The Impact of Artificial Intelligence on Nursing Services Across Care Settings
How hospital, outpatient, and long-term care might change via AI technologies
Nurses are the largest profession in health care services. While the focus in AI is often on how it will shape the physician’s role and work in the future, AI will have a marked impact on nursing care delivery across outpatient, acute care, and long-term and post-acute care. Given how nursing care is ubiquitous in health care services, these changes will also have a major impact on how patients will experience care. Nursing care is the proverbial” sausage making” of health care services, so it is a very large target for AI companies and innovations—especially given the prevailing beliefs around staffing shortages.
The Health Resources and Services Administration (HRSA) suggests that about 3.5 million registered nurses (RNs) are actively working and their average hourly wage in 2024 was $41.38 according to the Bureau of Labor Statistics. With benefits and taxes of 15% and assuming full-time employment of 40 hours per week, registered nurses represent $346.4 billion in healthcare spending annually from payroll and benefits costs alone. This makes efficiency and clinical effectiveness critical in how nursing care workflows are designed.
I have been working closely with nurses and software-based technologies and algorithms over the last five years. Understanding when something is actually useful versus “a cute idea” is very important. Useful software features and algorithms serve to make workflows, patient triage, and analysis easier, while “cute” features are often designed by non-nurses based on an idea of what might be helpful. These features or functions look appealing and sound cool, but are actually distractions. Technological innovation in health care services is riddled with examples of ideas that sound good on paper, but fail when meeting reality.
So, where will AI and ML fall on the scale of useful to distracting-cuteness?
The Nursing Profession
The nursing profession encompasses multiple levels of caregivers working across diverse healthcare settings. At its foundation are Certified Nursing Assistants (CNAs) who provide basic patient care including assistance with activities of daily living, vital signs monitoring, and mobility support. Licensed Practical/Vocational Nurses (LPNs/LVNs) perform more advanced technical care like medication administration and wound care under RN supervision. Registered Nurses (RNs) represent the largest segment of the nursing workforce and provide comprehensive patient care including assessment, care planning, care coordination, patient education, and complex clinical interventions.
Nurses perform very mixed roles across the continuum of care. Their daily activities involve data collection, data interpretation and analysis, patient condition monitoring, care coordination, clinical judgement calls, and hands-on-care.
The emergence of artificial intelligence technologies promises to transform how nurses at all levels deliver care across outpatient clinics, hospitals, and long-term care facilities. This paper is aided by a comprehensive review published in the Journal of Medical Internet Research examined how AI applications are being developed and implemented to support nursing practice. This review, however, is only based on use cases published in academic clinical journals and does not encompass newer innovations, which I will add from my own review.
Summary of Current Research
In 2021, researchers in Germany conducted a very comprehensive rapid review analyzing 292 publications on AI applications in nursing care between 2005-2020. Their findings revealed that most AI development has focused on acute care hospital settings (29.8% of studies), followed by a focus on independent living at home (22.6%), and skilled nursing facilities (11.3%). The majority of applications utilized machine learning approaches (78.1%) rather than rule-based1 (nested “If-Then” statements) systems (11.6%) or hybrid systems (1.7%). This paper was published before the contemporary large-language model revolution, so it does not focus on that use case, but this article will cover it.
These innovative systems focused on the following key issues:
Image and signal processing for tracking patient activity and health status
Care coordination and communication
Fall detection and prevention
Nurse scheduling and staffing optimization
Pressure ulcer prevention and management
Social integration support for patients
Categories of AI Applications in Nursing
Clinical Decision Support
AI systems are being developed to support nursing assessment and clinical decision-making. For example, machine learning algorithms can analyze electronic health record data to predict patients at high risk for complications like pressure ulcers or falls. This allows nurses to implement preventive interventions proactively. In ICU settings, AI helps filter and prioritize patient monitoring alarms to reduce alarm fatigue while ensuring critical alerts are addressed promptly. In other circumstances, CDS systems can help clinicians predict when a patient is likely to deteriorate and require intensive care at the beginning of admission.
In the outpatient setting or at the end of inpatient care, admission and readmission risk, respectively, is also an area of work in this space. Imagine being able to accurately and precisely pinpoint a patient who is going to be admitted to the hospital again in the next 60 days. What could we do with that information?
While these use cases are interesting and important for clinical care, payment models may incentivize organizations to focus first on AI technologies that reduce cost or improve efficiency. More on that tension here.
Workflow Optimization
AI applications can assist with staffing, scheduling, and workflow management. Intelligent scheduling systems consider factors like patient acuity, nurse competencies, and workload distribution to create optimal staffing plans. These plans can also be designed to minimize nursing cost for hospital facilities. Documentation assistance tools using natural language processing help streamline charting requirements. Documentation assistance tools are important to note as they will likely become very common in hospitals to alleviate workload on nurses. Positively, this may allow nurses to spend more time on direct patient care, but it may also allow facilities to increase the number of patients per nurse on a shift.
