differences, and the Index of Disparity. In 2023, Black non-Hispanic women experienced a maternal mortality rate of 50.3 per 100,000 live births — more than three times the national average of 18.6 — compared to 14.5 for White non-Hispanic, 12.4 for Hispanic, and 10.7 for Asian non-Hispanic women.

Presenter(s): Mahshina Khanam

Showcase Advisor: Thomas O'Grady

Abstract: Childhood obesity is associated with long-term metabolic dysfunction and increased risk of type 2 diabetes, yet less is known about whether adults who retrospectively report childhood obesity exhibit higher diabetes-related biomarkers in midlife. We examined the association between retrospectively reported childhood obesity and elevated fasting glucose among U.S. adults aged 40–65 using 2017–March 2020 pre-pandemic NHANES data. Childhood obesity was defined using self-reported heaviest weight at age ≤17 years. Elevated fasting glucose (≥100 mg/dL) was the primary outcome. Survey-weighted logistic regression adjusted for demographic, socioeconomic, behavioral, and health-related covariates. Among 1,410 adults, 68.8% had elevated fasting glucose, 12 had childhood obesity. Childhood obesity while not significant in bivariate analyses, was inversely associated with elevated fasting glucose in adjusted models (aOR=0.18, 95% CI: 0.03–0.96). The association attenuated after adjusting for adult BMI, suggesting current adiposity predominates. Findings underscore measurement limitations and life-course approaches to diabetes prevention.

Presenter(s): Bikash Pal

Showcase Advisor: Edward Valachovic

Abstract: Seasonal variation in dengue incidence is well documented in Bangladesh, yet conventional analyses often rely on predefined seasonal categories or parametric models that may overlook overlapping and nonstandard periodic structures. This study examines seasonal and harmonic patterns in reported dengue cases using the Variable Multiple Bandpass Periodic Block Bootstrap (VMBPBB). Daily dengue case data were analyzed using Kolmogorov–Zurbenko Fourier Transform filters to isolate dominant periodic components prior to bootstrapping. Each component was resampled independently to preserve periodic correlation while reducing noise-driven variability. Statistically significant periodicities were identified through bootstrap confidence interval bands and aggregated to construct an overall periodic mean estimate. Results reveal strong annual and weekly cycles, with substantially narrower confidence intervals compared with conventional periodic bootstrap methods. These findings demonstrate that VMBPBB improves precision in characterizing dengue seasonality and provides a flexible framework for analyzing infectious disease time series influenced by climatic and environmental variability.

Presenter(s): Padmore Nana Prempeh

Showcase Advisor: Edward Valachovic

Abstract: This study examines periodic structure in weekly Lassa fever data from Nigeria using frequency-domain bootstrap methods. Understanding seasonal behavior in infectious disease incidence is essential for surveillance and intervention planning. We apply the Variable Bandpass Periodic Block Bootstrap (VBPBB) to isolate and infer periodic components in suspected, confirmed, and mortality series. The method first extracts target frequencies using a bandpass filter, then performs cycle-aligned resampling to preserve periodic dependence while reducing interference from nonperiodic variation. Resulting confidence intervals reveal a stable annual pattern in suspected and confirmed cases, with sharper bands and clearer seasonal interpretation than standard periodic block bootstrap approaches. VBPBB also improves detection of sub-annual structure that is less apparent under traditional methods. These findings provide a more precise characterization of seasonal timing in Lassa activity, supporting improved planning of surveillance and intervention efforts and motivating continued data collection for future analysis.

Presenter(s): Ummat Safwat Sristy

Showcase Advisor: Eirinn Noorie

Abstract: Insufficient sleep and hypertension disproportionately affect socioeconomically disadvantaged populations in the United States. Using 2021–2023 NHANES data, we conducted a cross-sectional analysis of adults aged ≥18 years (N = 5,561). Hypertension was defined as systolic BP ≥130 mmHg, diastolic BP ≥80 mmHg, or antihypertensive medication use. Sleep duration was categorized as short (<7 h), normal (7–9 h), or long (>9 h). Educational attainment served as a proxy for socioeconomic status and a potential effect modifier. Multivariable logistic regression adjusted for age, sex, race/ethnicity, BMI, smoking, and diabetes while accounting for NHANES survey design. In crude models, short (OR = 1.28) and long sleep (OR = 1.34) were associated with higher hypertension odds. After adjustment, only long sleep remained significant (aOR = 1.22, 95% CI: 1.02–1.47). Lower education was independently associated with hypertension, but no significant sleep–education interaction was observed (LRT p = 0.887). These findings suggest shared pathways linking sleep, socioeconomic status, and cardiometabolic risk.

