According to Lisiane Pruinelli, PhD, MS, RN, FAMIA, associate professor at the University of Florida College of Nursing, if nurses had wider access to AI, they would be more satisfied, experience less burnout, and stay in their jobs longer.
“Many nurses leave the profession because of the hard work and physical exhaustion that comes with the job,” says Pruinelli, who has worked extensively on implementing machine learning and intelligent systems.
Recently, she co-authored an article on ChatGPT, an artificial intelligence chatbot, concluding that the technology shows promise for nursing education but there is still room for improvement.
The article, “ Transforming Nursing Education with ChatGPT Effects and Generative AI: A Discussion Paper ,” was published in Nurse Education in Practice and aims to demonstrate the technology’s potential to assist nurses with clinical documentation.
UF is embracing AI, launching an artificial intelligence initiative aimed at setting the standard for AI universities. University officials want to prepare students to work with AI in all disciplines taught at the university. As part of the university’s AI initiative, the School of Nursing has several prominent AI-focused researchers, including Pruinelli.
In Pruinelli’s article, the researchers presented ChatGPT with prepared patient case studies along with questions soliciting nursing diagnoses, nursing interventions, and nursing outcomes.
“We were trying to see if ChatGPT could do that kind of information gathering and provide a diagnosis,” Pruinelli said.
Although the program returned appropriate nursing diagnoses, most did not match the requested North American Nursing Diagnosis Association International (NANDA-I) classification, which uses standardized nursing terminology and diagnoses.
However, along with appropriate answers, there were also inappropriate answers and diagnoses not found in NANDA-I, and the diagnosis codes were incorrect.
“ChatGPT got a little lost and gave out incorrect information, but that doesn’t mean that training ChatGPT will give you better responses,” Pruinelli said. “The big problem is that ChatGPT is using publicly available information, not scientifically ‘proven’ evidence.”
Although chatbots have access to a lot of information, they don’t always return the most accurate information.
ChatGPT stands for Chat Generative Pre-training Transformer, which is based on generative artificial intelligence and creates responses to questions in a human-like conversation. Users can ask additional questions. The chatbot takes information from billions of parameters and millions of websites to create the response. Through continuous human interactions, including user feedback, the chatbot learns and improves.
Streamlining clinical documentation and administrative workflows will allow nurses to spend more time with patients, Pruinelli said, citing research showing that up to 41% of U.S. nurses’ time is spent on documentation.
Pruinelli believes generative artificial intelligence will eventually help with everyday tasks, but won’t replace nurses.
“The idea is not to replace our work, but to use technology to help us or as a decision aid,” she said. “It doesn’t matter how accurate ChatGPT or any other technology is. We need to make sure it’s actually the right intervention, the right diagnosis, the right recommendation for that patient, that problem, at that time. The endpoint is always us.”
Pruinelli recently published a blog post in the American Nursing journal Off the Charts discussing the nursing profession’s openness to AI.
She argues that nurses have generally been at the forefront of change, driving new technologies, from paperless documentation to smart pumps, and that adopting ethical and safe technology allows them to spend more one-on-one time with patients.
Pruinelli envisions a future where AI technology can help restock supplies at nursing stations or lift and transport patients.
“Instead of transporting the patient, the nurse can be at the patient’s bedside and use technology to coordinate these actions. Patients appreciate the personal attention and concern of a nurse at the bedside.”
Medication administration, both in hospital and at home, may soon be left to smart technology.
“At home, they need more help. Caregivers can be tired. They can be overwhelmed with how many medications to take and when to take them,” Pruinelli says.
But she also warned that AI medical technologies will need oversight as they develop.
“There is a clear pathway to making drugs safe before they are put on the market, but there is not the same pathway to making sure technologies are safe,” she says. “My question is, how can certain technologies be tested before they are introduced to see if they actually improve patient outcomes?”