Remote Monitoring
For home care and long-term care settings, AI-enabled remote monitoring systems track patient activity patterns, vital signs, and potential safety concerns. Computer vision and sensor technologies can detect falls or changes in mobility that may indicate declining health status. This supports earlier intervention while enabling more independent living for older adults. In the United States, the looming aging and senior care challenges will require technology to manage. Japan is an early leader in this space and you can read more about it here.
Care Coordination
AI facilitates better care coordination by analyzing clinical documentation to identify care needs, track interventions, and support handoff communication between care teams. Machine learning models help predict which patients may need additional support services or are at risk for hospital readmission. AI technologies for care coordination can automate repetitive administrative tasks allowing nurses time to engage with patients and manage higher case loads with the same quality of care.
Patient Education and Engagement
Intelligent virtual assistants and chatbots provide 24/7 patient education and support. These AI-based and digital tools can answer common questions, provide medication reminders, and alert nurses when patients need additional assistance. This extends the reach of nursing staff while improving patient engagement. For chronic disease management, discharge planning, and post-operative care, interactive patient care guides can have a major impact on readmissions and condition management.
Outcomes and Impact
Studies, as analyzed in the JMIR review article, examining real-world implementation of AI in nursing practice have demonstrated several promising outcomes:
Clinical Outcomes
Reduced hospital-acquired pressure ulcer rates through AI-enabled risk prediction
Decreased falls through enhanced monitoring and early intervention
Improved identification of clinical deterioration enabling faster response
Better management of chronic conditions through continuous monitoring
Operational Efficiency
Reduced time spent on administrative tasks and documentation
More optimal nurse staffing and scheduling
Decreased false alarms and alert fatigue
Enhanced care coordination and handoff communication through digital document exchange and patient information sharing
Patient Experience
Increased independence and safety for home-based patients
Better access to education and self-management support
More consistent monitoring and follow-up
Improved social engagement for long-term care residents
However, researchers note that more evidence is still needed on the real-world effectiveness of AI applications, particularly outside of controlled study environments. Important considerations include:
Need for nursing-specific outcome measures that reflect care quality
Integration of AI tools into existing workflows and systems (e.g., innovation may be hindered by EHR companies monopolizing this space)
Data privacy and security requirements
Staff training and technology acceptance (e.g., will clinicians trust new systems and how will mistakes affect future trust)
Ethical implications of AI-assisted care delivery
LLM-based Applications in Nursing Care
The recent study by Wan et al. (2024) demonstrates a specific use of large language models (LLMs) in nursing practice. Through their randomized controlled trial of a nurse-LLM collaboration model, they demonstrated that AI can effectively support nursing staff while improving patient satisfaction and reducing workload.
The study's findings revealed notable improvements across several key metrics: higher patient satisfaction (3.91 vs 3.39 in nurse-only group), fewer negative emotional interactions (2.4% vs 7.8%), and stronger empathy ratings (4.14 vs 3.27). Perhaps most significantly, 95% of nurses reported reduced workload and preferred working with the LLM system.
While this study focused on outpatient reception, it suggests broader applications across nursing settings. LLMs could assist with documentation in inpatient care, care plan development in long-term facilities, and patient education in ambulatory settings. However, the research also highlights critical implementation factors: the need for site-specific customization (90.5% of required knowledge was site-specific), maintaining human oversight, and robust safety protocols and continuous performance monitoring.
Ethical Concerns and the Paradox of Innovation
In health, everyone is rightly preoccupied with the safety, efficacy, and risk of bias associated with these new technologies. Mistakes resulting from AI technologies can result in death or serious harm. While we should have regulations aimed at validating technologies and mitigating risk, we should keep the following innovation paradox in mind.
I first learned of this AI innovation paradox when I was at an AI in health event in September 2023 and a leading AI scientist2 made a comment that has stuck with me for years. He presented a hypothetical situation using self-driving cars as an example.
“If self-driving cars result in 400,000 accidents and 20,000 deaths, we might reject this technology upon learning of the harm. But, in reality, cars operated by humans result in millions of accidents and 42, 514 deaths each year as of 2022.
The self-driving cars, like other AI-technologies far outperform humans from a safety perspective, but the public and certain professions are very sensitive to errors from new technologies and can quickly reject them from a single anecdotal story.
Conclusions
Looking ahead, AI is poised to augment rather than replace nursing roles. By handling routine tasks and providing decision support, AI allows nurses to focus more on the elements of care requiring human judgment, emotional intelligence, and hands-on skills. Success will require thoughtful implementation that considers workflow impact, staff engagement, and most importantly, enhancement of the nurse-patient relationship.
As AI continues to evolve, ongoing research will help identify the most valuable applications across care settings. The goal should be leveraging AI to support nurses in delivering safer, more efficient, and more personalized patient care while maintaining the essential human elements of the nursing profession.
This measured integration of AI into nursing practice has the potential to address many current challenges in healthcare delivery while elevating the nursing role to focus on the highest value activities that improve patient outcomes. Continued collaboration between nursing professionals, technology developers, and researchers will be essential to realize this potential.
It is important to note that while this review focuses on ML applications, “If-Then” based systems are still likely to grow in the nursing workflow space to aid in non-AI-based automation as much of health care services is still very manual.
His name escapes me and I wish I could provide credit where credit is due.