John Carroll: Inspiring the future of “collaborative intelligence” in nursing.

In the world of an ICU nurse, hospital staff and information flow like high and low tide, in 12-hour shifts. Within each shift, the ICU nurse updates vital signs for high-risk patients every 15 minutes. They do full-body assessment of the body systems (neuro, cardiovascular, respiratory, GI/ GU, skin) for all patients every four hours.

Artificial Intelligence in Nursing

At the peak of the pandemic, many states found more than one-third of their ICU beds occupied by high-risk Covid patients, according to data from The Commonwealth Fund’s 2022 Scorecard on State Health System Performance. In that environment, nurses raced to keep pace with information tracking, room to room, patient to patient. And each nurse shift change involved a transfer of data from one care team to another.

While the human side of nursing remained more critical than ever during the pandemic, it also became clear that the “intelligence” load might be better shared by humans and computers working together. Many call this partnership “artificial intelligence” or “collaborative intelligence.” 

Students pursuing the Bachelor of Science in Nursing at John Carroll will grow accustomed to this new era while training in a state-of-the-art, high-fidelity simulation lab located in the Dolan Center for Science and Technology. John Carroll nursing students will gain further exposure through clinical experience in Cleveland’s world-class medical institutions: Cleveland Clinic, University Hospitals, and the MetroHealth System. Early exposure to forward-looking, evidence-based nursing and new technologies will serve as a cornerstone of the John Carroll University BSN.

Risk Prediction Algorithms

In healthcare, artificial intelligence can cover a range of applications, from risk prediction algorithms to robotic assisted surgery and sensor-based technologies. Enter a product like the Rothman Index (RI), a clinical support system created by PeraHealth, Inc. and used in ICU units to spot the early stages of clinical deterioration. The Rothman Index algorithm uses time-updated physiological data (vital signs, lab values, nursing assessments) to calculate a single score that objectively quantifies a patient’s condition. Its creators describe the tool as real-time clinical surveillance.

The Rothman Index score is used to detect a patient’s declining health and predict risk of decompensation and death. Various risk thresholds exist to describe a patient’s status. “Medium” risk is defined by a 30% reduction in a patient’s RI score over 24 hours, “high” risk is defined by a 40% drop in the RI score over 12 hours, and “very high” risk” is having a score less than or equal to 20 on the RI. 

Hospitals create protocols to be followed by providers when an RI risk threshold is reached (such as triggering rapid response team mobilization). The inputs and thresholds can be customized to each patient, in effect creating another “set of eyes” in an otherwise urgency-filled clinical environment. At the administrative level, this kind of “artificial intelligence” system provides insights on ICU stress, a measure designed to help hospitals plan and manage their surge capacity.

System-Wide Information Sharing

As the medical knowledge base expands and technological advances in health care make for more and more specialized care and providers, the need arises for a more reliable way to hand off information from one expert to another, and across successive shift changes.

In an ICU unit, a care team can include the primary ICU nurse, ICU physician, charge nurse, respiratory therapist, physical therapist, registered dietician, wound care nurse, infectious disease. Pharmacists, case managers and clergy can also be involved. In this setting, patients encounter countless handoffs, each one posing a range of risk — including gaps in patient care and breaches of patient safety, including medication errors.

As new algorithms and types of artificial intelligence are applied to clinical care, nurses will be tasked with interpreting multiple data results and integrating new information into nursing practice. The balance of automated and semi-automated decision making will be a growing concern for patients, families and care providers.

JCU is a private Jesuit university located in University Heights, Ohio, near Cleveland.