1. Home
  2. Home Page Featured
  3. Artificial intelligence in nursing
Home Page FeaturedPractice MattersWeb Exclusives
AI

Artificial intelligence in nursing

Share
By: By Brian J. Douthit, PhD, RN-BC; Ryan J. Shaw, PhD, RN; Kay S. Lytle, DNP, RN-BC, NEA-BC, CPHIMS, FHIMSS; Rachel L. Richesson, PhD, MPH; and Michael P. Cary, Jr., PhD, RN

Practical implementation in clinical settings.

Editor’s note: This is an early release of a web exclusive article for the January 2022 issue of American Nurse Journal.

Takeaways:

  • Artificial intelligence (AI) in healthcare isn’t new. In fact, it’s currently used in many ways that are relevant to nurses.
  • Nurses should be involved in the conceptualization, development, and implementation of AI, especially when it impacts nursing practice.

Artificial intelligence (AI) comprises many healthcare technologies transforming nurses’ roles and enhancing patient care. In healthcare, AI typically refers to the ability of computers to independently convert data into knowledge to guide decisions or autonomous actions. However, precisely defining AI can be challenging because of its breadth of applications, including risk prediction algorithms, robots, and speech recognition—all of which augment nursing practice and are on a fast track to changing healthcare as a whole.

Nursing AI tools include clinical decision support, mobile health and sensor-based technologies, and voice assistants and robotics. (For an introduction to AI, including definitions of machine learning, deep learning, and other related terms, visit myamericannurse.com/how-artificial-intelligence-is-transforming-the-future-of-nursing.)

Clinical decision support

Clinical decision support tools (including alerts in the electronic health record [EHR], clinical practice guidelines, order sets, reports, and dashboards) enhance nurses’ ability to make clinical decisions. They may supply the end user with information or provide actionable options based on the data. Clinical decision support also may be integrated into other tools, including mobile health applications beyond the EHR. When coupled with AI, clinical decision support can offer predictions and suggestions with accuracy and specificity beyond human capacity. AI-based clinical decision support includes automatically generated nursing diagnoses, fall risk prediction, and guided decision trees to prevent catheter-associated urinary tract infections.

The concepts behind these tools aren’t new. Fall risk prediction, for instance, involves regular assessment and fall precaution implementation. However, manual risk calculation is time-consuming and vulnerable to human error, leading to inaccurate predictions. AI offers three advantages over traditional methods:

  • the ability to quickly consider large volumes of data in the risk prediction
  • increased intervention specificity (accurately flagging patients most at-risk)
  • automated adjustments in variable selection and calculation.

AI accurately identifies at-risk patients by considering more diverse patient information from the EHR and other information sources.

To ensure AI-based clinical decision support tools complement nursing workflows and benefit care delivery and patient outcomes, nurses must participate in their development and deployment. AI’s potential in decision support includes helping nurses advocate for patients and identify care gaps and challenges.

Mobile health and sensor-based technologies

The COVID-19 pandemic transformed patient care delivery, including an increased need to retrieve data from patients remotely and between clinic visits. Mobile health (mHealth) and sensor-based technologies provide opportunities to reshape a nurse’s ability to deliver care and monitor patients, particularly with limited resources and staffing. These technologies are particularly useful for managing chronic illnesses, which consume over 75% of healthcare spending in the United States, according to the Centers for Medicare & Medicaid Services. 

Mobile health technologies (smartphones, smartphone apps, and wearable technologies) help manage chronic illnesses by receiving and sending data directly between patients and providers, creating a comprehensive picture of the dynamic state of a patient’s health in their everyday environments. Sensor-based technologies, when placed in the home or hospital environment and used in combination, help nurses compose text and multimedia messages (for sharing photos and videos), measure body movement, and collect weight, movement, and environmental (temperature, light, sound, air quality) data

These technologies can be used across the care continuum, following patients as they transition from inpatient to outpatient care. Smartphones, ubiquitous across socioeconomic, racial, and ethnic backgrounds, can help improve access to care. In addition, wearable sensors that monitor activity, sleep, and heart rate and rhythm are increasingly affordable. These tools and their applications allow nurses to send and receive data from individual patients in near real-time and on a population health level for scalable noninvasive diagnosis and monitoring.

Mobile health technologies generate large amounts of continuous data, requiring software tools to aggregate and analyze for actionable insights. For example, if patients with diabetes use mobile health devices to monitor their condition, clinicians can analyze the data for pattern recognition. This allows for feedback loops that prompt patients to change behaviors based on data trends. The analyses also can identify patients who may need additional care and self-management support. During clinic visits, nurses can use the data to illustrate the day-to-day behaviors and physiologic changes their patients experience.

