For emergency room nurse Maranda Bradshaw, AI helps reduce the stress that she calls 'stacking cognitive fatigue.'
“Think of it this way: for any given procedure that I perform, there could be 20, 30, 40 steps that I follow, and each step is represented by a stackable plate. A procedure’s “stack of plates” symbolizes your memory load. The larger the memory load or stack of plates, the greater the cognitive strain. Strain leads to fatigue, and fatigue leads to mistakes. As an example, take a routine procedure of setting up an IV infusion.
I have to get the IV start kit, the needle, the chlorhexidine (an antiseptic that helps prevent infections), an IV extension, and the right fluids. Let’s say in the middle of the procedure the charge nurse sees that I need to draw a lactate to help with resuscitation. So, I take the plate that I was just working on out of the stack, put it to the side and add the lactate ‘plate’ to the top of my stack. Now the memory stack isn’t organized anymore.
AI helps me not to drop or misplace the plates. The technology captures the process steps that we as nurses are already manually doing, prompts me on the computer and helps keep me organized with checkpoints.”
- Maranda Bradshaw, RN MSN, nurse manager for emergency services at UC San Diego Health.
Bradshaw’s example is just one of the many ways in which artificial intelligence can help to improve nurses’ jobs: it augments clinical nurses’ practice skills and responsibilities, integrates different databases and technology systems, and ultimately helps make their work more efficient.
Kay Burke, MBA, BSN, RN, NE-BC, VP and chief nursing informatics officer at UCSF Health, shares, “Sometimes we refer to this as ‘back to the bedside’ — allowing nurses to get off computers, out of documentation software, and back to taking care of patients.”
Burke also notes how AI can help reduce nurses’ cognitive burden. “Nurses spend so much brain power on what to do next. AI can help prioritize these tasks so that nurses know what’s next in a procedure. It provides a quick and efficient checklist.”
UC Health recognizes the critical role that nurses play in the health care system
According to the American Association of Colleges of Nursing, nurses represent the largest segment of the health care workforce in the United States, with 4.7 million registered nurses in the field. In California, there are approximately 542,000 active registered nurses.
At UC Health, nurses are an integral part of the programs and initiatives that examine how AI can be integrated into the health care landscape to improve both their professional roles and the experiences of their patients. Nurses across the UC system sit on review and advisory committees that evaluate and select new technologies, and many of the use cases for applications are submitted by UC nurses working every day with patients.
The immediate key benefit of AI: decreasing the time documenting and performing administrative tasks
Across the board, clinicians say that decreasing the time they spend documenting charts in electronic health records is one of the chief benefits of having AI integrated into their workflow.
According to Donna Wellbaum, MSN, RN, NEA-BC, NI-BC, chief nursing informatics officer at UCLA Health, “In a 12-hour shift, on average a nurse spends 132 minutes documenting patient information in the EHR (electronic health record) system.”
That’s approximately 18 percent of a nurse’s time during one 12-hour shift. That time also doesn’t account for other research that a nurse may need to conduct outside of the EHR to consult other databases and systems on policies, patient education, standards of care or other procedures.
Credit: UC Health
Burke, from UCSF Health, offers an example that’s familiar to anyone who composes emails or text documents. “AI use in EHRs is similar to the predictive function that’s been used in many email and word processing applications. When you type a sentence, the system anticipates the next word or phrase that you may be thinking of using. In the EHR, the system suggests the next sentence or two based on the data that’s already in the patient’s chart or record. The nurse can then swiftly confirm or reject that suggestion. Each step saves time that that nurse can then spend with direct patient care.”
Wellbaum agrees. “Nurses spend a lot of time in the charts. AI helps them put that time back with the patients.”
Integrating protocols, policies and procedures
Frontline nurses and nurses in management roles both cite another key benefit that AI brings to the profession: being able to collect information from several sources and systems on a single screen to help them get what they need to help their patients.
Wellbaum says that “AI can help survey separate databases so that when a nurse needs to complete documentation, that nurse can ask the system to help consolidate the documentation required.”