Presenter(s): Asmaa Ahmad

Showcase Advisor: Edward Valachovic

Abstract: Incomplete time series data create major challenges for accurate analysis, especially when the data contain recurring seasonal or monthly patterns. Many common imputation methods overlook these temporal structures, which can distort trends and reduce reliability. This project introduces a structure-preserving imputation framework that improves how missing values are estimated by incorporating significant periodic components using the Variable Bandpass Periodic Block Bootstrap (VBPBB).

Simulated time series with annual and monthly cycles were generated under different noise levels to reflect real-world variability. Missing data were introduced at rates from 5% to 70%. VBPBB was used to identify dominant frequencies, which were then included in a multiple imputation model (Amelia II).

Results show that this periodicity-aware approach consistently outperforms standard methods, particularly in noisy conditions. By preserving underlying temporal structure, this framework enhances accuracy and offers a flexible solution for handling incomplete time series in public health, environmental, and other real-world data applications.

Presenter(s): Toria Udida

Showcase Advisor: Emily Lutterloh

Abstract: Temporary health care staffing agencies play a growing role in maintaining hospital workforce capacity across New York State, yet their financial impacts remain incompletely understood. During a 2025 internship with the New York State Department of Health, Article 29-K regulatory submission data from Q1–Q4 2024 were analyzed to examine statewide spending trends, agency utilization, and registered nurse bill-to-pay rate gaps. Findings revealed rising statewide expenditures and consistently higher per-bed agency costs in the North Country relative to New York City, suggesting rural regions bear a disproportionate financial burden when sourcing contingent labor. These findings were synthesized into a literature review and policy report with evidence-based recommendations targeting cost transparency, reporting requirements, and permanent staffing expansion. This work contributes to statewide efforts to build a sustainable, equitably distributed health care workforce across New York State.

Presenter(s): Alina Khan, Zara Syed

Showcase Advisor: Elizabeth Vasquez

Abstract: Our research aims to increase awareness regarding the health disparities present in communities lacking proper interpretation services. We want to shine light on its negative effects such as the over-reliance on Ad-hoc interpreters, as opposed to professional interpreters. Ad hoc interpreters, including children and untrained staff, make twice as many errors compared to professional translators. This leads to frequent miscommunication, lower-quality care, and unfair treatment for patients with limited English proficiency. Expanding access and improving current professional medical translation services will improve the doctor-patient relationship, ensure patient understanding, increase inclusivity, and create more equitable care for all communities.

Presenter(s): Denzel Edwards

Showcase Advisor: Thomas O'Grady

Abstract: Prior research has documented an immigrant advantage in select child health outcomes despite socioeconomic disadvantage. However, fewer studies have examined how generational status interacts with cumulative medical, social, and relational risks that shape child well-being. This cross-sectional study analyzed data from the 2023 National Survey of Children's Health (N=55,162) of children aged 6–17 years. Weighted, multivariable logistic regression estimated adjusted odds ratios (aOR) for neurobehavioral conditions across three generational subgroups, controlling for demographic, socioeconomic, and neighborhood support factors. A composite index of cumulative medical, social, and relational risk was examined as a moderator. Compared with United States-born children, first-generation (aOR=0.68; 95% CI: 0.27–1.72) and second-generation children (aOR=0.62; 95% CI: 0.38–1.00) had lower odds of nuerobehavioral problems. High cumulative risk was strongly associated with adverse outcomes (aOR=15.09; 95% CI: 13.24–17.20). No significant interaction was observed. Findings support an immigrant advantage while highlighting cumulative adversity.

Presenter(s): Jonas Barkevich

Showcase Advisor: Jemma Rowlands

Abstract: The Respiratory Virus Hospitalization Surveillance Network (RESP-NET) monitors hospitalizations associated with COVID-19, influenza, and respiratory syncytial virus (RSV). This analysis describes vaccination status ascertainment among RESP-NET cases in the New York Albany area during 2024-2025.

Surveillance data was stored in REDCap and analyzed using R. Vaccine status was ascertained through medical chart reviews, New York Immunization Information System search, and provider and patient interviews. About 25% of COVID-19 cases (n = 339), 25% of Flu cases (n = 372), and 30% of RSV cases (n = 163) were sampled, vaccine eligible, and included in this analysis. Vaccine status was documented for 68% of Flu cases, 52% of RSV cases, and 45% of COVID-19 cases. Only 40% of RSV cases under 2 years old had known vaccination status despite having the most thorough ascertainment process.  

Difference in vaccine status documentation rates highlight the need for continued surveillance improvement supported by standardization of RESP-NET processes.