Skip to main content
Data pattern

UCSF Center for Real World Evidence

Actionable answers from rich clinical data and incisive AI, guided by experts.

Our Approach


Real-world evidence begins with data drawn from a variety of sources beyond clinical trials. The Center for Real World Evidence (CRWE) draws insights from billions of de-identified data points that have been aggregated and harmonized on an ongoing basis since 2012 from both UCSF (3M patients) and the entire University of California Health system (7M patients). We have access to structured data from electronic health records, as well as many varieties of unstructured data, including clinical notes, cancer genomics, physiological waveforms, and diagnostic imaging. These proprietary datasets are the foundation for our efforts to improve health care outcomes by bridging the gap between educated guesses and inferences drawn from real-world evidence.


Cutting-edge tools translate real-world data into the evidence necessary to make key decisions. The CRWE employs cutting edge-data science techniques to analyze data in a way that is methodologically appropriate and reliable, often creating new software tools adapted to project demands. Our proprietary technologies include a suite of clinical NLP tools, a customized BERT deep learning model, a high performance and PHI-compliant supercomputing cluster called UCSF Wynton, as well as UCSF Information Commons, a state of the art data science platform supporting all forms of AI modeling, and a custom annotation environment to support gold-standard labeling of text and images. Robust policies and computational infrastructure allow us to safely and respectfully use patient data.


Data and tools require human intelligence and insights for successful deployment. Partners working with the CRWE tap the knowledge base of a leading institution with more than a century of global impact. Our domain specialists include expert data scientists, epidemiologists, biostatisticians, and faculty clinicians in all fields of medicine. The CRWE collaborates with life sciences companies, molecular labs, and regulators, among other partners. Capitalizing on the broadest spectrum of know-how in RWE, we are able to meet the evolving needs of our partners. It is an honor to bring the strengths of a global community in an emerging field to respond to today‚Äôs complex health problems.

Quick Facts

Expertise: Partner to the U.S. FDA and top global biopharma firms

Breadth: Electronic health records data for 3 million patients at UCSF, 7 million patients through University of California Health

Scholarship: Publications each month in leading medical journals

Depth: 110 million de-identified clinical notes, images and more

Diversity: We access data from the University of California Health system, representing the diversity of California

Our Partners

UCSF logo
UCLA logo
UC Davis
UCSD logo
UCI logo
UC Riverside

We use a central repository for data from all six University of California health systems

Making reliable inferences from the unruly world of observational clinical data isn't easy. To draw the right conclusions, we must integrate expertise across multiple domains, from clinical to technological.

Vivek Rudrapatna, MD, PhD, CRWE co-director

Our team

Selected investigators who specialize in using real-world evidence to generate actionable insights include:

Atul Butte
Atul Butte, MD, PhD
CRWE co-director
Julian Hong
Julian Hong, MD, PhD
Assistant Professor of Radiation Oncology
Vivek Rudrapatna
Vivek Rudrapatna, MD, PhD
CRWE co-director
Marina Sirota
Marina Sirota, PhD
Associate Professor, Pediatrics
Julia Adler
Julia Adler-Milstein, PhD
Professor of Medicine
Alejandro Sweet-Cordero
Alejandro Sweet-Cordero, MD
Associate Professor of Pediatrics
Rima Arnaout
Rima Arnaout, MD
Associate Professor of Cardiology
Rohit Vashisht, PhD
Clinical Data Scientist

Problem-solving with our Partners

Today, we are proud to be verifying product value and allowing partners to make complex decisions with confidence: 

  • Real-world effectiveness and safety of COVID-19 vaccines
  • Decoding patient journeys to reduce diagnostic delays in individuals with rare and unrecognized conditions
  • Comparative effectiveness and precision therapy for patients with inflammatory bowel disease
  • Computer vision algorithms to diagnose congenital heart disease
  • Risk stratifying patients with metabolic-associated fatty liver disease
  • Monitoring drug safety using sentinel programs that continually mine data from clinical notes