Adverse Events
Akin to the "unforced error" in sports culture, adverse events (AEs) are defined as unintended injuries during hospitalization rather than the original disease process. According to a recent study New England Journal of Medicine, the likelihood of a patient experiencing at least one adverse event is roughly 25%, and with 32M hospitalizations per year, adverse events are becoming an increasingly prevalent issue in today's healthcare institutions. On average, adverse events annually contribute to both 210,000 deaths and cost the US healthcare market $20B; this has led some experts to dub adverse events as the $20B problem.
Fortunately, many adverse events are believed to be preventable. In many cases, unusual fluctuations in a patient's vital signs often serve as an indication of an impending adverse event. For example, heart rate variability and increases in respiration rate and temperature levels have been identified to be common precursors to sepsis or infections.
Sepsis
Marked by heart rate variability, and increases in respiration rate and temperature
Patient Restlessness
Marked by rapid fluctuation in in-bed activity or sudden increases in heart rate
Opioid Overdose
Marked by declining oxygen saturation levels and in-bed activity
Drug Interactions
Marked by heart rate variability, increases or decreases in respiration rate and blood pressure
Cardiac Arrest
Marked by heart rate variability, and changes in respiratory rates and systolic blood pressure
Increased Stress
Marked by declining heart rate beat-to-beat variability
Unfortunately, many hospitals, especially in non-ICU units, face challenges in identifying these adverse event precursors as they are occurring; thus, many adverse events are reacted to at their onset, rather than actively being prevented.
Lack of Real Time Data
Systems outside of the ICU often do not provide data to healthcare practitioners remorely and in real time. Systems that do are expensive, have limited or no interoperability or rely heavily on EMR data
Demanding Patient Ratios
Non-ICU nurses may care for up to five or six patients at once. This makes it impossible to solely rely on manual monitoring of patients to observe event markers.
Poor Technology
Existing AI/Machine Learning adverse event anticipation algorithms often rely on poor datasets that have inconsistent timescales, miss values, and have no label for AE onset
As a result of these limitations, the most reliable method to understand a patient's physiology is to manually and physically assess the patient and their medical devices. If a nurse does not have the ability to do so, nurses often rely on the sound of various medical device alarms to prioritize their workflow, and this work flow becomes particularly problematic if nurses have multiple patients to care for, and worse still, if a nurse has multiple patients declining at once. Alarm fatigue paired with the immense task of adverse event prevention often makes patient care a herculean task.
With the growing shortage of nurses in the US, high levels of nurse burnout, and the aging nursing workforce, hospitals are in dire need of a robust, cost effective solution that enables non-ICU nurses to be alerted about changes in the condition of their patients regardless of their location in the hospital.