As we dive into the realm of emerging technologies in healthcare, we find artificial intelligence (AI) defined as the aptitude exhibited by smart machines broken down into perceiving, thinking, planning, learning, and the ability to manipulate objects (NITI Aayog, 2018), its applications and benefits to nurses in care delivery environments are still vague. Understanding how AI and its use can enhance nurses’ decision making, by supporting critical thinking and positively impacting the nursing process is necessary. The need for nurses’ comprehension of the foundations of AI and the symbiotic nature of it with nursing practice is essential with its increased use in practice in today’s value-based care environment. It's to nursing's advantage as health systems, providers and payers look to health IT innovations, including AI, to disrupt the business model of healthcare and the products and solutions offered to meet the Quadruple Aim.
The foundation of AI can seem overwhelming and wildly complicated. To break down the intricacy of AI, nurses can begin to understand that at the technology’s core is algorithms. Algorithms are sequential instructions that ensure particular task completion. They are a set of well-constructed rules given to an AI program to help it learn on its own. When many algorithms are put together and layered in applications, they become the backbone of AI (Polson & Scott, 2018). While AI is machine-driven, it is important to note that it is assistive. The transforming technology outputs suggestions requiring human judgment to apply recommendations, to be helpful and strike a balance with data-manipulated science and decision making. In healthcare it is “Clinical Intelligence”: machine algorithms designed for diagnostic and treatment processes utilized in the appropriate use cases for everyone (patients, health professionals and payors) as an extension of care for the right treatment, to the right person at the right time (Health Information and Management Systems Society, 2018).
One type of AI readily applied in healthcare today is predictive analytics and machine learning (ML). Predictive analytics is mathematical computations that analyze historical data from multiple sources to predict future events. It is a machine approach to refine those data, using knowledge to extract hidden value from newly discovered patterns, and dynamically informs data-driven decision-making to know what will happen, when and what to do about it (Carroll & Hofmeister, 2018). Multiple industries use predictive analytics and ML in internet search engines, app streaming services– both audio and video – and in social media. In nursing, EHR clinical decision support tools, radiology image recognition, and disease progression prediction are applications of this type of AI.
Another AI program is natural language processing (NLP) coupled with Automated Speech Recognition (ASR), which help computers better understand and process human (unstructured) languages, to move smart machines closer to a human-level understanding of language (Seif, 2018). Combined with predictive analytics and ML, both NLP and ASR enable scientists to develop algorithms for language translation, semantic understanding, and text summarization, making it easier to understand and perform computations on volumes of text with less effort (Seif, 2018). Examples of NLP and ASR are virtual assistants, chatbots, cell phone voice texting/messaging, and nursing applications include extracting EHR text from notes in non-discreet fields, vocal charting, and speech-activated paging devices.
The data used for effective decision making in AI feed programs for algorithms building, stems from many healthcare data sources, include EHRs, medical claims, voice/audio files, images, and workforce and hospital throughput data such as staffing and bed management. Quality of the data for successful AI use is vital as they must be complete, structured, cleaned, unbiased, and more data is best as the volume of robust data sets help machines to learn, predict and process semantics languages better. AI programs have moved from merely searching for results to smart suggestion and recommendations machines (Polson & Scott, 2018), and for nurses’ benefit, they now guide decision-making by enhancing critical thinking and the nursing process which enables patterns and semantics detection that ultimately thinks more like humans and in clinical environments, nurses.
Critical thinking applied in nursing is complex and multifaceted. It is a nonlinear practice that focuses nurses on clinical decision-making (Paul & Heaslip, 1995). Overall AI such as machine-driven prediction and semantic technology aids nurses’ critical thinking process. It helps meet criteria for evaluation, the standards of care nurses strive for by national nursing associations and entities such as the American Nurses Association, the National League of Nursing and The Joint Commission. Critical thinking is a flexible, fluid process, guiding nurses to learn to think and to anticipate what will happen next and what the intervention should be, by asking “What, Why, & How” questions. Machine-driven prediction is a good example since it can help expedite actionable decisions. As well. AI assistants can assist using information synthesis, the reasoning process by which nurses reflect on and analyze personal thoughts, actions, and decisions, and those of others (Paul & Heaslip, 1995) including machine-driven and AI programs. AI used with critical thinking equates to strengthening the guidance of nurses to generate new knowledge, look beyond the obvious, and challenge traditional ways of thinking (Paul & Heaslip, 1995). This fluid process of reasoning is enhanced by AI, with a healthy mix of nursing knowledge and judgment throughout.
