2022 Annual Symposium

Pipelines Into Biostatistics Annual Symposium

Harvard T.H. Chan School of Public Health
Thursday, July 14, 2022

Opening Remarks and Introductions
10:00 – 10:30am

Marcello Pagano
Professor of Statistical Computing and Principal Investigator, Harvard Chan School

John Quackenbush
Henry Pickering Walcott Professor of Computational Biology and Bioinformatics
Chair, Department of Biostatistics, Harvard Chan School

Amarildo “Lilu” Barbosa
Chief Diversity, Inclusion & Belonging Officer, Harvard Chan School

Keynote Speaker
10:30 – 11:30am

“LORD8DEERJAGUARCLAW: Embracing Uncertainty in Your Journey
Miguel Marino, Ph.D.
Associate Professor of Family Medicine, School of Medicine
Associate Professor, OHSU-PSU School of Public Health
Oregon Health & Science University

Summer Program Research Project Presentations

11:35 – 12:00pm       The Effect of Neighborhood Socioeconomic Status on the Single Ventricle Heart Disease Surgical Outcomes in Infants

Infants born with hearts unable to efficiently pump blood to the body must undergo corrective surgery in the first days of life. Single ventricle heart defects affect ~1 in 10,000 newborns, have relatively high surgical mortality compared to other congenital heart defects, and require costly hospital stays and extensive follow-up care. We used data from Boston Children’s Hospital between 1997 and 2017 to determine, after accounting for known risk factors such as low birth weight and prematurity, whether socioeconomic status has an effect on surgical outcomes. We used a socioeconomic status neighborhood score previously derived from US census data that incorporates income, educational attainment, and employment data. In-hospital mortality or transplant was analyzed using logistic regression. Post-discharge mortality or transplant was analyzed using Cox proportional hazards regression. Postoperative ICU length of stay was analyzed using a generalized linear model with a gamma distribution. The results highlight that low SES patients were at higher risk for mortality or transplant and longer ICU stays, even after adjusting for known risk factors.

Adina Cazacu-De Luca, Columbia University ‘24
Yuki Low, University of Michigan ‘23
Asma Asghar, Barnard College ‘22

Faculty Mentor: Kimberlee Gauvreau, Associate Professor in the Department of Biostatistics, HSPH, Associate Professor of Pediatrics, HMS
Graduate Student Mentor: Kexin Yang, HSPH

12:05 – 12:30pm     Can Antigen Home Tests be Used for Population-Level COVID-19
Surveillance?

COVID-19 is an infectious respiratory virus that has had a large impact on global health during the last 2 years. For this research project, we investigated if COVID-19 antigen home tests, which are much more economical than molecular tests, can be used for population-level COVID-19 surveillance. Molecular test results are used to inform the positivity rate of the island. However, since December 2021, the Department of Health in Puerto Rico has allowed citizens to report their COVID-19 hometest results. We studied both tests to find similarities in the daily and weekly trend of the positivity rate. In addition, we looked at age populations and certain time periods in order to find similar patterns between tests. Although not yet as reliable as molecular test, we find that self-reported antigen home test results can potentially be used for COVID-19 surveillance.

Michael H. Terrefortes Rosado, University of Puerto Rico ‘23
Javlon Nizomov, University of Florida ‘24

Faculty Mentor: Rafael Irizarry, Professor of Biostatistics, DFCI, HSPH
Graduate Student Mentor: Sijia Huo, HSPH

Lunch Break

12:30 – 1:00pm

Where are They Now: Alumni Panel

1:00 – 2:00pm

Morjoriee White, Summer Program Alumni ’05
Homeless Administrator, City of San Antonio Department of Human Services (DHS)
MPH from Emory University

Quincy Greene, Summer Program Alumni ’06
Data Manager, Center for Disease Control and Prevention/IHRC
MPH Student at Emory University

David Angeles, Summer Program Alumni ’17
PhD Student in Biostatistics at Ohio State University

Ula Widocki, Summer Program Alumni ’17
PhD Student in Network Science at Northeastern University

Summer Program Research Project Presentations (continued)

2:00 – 2:25pm           Analyzing the Ovarian Cancer Disparity based on Biological Patterns and Socioeconomic Status

Ovarian cancer is a known treatable disease if caught in time, but there is a disparity in what type of women are diagnosed with the disease. For our research project, we are looking at the biological and social mechanisms behind the ovarian cancer disparity. We used a K-means clustering analysis in order to identify groups of women that have similar biological patterns in their microRNA. We also used the Socioeconomic Disadvantage Index (SDI) scores to look at geographical and demographic similarities between those diagnosed with ovarian cancer and those not diagnosed with the disease. Our purpose is to see if there is any significant association between the biological patterns or socioeconomic status (SES) of patients with ovarian cancer compared to those that do not have an ovarian cancer diagnosis. 

Maya Lightfoot, Tufts University ’23
Sean O’Connor, Lake Forest College ’24
Amia Graye, Georgetown University ’23

Faculty Mentor: Briana Stephenson, Assistant Professor of Biostatistics, HSPH
Graduate Student Mentor: Carmen Rodriguez Cabrera, HSPH

2:30 – 2:55pm      Statistical Methods for Exposome Wide Analysis Studies

The exposome is defined as the totality of all the external, non-genetic factors an individual is exposed to from conception and onwards in their lifetime, and how it relates to health. Geneticists have delivered various approaches to comprehensively analyze our genome, from genome-wide association studies (GWAS) to deep-sequencing, yet genetics can only account for about 10% of diseases. The analysis of the environmental drivers of human health has lacked a similarly comprehensive approach. Ultimately, the analysis of the environmental drivers of human health can aid in identifying unknown exposures, revolutionize our understanding of the underlying causes of disease and aid in the development of preventions and cures for more diseases. Our aim was to provide statistical approaches to study the health effects of complex external factors from a dataset which included multiple health outcomes (continuous and categorical), multiple exposures, -omics and additional non-exposure variables (e.g., potential confounders). Linear or logistic regression models were used to identify significant variables within subgroups in the categories and overall. This process was aided by the correlation analysis, existing literature of the outcomes, and the FOCI (Feature Ordering by Conditional Independence) algorithm.

Samantha Reynoso, University of Connecticut ‘23
Dayanlee De León Cordero, University of Puerto Rico ‘23
José Constantino Sánchez Curet, University of Puerto Rico ‘23

Faculty Mentor: Rajarshi Mukherjee, Assistant Professor of Biostatistics and Assistant Director of Graduate Studies, HSPH
Graduate Student Mentor: Sean McGrath, HSPH

2:55 – 3:00pm             Closing Remarks

Marcello Pagano
Professor of Statistical Computing and Principal Investigator, HSPH