Doctor creating artificial intelligence interface 3d rendering
HealthDecember 21, 2020

AI in healthcare: What nurse leaders should know

By: Sarah Handzel, BSN, RN
As clinicians seek better, faster ways to effectively manage patient care, artificial intelligence (AI) research in healthcare has taken off as a potential solution. AI is growing rapidly in diagnostics and treatment management1 as more data from each patient becomes available, computer power increases, and AI tools and algorithms improve.

Now, it’s imperative that nurse leaders and nursing staff understand AI and its potential applications with clinical models, critical thinking skills, and evidence-based practice.2 A review in the journal Nursing Management2 details AI frameworks and tools to supplement them, how these models may impact nursing practice, and what nurse leaders should consider before implementing specific AI tools.

The components of AI

In nursing practice specifically, many AI models utilize a specific framework to make decisions that affect care delivery. This framework, known as the data, information, knowledge, wisdom (DIKW) framework, helps nurse informaticists determine how data, information, and knowledge are used in the decision-making process. The DIKW framework has specific definitions that help shape each concept:2

  • Data – Little or no meaning in isolation from other factors
  • Information – Data plus meaning
  • Knowledge – Derived by discovering patterns and relationships between types of information
  • Wisdom – Understanding and internalization of knowledge patterns and relationships

In many cases, AI models are paired with clinical decision support (CDS) applications to help practitioners advance from knowledge to wisdom, which involves exercising judgment. When using AI for clinical purposes, the DIKW framework can be used to categorize CDS tools—this ultimately allows nurse leaders to strategically evaluate relationships between AI and CDS.

But AI and CDS are not the same thing. According to the article, CDS provides clinicians with computer-generated clinical knowledge and patient-related information that is intelligently filtered and presented at appropriate times to enhance patient care.2 AI, on the other hand, may be thought of as the fuel that powers more intelligent CDS systems.

Current AI/CDS-based tools in healthcare

Already, AI/CDS-based clinical tools are used to help support nurse decision making. For example, some of these tools provide predictive analytics for staffing and bed management. In many hospital systems, AI-based sepsis CDS alerts are already integrated into electronic health records (EHR) systems—these alerts use patient vital signs and laboratory results to build AI-based rules predicting patients who are risk for sepsis.2

Other tools, such as the Communicating Narrative Concerns Entered by RNs (CONCERN) CDS system are being rolled out in healthcare systems around the country. CONCERN identifies nursing documentation patterns that are a proxy of a nurse’s concern for patient deterioration. Using this information, the system generates a predictive early warning score for each patient, which is intended to augment clinical expertise. This system flags patients at risk for clinical deterioration using a simple green-yellow-red color system, enabling nurses to quickly determine which patients may require more stringent observation and care.

Questions nurse leaders should ask

According to the article, the success of new AI/CDS-based systems requires careful planning, evaluation, and decision-making on the part of nurse leaders. Such questions can help leaders identify which systems and tools may be more likely to help connect clinical practice with informatics solutions to help nurses provide safer, more effective patient care.

When considering AI/CDS-based systems, nurse leaders may ask:

  • What data does the AI/CDS tool use?
  • How is that data captured, and does the data capture fit into the existing clinical workflow?
  • Does the AI/CDS information take into account the clinical context?
  • Does the information produced make clinical sense and have clinical relevance?
  • Does the AI/CDS-based system actually help solve a clinical problem?
  • Does the CDS fit nursing processes?
  • Is the AI/CDS tool supplementing or taking over clinical decision-making?
  • Is the AI explainable to clinicians?
  • Is the necessary short- and long-term training in place?

The answers to these questions can help leaders organize strategic planning and prioritization of potential AI/CDS-based tools. The answers help give insight into AI systems and can help organizations implement current and future AI/CDS applications.

Sarah Handzel, BSN, RN
Freelance Health and Medical Content Writer, Wolters Kluwer Health
Sarah has over nine years’ experience in various clinical areas, including surgery, endocrinology, family practice, and pharmaceuticals. She began writing professionally in 2016 as a way to use her medical knowledge beyond the bedside to help educate and inform healthcare consumers and providers.
  1. Rice, Michelle. “The Growth of Artificial Intelligence (AI) in Healthcare.” HRS, Health Recovery Solutions, 15 Apr. 2020, healthrecoverysolutions.com/blog/the-growth-of-artificial-intelligence-ai-in-healthcare.
  2. Cato, Kenrick D. PhD, RN, CPHIMS, FAAN, et al. “Transforming Clinical Data into Wisdom: Artificial Intelligence Implications for Nurse Leaders.” NursingCenter, 2020, nursingcenter.com/journalarticle?Article_ID=5689990.
Solutions
Lippincott Solutions
Our best-in-class suite of evidence-based, institutional software can help you to balance clinical and business needs by streamlining workflow, standardizing care, and improving reimbursable patient outcomes.