At Stanford Medicine we are committed to our tripartite mission — to educate tomorrow’s leaders; make fundamental scientific discoveries; and transform patient care — using the most cutting-edge technology. Technology & Digital Solutions is committed to harnessing Artificial Intelligence in serving this mission for our researchers, care teams,  learners, and staff as we drive medicine forward together.

Stanford Medicine has been recognized as an award-winning leader in health care AI. Our leaders are at the forefront in shaping the national guidelines for responsible AI in healthcare and in guiding the adoption of AI in clinical practice.
 

In healthcare, AI has to be fair, usable, and reliable. Stanford Medicine’s Chief Data Scientist, Dr. Nigam Shah, developed a framework called FURM (Fair, Useful, Reliable Model) to help us decide what and how to implement AI at scale. We’ve adopted this model at Stanford Medicine for all AI initiatives.

RAISE Health, a joint initiative between Stanford Medicine and the Stanford Institute for Human-Centered Artificial Intelligence (HAI) guides the responsible use of AI across biomedical research, education, and patient care throughout our organization.

To view FURM framework diagram, visit the Shah lab
 

FURM stands for Fair, Useful, Reliable Models. FURM Assessments (see stage 1 in the figure) help articulate the potential value of using a classifier or predicter (i.e. a model) to drive a specific clinical or operational workflow. The assessment quantifies whether a specific AI-guided workflow is likely to be sustainable and have enterprise value.

The process is based on 23+ peer-reviewed publications by members of the data science team. The output of the assessment is a report summarizing the eligible patient population, the workflow driven by the model, the anticipated business case, and ethical considerations. The report helps different stakeholders understand the overall impact of a specific model-workflow pair to ensure responsible use of AI at Stanford Healthcare.