As a culminating experience, students will put into practice the knowledge and skills they learned during their coursework through a capstone project. Students will have the opportunity to develop and implement a biomedical informatics project with an industry or University partner, or within their workplace.
The capstone project is a degree requirement for students and is completed during the last three quarters of their program. Students work in small teams with a business partner to address key problems the company needs to solve. The program aids students in identifying viable projects, and establishing a scientific advisory panel for oversight and mentorship. At our Capstone Showcase events, all projects are presented to faculty and sponsors for review and evaluation.
The capstone program is designed to offer both students and capstone partners an opportunity to gain experience working on real-life biomedical informatics related problems, enhance their research agenda, and explore potential employment partnerships.
Capstone teams engage with problems that may have wide-ranging effects in a variety of settings: clinical, research, and industry. Students identify the knowledge and framework required to address the problem and use the methodologies learned in the Biomedical Informatics program to develop strategies which may involve creating new information management resources, optimizing current data systems, conducting data analysis, and scoping new solutions.
The three quarter sequence includes the following elements:
In the Capstone Proposal (MSBI 39901) course, students will learn:
In the Capstone Implementation (MSBI 39902) course, students will learn:
In the Capstone Writing and Presentation (MSBI 39903) course, students will learn:
Description: Developed web-based database management system for acute care surgical residents.
Impact: Improved data collection and analysis for tracking patient status and estimate operative complication risks. Improved resident workflow and quality measures, provided residents with individual complication rates.
Description: Developed analytic template leveraging grouper methodology to examine health expenditures of a large corporation’s population.
Impact: Identified major drivers of population costs utilizing data analytics and visualization tools.
Description: Evaluated the frequency and causes of duplicate computed tomography (CT) scanning in receiving pediatric and adult trauma centers and considered use of electronic methods for image exchange.
Impact: Utilized scholarly research database to conduct literature review and concluded an industry-wide standards-based framework to facilitate the seamless electronic exchange of images is necessary to reduce duplication.
Description: Evaluated correlation between pre-operative lab data and post-discharge adverse outcomes in elective hip and knee joint replacement.
Impact: Identified significant laboratory tests, risk adjusted data, and used logistic regression to predict an adverse event. Concluded abnormal values of Albumin and Hemoglobin were significant predictors of prolonged length of stay in both hip and knee patients.
Description: Developed tool to assist clinical genomics group in handling the increasing volume of patient genetic data for a large healthcare system.
Impact: Utilized programming scripts to extract, transform and load data from dbSNP, ClinVar and COSMIC into postgreSQL database. Genetic information is now available through a single resource which helps with repeatability, documentation and incidental reporting.
Description: Gastroesophageal adenocarcinoma has a poor prognosis, high molecular heterogeneity and few targeted therapeutic options. Guardant360 is a clinical 73-gene next generation sequencing (NGS) panel for plasma circulating tumor (ct)DNA. Evaluated a global cohort of 1314 Guardant360 tests to determine correlations between allele frequency of ctDNA, median overall survival and immunotherapy-treated survival.
Impact: Concluded ctDNA analysis merits further evaluation as a prognostic and predictive biomarker and in evaluating molecular heterogeneity.
Description: A cancer center at a large university has developed a research data warehouse for translational research. Data is generated across multiple domains and stored in a centralized repository. Robust Extract-Transform-Load capabilities have been missing. Evaluated and made recommendations for ETL workflow.
Impact: Identified ETL workflow, informatics pipeline and data quality control strategies. Reviewed data collection process and documented risks to data quality. Proposed learning system approach for continuous data collection.
Description: The need exists to characterize disease occurring in population with moderate-to-severe psoriasis (PsO) that may not be applicable to mild PsO or the general population. Evaluated and identified cohorts based on EMR information.
Impact: Utilized EMR data to identify and stratify cohort of patients with PsO by severity based on their medication. Conducted descriptive and regression-based tree analyses to characterize each cohort. Concluded characteristics of those within the moderate-to-severe PsO cohort included advanced age, cardiovascular disease, diabetes, consistent with literature describing patients with more severe forms of PsO.
© Copyright 2016–2019 University of Chicago