Like many organizations, UCSF Health is rapidly growing. Helendale Hospital, a $3 billion hospital tower, is set to open at the flagship San Francisco campus in 2030, while a second tower at the children’s location in Oakland is expected to open the following year. 

It’s a heavy lift – both literally and figuratively. And what’s just as critical as the physical design is the Connected Care Roadmap that incorporates the “enterprise decisions for a digitally enabled care environment,” according to CNIO Kay Burke. 

However, with AI and other technology advancing so quickly, it’s becoming increasingly difficult to ensure the right infrastructure is in place, particularly when architectural blueprints are finalized years in advance.

What does that mean for healthcare leaders?

“The roadmap is very fluid,” said Burke. “It’s a reflection of where we are now, which is going to change.”

Recently, she spoke with This Week Health about how the Connected Care roadmap is laid out (for now), along with other key priorities, including an initiative designed to reduce cognitive burden on nurses, a research-backed AI literacy program, and a fundamental rethinking of health IT training.

A Map Without an End

In addition to new construction, UCSF is also acquiring community hospitals and adding several outpatient centers, which is why having a roadmap is so critical. And not just a fixed plan with a completion date, but a living framework that can accommodate an evolving digital strategy. 

The roadmap, according to Burke, covers four domains: an ambient intelligent care environment built on AI automation and passive workflows; a frictionless clinician experience centered on smart clinical communications and alarm management; an integrated patient journey that connects in-room interactive tools to the EHR; and enterprise-scale platform standardization to replace the point-solution sprawl that can result in uneven experiences across UCSF’s locations.

“The connected care roadmap doesn’t have an end. It’s about what are the tactics, capabilities, or areas of focus for fiscal years 27, 28, 29 – what we need to implement enterprise wide before any new hospitals open,” she said.

One example is AI-enabled cameras that can detect wounds based on pupil size, a functionality that is just starting to proliferate on the market. “This is a frenetic space, and we need to be positioned to make the right choices for our patients and our clinicians,” she said. 

Care Before Clicks

Alongside the roadmap work, Burke is running a parallel effort to undo fifteen years of EHR accumulation. The “Care Before Clicks” project is, in her words, an un-build, actively removing documentation requirements and configurations from a production environment rather than adding to them. UCSF’s Epic instance has been live since approximately 2010, and has grown continuously without a corresponding effort to decommission what no longer serves clinical practice.

The design of the EHR has diverged from the reality of clinical practice, and closing that gap requires going to gemba: shadowing frontline staff, running design workshops that interrogate the status quo, and challenging the assumption that every regulatory requirement mandates nurse documentation.

“There’s a lot of research stating that after so many hours and so many alerts, the cognitive burden actually creates patient safety risks because they’re no longer actually reacting to the right alarms, the right alerts, the right BPAs,” Burke said. “Care Before Clicks can help by scrutinizing the amount of clicks and documentation requirements and ratcheting those back significantly.”

Reimagined Epic Training

A critical aspect in improving usability, of course, is in education, which has been a top priority for Burke. When she took over what was called the Epic Apex training team, she found it to be narrow and implementation-centric. The way it was situated, trainers could walk a nurse through point-and-click Epic workflows but could not answer basic questions about clinical policy or standards of practice.

Burke has since renamed it as the Health IT Learning Center of Excellence, broadening its scope to cover the full range of digital tools used by nurses and allied health professionals, including clinical communications, patient-facing applications, and ambient documentation tools alongside the EHR. The redesign introduced personalized learning through a platform called Amplifier, which assesses each user’s confidence and competency before tailoring the module to their actual need. 

As a result, a nurse who needs one hour of targeted reinforcement no longer sits through an eight-hour class, which can lead to improvements in efficiency and cost savings.

“We re-envisioned EHR training to focus on learning as an ongoing journey as opposed to training as a point in time,” she said, while also embedding informatics into the curricula.

UCSF’s enterprise relationship with OpenAI has also opened a new channel. The team is ingesting learning materials, knowledge base content, and course curricula into custom ChatGPTs, enabling just-in-time self-service support that does not require scheduling or pulling staff off the unit. 

In addition, Burke is piloting Purposeful Rounding, which uses Epic time-motion data and Slicer Dicer analytics to direct informaticists and trainers to specific problem units and workflows, replacing open-ended check-ins with targeted, measurable interventions. Outcomes will be measured through UCSF’s participation in KLAS’ Arch Collaborative EHR satisfaction survey, which recently went out to 14,000 users across nursing and allied health.

The Patient Voice

What’s just as important, however, is ensuring patients and caregivers are able to utilize technology safely and effectively. The problem is that patient-facing digital experiences are often designed without patients at the table, a trend that UCSF is looking to buck.

“Patients expect convenience. They expect digital access. They expect real time updates, financial transparency and personalization. It's a very consumer centric world,” she said.

Through the Patient Experience Technology portfolio, her team is working to become more intentional about including patients in the decision making processes.

“They have a choice in where to seek care, and if they don’t feel they have a convenient option, they’ll go elsewhere,” Burke noted. 

As such, her team recently launched an initiative to publish ED wait times across UCSF’s locations on a public website, offering a level of transparency that hasn’t necessarily been available in the past. There are also plans in place for in-room interactive experiences that integrate with the EHR rather than functioning as infotainment layered on top of it.

“These are the types of thighs we’re prioritizing to support the expectation of convenience,” she said. “There’s a lot of work going on in that space.”

Nursing & AI

Of course, as digital and AI tools become more pervasive across healthcare, another cohort is greatly impacted. “I spend a lot of time thinking about the nursing vision within the very busy AI landscape,” said Burke, who is pursuing a doctorate with a focus on nurses’ perceptions of AI in clinical practice. According to her research, the common assumption that younger, less experienced nurses are more receptive to AI is not supported by data.

In fact, she has found no meaningful correlation between years of experience and willingness to adopt AI. What she has found is that an educational intervention that encompasses the human-in-the-loop imperative, critical thinking as a clinical contribution, ethical and safety considerations, and governance participation can yield results. At UCSF, it has produced a 22 percent improvement in AI perceptions and attitudes among participants.

“As nursing continues to become a more prominent figure in governance, we’re moving away from experimenting and toward operationalization to support them,” especially when it comes to leveraging AI. “There’s a lot of fear, misinformation and propaganda out there. A lot of what I’m trying to do is demystify and make sure it’s working for our frontline nurses.”