With an AI overlay, nurses now no longer need to “hunt and peck” across different systems to find patient education, medication information, disease processes, policies and procedures, standards of care models or even a refresher on the steps of a routine clinical task they may not have done in a while. But even with those time savings, nurses are still the ones evaluating the suggestions.
Burke states: “Nurses will always need to apply clinical judgment and critical thinking in how they take care of their patients…. Every AI tool necessitates a clinician to review and validate what was suggested.”
Whether it's in an inpatient, outpatient, clinical or emergency department setting, every minute that a nurse can spend directly with a patient and not in a technical system can result in a better outcome for the patient.
AI in nursing and health care: This is not new
Seasoned nurses point out that AI-type models have been in health care long before the current buzz around the new technology, its offerings and its possibilities. To support and recommend nursing interventions, Burke points to various AI applications that have already assisted nurses in delivering care, including decision support systems, risk prediction scores, and early deterioration models.
Burke adds, “As we know, AI is – put simply – computer programs that process reasoning, designed to act rationally by using intelligent software to support clinical decision-making. This is not new.”
Bradshaw, an ER nurse, agrees.
Over her 25-year career as a nurse, Bradshaw remembers when nurses created paper-based processes and decision trees to document best practice alerts, and in particular around steps to take for addressing sepsis. (Sepsis is a condition where the body responds improperly to an infection and that infection spreads to other organs and vital systems.) This type of logic work was the forerunner to today’s computer-driven algorithms.
For a time, sepsis was a clinical priority where every patient in an emergency room setting would be subject to the labor-intensive and multi-step protocol to treat it, including repeated check-ins to give fluids and monitor vital signs at regular intervals. But every patient did not require that protocol.
“Now, an AI model goes beyond what we could work through with paper charts and decision trees to look at all the data about the patient and helps us weed out those who aren’t seriously at risk for sepsis. This technology-supported model allows us to focus on caring for those who are most in need of care.”
Nurses drive AI innovation in health care
Across UC’s health system, nurses are at the center of the conversations around AI, its possibilities, its applications and its limitations in current clinical settings.
At UCSF Health, Burke notes “We don’t make top-down decisions when it comes to nursing practices. We ensure we have frontline nurses who participate and drive decision-making. The nurses drive practice changes and innovation.”
At UC San Diego Health, there’s both an AI Thinkshop program and a Health AI Committee that evaluates all ideas for on their feasibility and safety, many of which come from frontline nurses. All ideas go through a thorough review process to ensure that any implemented AI meets the standards of the health system and is thoughtfully implemented with input from a diverse group of stakeholders. In addition, Bradshaw, a committee member and past submitter of potential project ideas, said that the committee evaluates each submission to determine if the use of AI is appropriate, transparent, accurate, reliable, safe, and fair. Through coordination with legal, compliance, and information security teams, the committee is also able to ensure that patients’ privacy is protected in the use of AI.
Ellen Pollack, MSN, RN-BC, chief information officer at UCLA Health Sciences, stated that “we will never implement any AI technology involving patient care without nurses giving input.”
Another example of how nurse managers and nurses are leading AI innovation is UCSF Health’s “conversational care” project. A pilot program currently in discovery, the project examines how a voice-activated enhancement could file information in the appropriate place in the EHR. If it meets the standards of care, the nurses also would lead its implementation and measure its success.
Looking forward
The nurses who shared their thoughts all agreed that AI currently is opening up opportunities for the profession and will continue to do so moving forward.
Burke believes that “in five to seven years, nurses will not remember life without AI as part of their everyday workflows. Something as simple as having that intelligence at their fingertips to help them draft messages and notes can increase the time they can spend with patients.”
Bradshaw agrees. “It’s wild how much a simple machine that helps remind you what we know is a clinical standard in the clinical decision-making helps you correct all those things.” Nurses continue to play a central role in the decision-making processes around AI in health care.
Bradshaw concludes: “Nurses are writing the new history of health care.”
