Harvard Catalyst Profiles

Multiple projects are available for highly motivated student in the field of ophthalmic epidemiology and data science. Background in statistics and knowledge of statistical and computer programming is helpful. 1. Genomics and disease risk prediction. Using two large biobanks and results from a large ophthalmic clinical trial, these projects aim to understand the influence of polygenic risk on glaucoma disease risk, phenotype, and progression. 2. Machine learning. Using large existing datasets these projects aim to use machine learning for better disease prognostication. Unsupervised and semi supervised deep learning models will be used on large repositories of ophthalmic imaging to understand subtypes of disease and characterize progression. The goal is to define structural and longitudinal endophenotypes that are closely aligned with disease type, severity, and progression. 3. Outcomes and health equity research. Multiple datasets are available for use including a large ophthalmic registry, large EHR database and Medicare datasets. We have previously used these datasets to understand the effectiveness of new surgical procedures in the US and racial healthcare inequities among glaucoma patients. Student role: The primary goal of these projects is for the students to learn how to handle large amounts of healthcare data and perform appropriate statistical analysis. The student will formulate the research question and use existing data to answer these questions. The goal is that the student will take ownership of a project and bring it to fruition by planning and completing at least one manuscript. The student can decide on the question most interesting to them or work with the mentor to join an existing project. Student and mentor will meet at least weekly (and more often as needed) to ensure adequate progress.