AHCJ – Association of Health Care Journalists

08/09/2024 | News release | Distributed by Public on 08/09/2024 15:23

California nurses protest ‘untested’ AI as it proliferates in health care

In April, hundreds of registered nurses and members of the California Nurses Association from across the state held a protest at Kaiser Permanente's San Francisco Medical Center to highlight their concerns with the hospital industry implementing untested and unregulated AI. Photo by Jaclyn Higgs/California Nurses Association

Back in April, about 200 California nurses and members of the California Nurses Association (CNA) protested outside Kaiser Permanente's San Francisco Medical Center to express their patient safety concerns about artificial intelligence (AI), and what they say is the hospital industry's rush to implement "untested and unregulated AI." They held signs saying, "Trust Nurses, Not AI," and "Patients are NOT Algorithms."

As AI becomes more rapidly integrated into health care settings, the protest serves as an important reminder for both health care leaders and journalists that all stakeholders - including nurses - should be involved in the testing and deployment of new health care technologies that impact their area of practice. Reporters writing about such technologies should ask health care leaders how particular technologies were selected and tested before use, and get feedback from users about how it's going.

More about the protest

"Nurses are all for tech that enhances our skills and the patient care experience," said Michelle Gutierrez Vo, a registered nurse at Kaiser Permanente Fremont Medical Center and president of the CNA, in a prepared statement. "But what we are witnessing in our hospitals is the degradation and devaluation of our nursing practice through the use of these untested technologies." The statement demanded that workers and unions be fully involved with developing data-driven technologies and "be empowered to decide whether and how AI is deployed in the workplace."

Human expertise and clinical judgment "are the only ways to ensure safe, effective, and equitable nursing care…No patient should be a guinea pig and no nurse should be replaced by a robot," added Cathy Kennedy, a registered nurse at Kaiser Permanente Roseville Medical Center and a president of CNA, in the statement.

A similar but smaller protest occurred at HCA Healthcare in Nashville, WZTV reported.

Among nurses' specific complaints to the press, as reported by Becker's Health IT:

  • Concern about an AI chatbot that patients can talk to instead of a nurse, citing worries about what might happen if a patient has a heart attack and the chatbot doesn't understand because of the language in which it was trained and sends them to a pharmacy instead of a hospital.
  • An AI tool that monitors patient vital signs often issues false alarms.
  • A platform in the Epic electronic health record determines nurse staffing based on real-time charting. One nurse said that if nurses don't log charts immediately, the next shift could be short-staffed. Other nurses said it might not account for time-sensitive work that can't be measured, such as educating family members or preparing for chemotherapy treatments before a patient arrives.
  • A monitor in place at 21 Northern California hospitals that analyzes electronic record data to detect patient deterioration can produce erroneous notifications or miss patients who are declining.

In response, a Kaiser Permanente-issued statement, reported on by Becker's Health IT, said its technologies empower nurses and enable them to work more effectively, "resulting in improved patient outcomes and nurse satisfaction." AI tools "don't make medical decisions - our physicians and care teams are always at the center of decision-making with our patients," the statement said. "We believe that AI may be able to help our physicians and employees and enhance our members' experience."

Healthcare IT News reached out to Kaiser nationally and in California to ask how nurses were involved in AI use in patient care but didn't hear back. A report on the future of nursing, released by Florida Atlantic University's Christine E. Lynn College of Nursing and Cross Country Healthcare, a technology and workforce advisory firm, showed that most nurses are uncomfortable using AI and have concerns about its safety and efficacy. The survey of over 1,100 nursing professionals and students found nurses' top concerns about using AI related to factors like a lack of empathy and patient connection, job replacement, data security and regulation of emerging technologies.

"Engaging nurses in AI implementation is vital to acceptance and adoption," the report authors said. "By soliciting and incorporating nurses' feedback, healthcare organizations can tailor AI solutions to address specific pain points and enhance the nursing experience."

Roles for AI in health care

Meanwhile, health care institutions continue to implement AI technologies for various purposes.

  • Phoebe Physician Group in Albany, Ga., uses an AI tool to analyze years of patient visit data and predict the probability of patients not showing up for appointments. If it finds a person unlikely to show, it automatically creates an adjacent appointment slot for someone else, Healthcare IT News reported. The system led to an average increase of 168 patient encounters per week.
  • Cedars-Sinai Medical Center in Los Angeles launched a 24/7 virtual care app that uses AI to help evaluate patients. Through the app, an AI chatbot collects health details and symptoms, then generates a summary for clinicians, Becker's Health IT reported.
  • Researchers at Mass General Brigham in Boston published a study earlier this year showing that AI can automatically extract information about social determinants of health from clinician notes in electronic health records, potentially helping connect patients to resources such as social work support. The large language models caught nearly 94% of patients with relevant social determinants of health information.

The general public, too, is becoming more accepting of AI, according to a recent survey by the market research firm Ipsos and Axios, a news site. Overall, 72% of people surveyed said they would be at least somewhat comfortable with their provider using AI to schedule appointments, and 53% said they would be comfortable with their provider using AI to deliver test results. Some 45-47% said they had no problem allowing their doctors to use AI for assistance in triaging or asking basic health questions, or assisting a physician in diagnosis. However, only 16% said they would be comfortable with AI making a diagnosis alone.

Analysis: Where is the disconnect with nurses?

Did the nurses' concerns have to escalate to the level of a protest? Without a behind-the-scenes peek at how Kaiser Permanente - an early adopter of AI in health care - has rolled out AI tools, or how they have previously addressed any nurses' concerns about the technologies, it's challenging to know. However, a spokesperson for UC Davis Medical Center told the San Francisco Standard that one nurse's concerns about a patient monitor issuing false alarms were akin to "fear-mongering." If that indicates how nurses were treated, it's unsurprising they organized a protest to feel heard.

I have covered plenty of medical and pharmacy conferences over the past decade for assorted trade magazines, including sessions where health care providers discuss their implementation of new technologies. They almost always issue a list of best practices that include the following:

  • When evaluating new technologies for implementation, always get buy-in from health system leadership.
  • The implementation team should be multidisciplinary, involving representatives from all stakeholders such as doctors, nurses, pharmacists, IT professionals, etc.
  • New technologies should be beta tested in one department or unit (or in a multi-hospital system, at one hospital) to work out all kinks before deploying them more widely.
  • Identify one or more champions for each group such as nurses, doctors, etc. who can introduce their colleagues to the new technology and address concerns.
  • Allow time for training in the use of each new technology.
  • Once a technology is rolled out, review its performance periodically and address any issues.

Resources