To evaluate a clinical decision support system (CDSS) designed to predict hypotension in real time in ICU patients.
Approach:
Study Design: Pilot evaluation involving 20 adult ICU patient-subject/nurse-subject pairs with continuous vasopressor therapy and BP monitoring.
CDSS Functionality: The CDSS linked to bedside monitors to predict hypotension events and issue notifications based on MAP data.
Data Collection: Data were collected from 18 patients over 3.3 patient-days, focusing on hypotension-clusters.
Key Findings:
The median patient experienced hypotension for 119.7 min/day.
The CDSS detected every hypotension-cluster (100%) and predicted 91.7% of vasopressor dose increases.
44% of hypotension-clusters were predicted before they began.
Interpretation:
The findings validate that MAP frequently falls below target in temporal clusters, and early flickering hypotension often precedes deeper hypotensive episodes.
Limitations:
The study involved a small sample size of 20 subjects.
The CDSS did not generate predictions during active hypotension.
Conclusion:
Timely intervention after predicted hypotension may have reduced the duration of hypotension.
by Iain Edmund Kehoe, Jeremy Alanano Pepino, Jin-Oh Hahn, Mary McAuley, Laura Lux, Abigail O’Connell, Noelle Saillant, Haitham S. Alabsi, Jarone Lee, Andrew Tomas Reisner
