Cardiorespiratory instability in monitored step-down unit patients: using cluster analysis to identify patterns of...

Area of Science:

• * Cardiorespiratory instability (CRI) research
• * Clinical data analytics
• * Patient monitoring systems

Background:

• * Cardiorespiratory instability (CRI) in step-down unit (SDU) patients presents diverse etiologies and vital sign (VS) change patterns.
• * Understanding these patterns is crucial for timely clinical intervention and improved patient outcomes.
• * Clustering techniques offer a novel approach to identify distinct patient subgroups based on VS dynamics.

Purpose of the Study:

• * To identify distinct patterns of vital signs during the initial cardiorespiratory instability (CRI) epoch in SDU patients using clustering.
• * To analyze inter-cluster differences in patient admission characteristics and clinical outcomes.
• * To explore the potential of VS patterns in informing clinical decision-making for unstable patients.

Main Methods:

• * Continuous, noninvasive monitoring of heart rate (HR), respiratory rate (RR), and pulse oximetry (SpO2) data were collected at 1/20 Hz.
• * Hierarchical and k-means clustering techniques were employed to identify patterns in the initial CRI epoch (CRI1).
• * 10-fold cross-validation was used to select the optimal clustering solution, followed by assessment of inter-cluster differences.

Main Results:

• * Three distinct clusters of CRI1 vital sign patterns were identified among 165 patients.
• * Cluster 1: Normal/high HR and RR, normal SpO2 (n=30).
• * Cluster 2: Normal HR and RR, low SpO2 (n=103).
• * Cluster 3: Low/normal HR, low RR, normal SpO2 (n=32).
• * Significant differences were observed in patient age (p<0.001), number of comorbidities (p=0.008), and hospital length of stay (p=0.006) across clusters.
• * Older patients and those with more comorbidities were predominantly in Cluster 2.
• * Cluster 1 patients experienced longer hospital stays.
• * No significant differences in SDU length of stay or mortality were found between clusters.

Conclusions:

• * Distinct vital sign pattern clusters characterize cardiorespiratory instability (CRI) in SDU patients.
• * These clusters correlate with variations in patient age, comorbidities, and overall hospital length of stay.
• * Further research is warranted to investigate the physiological underpinnings of these VS clusters and their clinical utility.