Voice assistants and robotics

Voice assistants (think Amazon Alexa and Google Assistant) may have a future in EHR applications, collecting patient data in the home and delivering interventions to augment care. Imagine a scenario in which a nurse uses Alexa to remind older adults to take their medications and measure their blood pressure. Alexa then records patient data in the EHR for the nurse to review. For older adults and patients with certain disabilities, such as poor eyesight, these tools may be especially useful given their voice-based interaction. The benefit of voice assistants depends on nurse involvement in technology selection and its application in practice and patient care.

As robotics technology advances, it’s being used to provide care companions and create remote-controlled tools, such as telepresence robots (where a nurse can drive a wheeled robot using a voice and video application), to deliver care. Hospitals increasingly use telepresence robots to augment face-to-face patient care.

During the COVID-19 pandemic, nurses have frequently interacted remotely with patients via voice assistants and robots to reduce personal protective equipment use and repeated exposure to the virus. These technologies also may reduce the time nurses spend per visit on data collection and documentation. However, they’re limited to video and voice interaction. Ongoing research by Li and colleagues may lead to robots with arms that nurses can drive remotely to manipulate items in the environment, such as pressing infusion pump buttons and assisting with feeding and medication delivery.

Implementing AI in nursing

Researchers have been leveraging AI for several decades, but its use in practice remains relatively new. When nurses implement AI, such as clinical decision tools, they can process large amounts of data quickly to identify risks, recommend interventions, and streamline workflow. However, for AI to truly transform nursing practice, limitations must be addressed with input from nurses.

Understanding how AI functions compared to traditional tools can help nurses choose the best option based on the specific care situation. When assessing pressure injury or fall risk, for example, AI tools consider risk over time and may change how they perform calculations to improve accuracy. Traditional risk assessment tools consider a limited number of variables at a single point in time, and they can’t account for individual variation. In other words, AI tools “learn” to be more specific and accurate, identifying at-risk patients who would otherwise be overlooked.

AI tools also can analyze more data points than traditional tools, including data not readily available in the EHR, such as purchased consumer and publicly reported data. This additional information, when combined with a subset of EHR data, can result in more robust likelihood or risk analysis than traditional point-in-time risk assessment tools. One such example is pharmaceutical research. McKinsey & Company partnered with healthcare systems to combine their data with longitudinal EHR data to develop oncology medication. Nursing science also would benefit from integrating multiple data sources (including patient-reported outcomes, patient preferences, and socio-economic considerations) with the EHR.

To help make the best decisions about AI tool use, nurses should engage in all project development phases, from defining the problem to be solved by the tool to evaluating its impact. AI application training should focus on why AI is needed, what drives the recommendations, and how it can inform patient care. Training on workflow within the EHR is secondary to this more fundamental understanding of AI. Nurses who understand the differences in using AI clinical decision support, along with AI identified risk factors, are critical to the development and use of reliable solutions. AI implementation in nursing isn’t a perfect science. Success requires careful consideration of the most useful tool, engagement with the nurses who will actually use the tool, and nurse involvement in its implementation and evaluation. (See AI: Nursing impact.)

AI: Nursing impact

Researchers continue to investigate artificial intelligence (AI) and nursing. Here are a few examples:

Bose E, Maganti S, Bowles KH, Brueshoff BL, Monsen KA. Machine learning methods for identifying critical data elements in nursing documentation. Nur Res. 2019;68(1):65-72. doi:10.1097/NNR.0000000000000315

Impact on nursing: AI can help detect which patient features are most important in public health applications, allowing for more focused preventive interventions.

Li Z, Moran P, Dong Q, Shaw RJ, Hauser K. Development of a tele-nursing mobile manipulator for remote care-giving in quarantine areas. Paper presented at: 2017 IEEE International Conference on Robotics and Automation.

Impact on nursing: Robots may aid nursing care tasks in hazardous clinical environments, and they have the potential to automate some tasks.

Park JI, Bliss DZ, Chi C-L, Delaney CW, Westra BL. Knowledge discovery with machine learning for hospital-acquired catheter-associated urinary tract infections. Comput Inform Nurs. 2020;38(1):28-35. doi:10.1097/CIN.0000000000000562

Impact on nursing: Automated notifications may facilitate safe catheter removal and urinary tract infection treatment.

