Modern data science and analysis capabilities applied to clinical data sets has played a significant role in enabling many recent advances in medical research. However, each distinct clinical data set has inherent traits that dictate its utility for providing relevant answers to particular clinical questions. Characteristics such as the types of data elements being collected, frequency of collection, structured vs. unstructured format, consistency, completeness, and accuracy all play a major role in determining the overall effectiveness of the data for specific research studies. Quantifying and measuring data set attributes requires a deep understanding of the underlying clinical workflows that generate the data as well as a working knowledge of modern data modeling, capture, and curation practices. This elective will introduce students to techniques for identifying and analyzing the characteristics of patient data sets and provide hands-on learning opportunities for students to directly assess a data set's capacity for deconstructing various classes of clinical questions. Students will utilize patient data directly MASracted from the VA healthcare system's electronic health record (EHR) that has been secured in a pre-approved repository, as well as other well-known open source data sets. The VA EHR data contains clinical data such as demographics, medications, laboratory values, hospitalizations, surgical procedures, progress notes, and radiology reports. Examples of open source data include The Cancer Genome Atlas (TCGA) clinical and molecular data and the MIMIC-III critical care database. Questions that have been asked of the clinical data in the past have included determining the number of patients with lung cancer who had been identified prior to and after the LDCT national recommendation. Students will receive didactic instruction on the VA's clinical workflows as well as basic database principles such as relational data modeling and the Structured Query Language (SQL). Didactic instruction covers several introductory data science topics including database systems, the Structured Query Language, and standard analysis techniques. Laboratory sessions provide students with practical mentorship to ensure projects make timely progress. Students are familiarized with the workflows, processes, and practices involved in data generation that dictate the usefulness of the data. No previous data science knowledge is required but fundamental computer skills are necessary. Students will be required to work with the research mentors to obtain access to clinical data sets as a prerequisite approximately 8 weeks before the elective begins. At the beginning of the elective, students will meet with the research mentors to discuss research aims and formulate a preliminary timeline for the elective's major milestones. The timeline for the elective will generally consist of: finalization of research questions to be addressed and initial access to and familiar

Note that this information may change at any time. Please visit the MyBU Student Portal for the most up-to-date course information.