From the vantage point of an active CDHI data scientist, we will demonstrate how a highly modern, zero-trust, cloud-native, data science environment can be used for active distribution of complex data science tasks in a heterogenous research compute environment. Expect an introduction and real-world examples of how technologies that include Kubernetes, Jupyter, and Kubeflow can be used to improve data science. Time permitting, we will close by explaining how the use of Kubernetes provides broad-based compute freedom to embrace existing cloud, enterprise, and national resourcing.

As neighborhood factors and built environment become increasingly relevant for understanding the pathways to certain health behavior and outcomes, Geographic Information Systems (GIS) offers a variety of opportunity to study these linkages. The availability of location data at the individual level, such as residential locations, along with census data and freely available spatial data from a variety of resources provide an opportunity to assess the context of place and the people that drive differentials in health outcomes. 

Flipped classrooms is a process to help address the needs of adult learners by developing learning opportunities which blend e-learning and classroom learning. At UCSF Health the Clinical Nurse Educators have embraced this concept and added the theoretical framework from the Six Disciplines of Breakthrough Learning with the digital age learning theory of connectivism. Participants who attend an in person learning event will no longer expect to be stagnant thinkers while listening to a presenter read bullet points from PowerPoint slides.

UCSF performs over 300,000 imaging studies annually. UCSF Radiology has been invested in innovation, and digital transformation with a rich history dating back over 20 years. We are continuing this trend by driving toward improving patient access through improved scheduling processes and access to imaging studies.

Come learn how to securely use, store, and share UCSF data sets in accordance with UCSF and UC policies and state and federal laws. This session will also include an overview of the new Acknowledgement of Liability of the Use of Enterprise De-identified Data sources to introduce key concepts in the use, sharing, and securing of de-identified clinical data sets.

Substantial effort over the past decade has resulted in greater health system connectivity by enabling electronic availability of outside records for frontline clinicians. However, most approaches to enabling such availability require clinicians to go outside of their standard workflow.

Slack is the leading workstream collaboration (WSC) tool, and has been adopted in a grass-roots fashion by many groups across UCSF. Currently, there are over 500 Slack teams at UCSF, including over 28 paid teams with over 1,000 paying user licenses, and over 10,000 unpaid licenses.

This session will examine the potential sources of bias when building artificial intelligence (AI) systems in health care and lay out a framework for proactively identifying and preventing or mitigating bias . Topics will include identifying biases due to missing data and groups not identified by machine learning algorithms, sample size and underestimation, and misclassification and measurement error.

National registries offer tools to measure and improve care, conduct novel research studies, and participate in federal quality reporting programs. There are now over 100 federally Qualified Clinical Data Registries in the United States, and many more local and statewide registries. At a time when the need for registries is growing, so too are the legal and regulatory challenges.

In 2012, the UC Office of the President awarded an innovation grant to the Patient Support Corps (PSC) as a collaborative project between UCSF and UC Berkeley. Our program trains UC Berkeley students to serve as patient scribes who connect patients with valid information online; list patient questions; take notes; and make audio-recordings of pivotal medical consultations.