Safavi KC, Khaniyev T, Copenhaver M, et al. Development and validation of a machine learning model to aid discharge processes for inpatient surgical care. JAMA Netw Open. 2019;2(12):e1917221. doi: 10.1001/jamanetworkopen.2019.17221

Impact on nursing: Early detection of which patients may experience complicated discharges after surgery may help focus care.

Wang L, Xue Z, Ezeana CF, et al. Preventing inpatient falls with injuries using integrative machine learning prediction: A cohort study. npj Digit. 2019;2(127). doi:10.1038/s41746-019-0200-3

Impact on nursing: AI may more accurately predict fall risk without manual calculation and provide automatic warning systems.

Challenges

Although AI offers promising solutions to nursing, it’s not without its drawbacks. For example, just because you can use an AI tool, doesn’t mean you should. Many traditional tools actually perform similarly (or outperform) their AI counterparts depending on the application, such as in predicting mortality of older patients who have undergone hip fracture treatment. Data quality and sources, as well as modeling validation have shown mixed results. Novel robotics can be met with resistance. Cultural change is always a factor when introducing something new, but robotics may be misunderstood or considered invasive if not implemented with caution. Healthcare professionals also may be fearful that AI will result in job loss. This may be true in the future, but current tools and those under development don’t replace human jobs; they’re intended as enhancements. In addition, many have concerns about confidentiality and privacy related to AI use. As with technology that handles sensitive information, risks exist. However, with careful planning and implementation, these risks can be mitigated.

Transforming care

AI is transforming healthcare and nurses’ roles in care delivery. As we face a global pandemic, the development and implementation of AI-driven technologies aim to address the unique issues we’re facing, such as disease presence in asymptomatic patients. Nurses must contribute to the advancement of AI to ensure development that advances the nursing role and focuses on providing person-centered care.

Brian J. Douthit is a postdoctoral associate at the United States Department of Veterans Affairs and Vanderbilt University in Nashville, Tennessee. Ryan J. Shaw is an associate professor at Duke University School of Nursing in Durham, North Carolina. Kay S. Lytle is chief nursing information officer at Duke University Health System. Rachel L. Richesson is a professor at the University of Michigan Medical School in Ann Arbor. Michael P. Cary, Jr., is an associate professor at Duke University School of Nursing and a member of the Duke University Center for the Study of Aging and Human Development.

References

Ajani R, Chatterjee A, Talwai A, Zhang J. How a pharma company applied machine learning to patient data. Harvard Business Review. October 25, 2018. hbr.org/2018/10/how-a-pharma-company-applied-machine-learning-to-patient-data

Cary MP Jr, Zhuang F, Draelos RL, et al. Machine learning algorithms to predict mortality and allocate palliative care for older patients with hip fracture. J Am Med Dir Assoc. 2021;22(2):291-6. doi:10.1016/j.jamda.2020.09.025

Centers for Medicare & Medicaid Services. National health expenditures 2017 highlights. cms.gov/research-statistics-data-and-systems/statistics-trends-and-reports/nationalhealthexpenddata/downloads/highlights.pdf.

Douthit BJ, Hu X, Richesson RL, Kim H, Cary MP. How artificial intelligence is transforming the future of nursing. American Nurse J. 2020;15(9):100-2.

Li Z, Moran P, Dong Q, Shaw RJ, Hauser K. Development of a tele-nursing mobile manipulator for remote care-giving in quarantine areas. Paper presented at: 2017 IEEE International Conference on Robotics and Automation.

Woo M, Alhanti B, Lusk S, et al. Evaluation of ML-based clinical decision support tool to replace an existing tool in an academic health system: Lessons learned. J Pers Med. 2020;10(3):104. doi:10.3390/jpm10030104

Yang Q, Hatch D, Crowley MJ, et al. Digital phenotyping self-monitoring behaviors for individuals with type 2 diabetes mellitus: Observational study using latent class growth analysis. JMIR Mhealth Uhealth. 2020;8(6):e17730. doi:10.2196/17730

Key words: Artificial intelligence, Clinical decision support, Mobile health, Sensors, Voice assistants, Robotics

1 Comment. Leave new

Leave a Reply

Your email address will not be published.

Fill out this field
Fill out this field
Please enter a valid email address.

Get your free subscription to NurseLine!

NurseLine Newsletter

  • Hidden

*By submitting your e-mail, you are opting in to receiving information from Healthcom Media and Affiliates. The details, including your email address/mobile number, may be used to keep you informed about future products and services.

Recent Posts