Artificial intelligence in nursing

Can AI ever be called a nurse? Are there parts of a RN’s job that can be delegated to artificial intelligence? Will it improve patient outcomes? Will it create more medical errors and challenges?  Let’s dive into AI and nursing practice. 

When you think of AI, you probably think of a robot that can do just about anything a human does – but what about artificial intelligence in nursing? The goal isn’t to replace nurses but to complement them by automating tasks and streamlining workflows, so nurses have more time to devote to patients or other important matters. To better understand how AI fits into nursing, and how it has evolved in healthcare, let’s first unpack what AI actually is.

Artificial intelligence is a computer system or machine that is capable of learning, recognizing patterns, making decisions and understanding language. Going beyond simple computations, AI is more advanced (hence the term “intelligence”) and able to acquire more knowledge as it encounters more data, and improve itself, just as a human would.

Artificial intelligence was introduced in healthcare in the 1970s, when the computer system MYCIN was created. It was able to provide a list of pathogens and corresponding antibiotic treatments that were tailored to a patient’s weight. Artificial intelligence and nursing made its debut in 1985, with uses in nurse scheduling and clinical decision support. From there, AI in healthcare progressed:

●      1986: DXplain was created to generate potential diagnoses based on symptoms.

●      2007: IBM Watson, a question-and-answer system, was developed and laid the foundation for future healthcare AI applications.

●      2015: A chat-based AI program, PharmBot, was created for medication education for pediatric patients.

●      2016: Google created an AI tool, DeepMind, that partnered with the NHS in the U.K. to detect eye diseases for over 50 conditions with 94% accuracy.

●      2017: Mandy was created to streamline patient intake workflow.

●      2017: CardioAI was cleared for clinical use to analyze cardiac MRI scans, computing ejection fraction in seconds.

●      2017: DeepScribe, a medical dictation microphone using AI, was founded.

The short answer is yes. From nurse scheduling models to CAUTI (catheter-associated urinary tract infection) and fall-risk predictors, artificial intelligence in nursing practice is an active part of healthcare today. It just may not always appear in the futuristic way you may envision. For example, you may be familiar with clinical decision trees – many of these are now generated with AI to make predictions and suggest interventions. 

AI is used in patient monitoring by means of computerized models to track vital signs and analyze trends in real time. It can detect changes in clinical status early, alerting nurses and healthcare techs to intervene. There are also predictive analysis AI tools that identify patients at risk for complications, which helps prompt the nurse to implement preventive measures.

Along with the rise of telemedicine, AI plays an important role in virtual nursing. AI tools and platforms, such as Care.AI, enable nurses to monitor patients from a different location — whether from a separate location in the hospital or from home. Additionally, artificial intelligence chatbots are performing some of the more administrative tasks that fall into a nurse’s responsibility, like scheduling appointments. This can free up more time for nurses to perform direct patient care. 

Being that AI consists of computer-based models and machines, its use in healthcare generates patient-sensitive data. Thus, a common concern regarding artificial intelligence in nursing is privacy and security. There may also be inadvertent algorithm biases, which may create disparities in medical diagnosis or treatment. However, a bias in AI is a reflection of bias in data and the model design, which can be improved through ongoing testing and updating.

Another question that arises in this conversation is: When AI makes a mistake, who is held accountable? Transparency and responsibility are crucial to fostering and maintaining trust. There may also be legal considerations, as determining liability becomes complex when a computer system or model, rather than a human, makes a decision that results in significant harm or negligence. Some of the unease may come from the fact that many of these questions remain unanswered, and the conversation is ongoing. AI in nursing is still unfolding, there’s not a level of certainty that could alleviate anxiety among all nurses, providers and patients. 

As with anything new, there will be innovators, early adopters and those who are simply late to the game. When it comes to nursing, some nurses will be eager to implement new tools and systems immediately. Others may be more cautious and reluctant, wanting to see more proven results. As a result, there may be some inconsistency in AI implementation, which may create challenges of standardization and collaboration. 

Simulation is a part of nursing education. It can provide context and opportunities for nursing students to practice clinical skills and decision-making in a safe, controlled environment, prior to implementing skills in real clinical situations. Where does AI come in? An AI robot can interact with the nursing student, simulating a real-life human experience, in a way that a manikin cannot.

Virtual and augmented reality can also be enhanced with artificial intelligence, creating immersive experiences tailored to potentially strengthen each nursing student’s competencies. For example, if a student has not yet encountered an obstetric emergency in their clinical rotation, the clinical instructor could utilize AI to create a scenario for the student to practice the necessary skills and critically think through nursing decisions. 

Nurses already need to have some level of technical literacy due to electronic medical records, and the integration of artificial intelligence in nursing will further amplify this need.

Other skills may include:

●      Adaptability: Continue evolving practices to stay updated with new AI tools

●      Decision-making: Pair personal expertise with AI-generated predictions

●      Problem-solving: Troubleshoot AI issues when technology malfunctions

●      Critical thinking: Just because AI says it, doesn’t make it 100% applicable, 100% of the time

●      Data interpretation: Review and analyze AI results

As the adoption of AI increases, nurses will need to be familiar with tools that patients may use at home for self-monitoring. Things like wearables (e.g., smartwatches) and health apps used by patients may call for the nurse to answer questions, interpret data or show patients how to use the technology properly (even if this isn’t in their job descriptions). Some hospitals are already using these self-monitoring tools — with patients desiring to enhance their ability to self-monitor, and nurses wanting to ensure accurate data that can be easily integrated into their workflow.