Underlying the progression of critical thinking is the Nursing Process, a systematic approach that is used by all nurses to gather data, critically examine and analyze the data, identify patient response, design outcomes, take appropriate action, then evaluate the effectiveness of actions (Johansen & O’Brien, 2016). AI, together with critical thinking and human judgment, can guide all phases of the Nursing Process by heightening the speed and accuracy of evaluation, anticipation, synthesis and knowledge generation (See Figure 1). The Assessment phase of the nursing process is the organized collection and verification of data that is analyzed and communicated to members of the healthcare team and patients. NLP and ASR can aid this step by the use of hands-free input of data and mechanical extraction of patient-specific data from notes and indiscreet fields of the EHR to more accurately determine the patient’s condition. Diagnosis which is the definition of the disease state that allows a nurse to individualize care, Planning to the stage where nurses set care priorities and define patient goals and Intervention, is when the actions necessary for achieving the goals and outcomes to assure they get initiated and completed occurs (Johansen & O’Brien, 2016) All of these stages are made more precise and are expedited by predictive analytics and machine learning. By using data from multiple sources including the EHR, claims, geographic, demographics and socioeconomics ML can make actionable recommendations specific to each patient. Finally, Evaluation helps nurses refine priorities and collect data in an ongoing basis (Johansen & O’Brien, 2016). The data is collected by NLP and ASR through charting, clinical results, and communication to complete the process ensuring the most precise, highest quality of care for each patient throughout the care continuum.
Together algorithms that drive AI programs are aided by machines to meet the specific needs of both patients and nurses. The technology is designed for decision making to provide person-centered, precision care to improve their experience and engagement, and focus on patient outcomes to affect care quality. It directly decreases costs by reallocating menial tasks, improving productivity and reducing waste. Finally, AI used in the clinical environment impacts human factors such as alarm fatigue, knowledge deficiencies, and understaffing, thus improving nurses’ wellbeing. AI assistive outputs used judiciously in the care context enhance critical thinking and the nursing process, leading nurses to serve the Quadruple Aim by adding value through expedited, more precise, enriched decision making.
The views and opinions expressed in this blog or by commenters are those of the author and do not necessarily reflect the official policy or position of HIMSS or its affiliates.
Powered by the HIMSS Foundation and the HIMSS Nursing Informatics Community, the Online Journal of Nursing Informatics is a free, international, peer reviewed publication that is published three times a year and supports all functional areas of nursing informatics.
Carroll, W. & Hofmeister, N. (2018). Predictive Analytics and the Impact on Nursing Care Delivery [Slide Show File]. Retrieved December 12, 2018 from http://365.himss.org/sites/himss365/files/365/handouts/550229529/handout-NI2.pdf
Health Information and Management Systems Society. (December, 2018). HIMSS Insight Series: Clinical Intelligence. [eBook].
Johansen, M. and O'Brien, J. (2016). Decision making in ursing practice: A concept analysis. Nursing Forum, 51: 40-48.
National Institution for Transforming India (NITI) Aayog. 2018. National Strategy for Artificial Intelligence. Retrieved from: http://niti.gov.in/writereaddata/files/document_publication/NationalStrategy-for-AI-Discussion-Paper.pdf
Paul, R. & Heaslip, P. (1995). Critical thinking and intuitive nursing practice. Journal of Advanced Nursing, 22: 40–47.
Polson, N. & Scott, J. (2018). AIQ: How artificial intelligence works and how we can harness its power for a better world (Hardback): Bantam Press.
Seif, G. (October, 2018). An easy introduction to Natural Language Processing: Using computers to understand human language. Retrieved January 20, 2019 from https://towardsdatascience.com/an-easy-introduction-to-natural-language-processing-b1e2801291c1