(362) Unsupervised Machine Learning of ICU Medication Use Patterns to Predict Fluid Overload

Introduction: Intravenous (IV) medications are a fundamental cause of ICU fluid overload (FO); however, the association between IV medication use (including volume), administration timing, and FO occurrence remains unclear. We conducted a time-dependent, unsupervised machine learning (ML) analysis to determine patterns of IV medication use associated with FO.

Methods: This retrospective cohort study included consecutive adults admitted to an ICU ≥72 hours at one academic center with fluid balance data. FO was defined as a positive fluid balance ≥7% of admission body weight within 72 hours of ICU admission. After reviewing medication administration record (MAR) data in three-hour periods, IV medication exposure was categorized into clusters using principal component analysis (PCA) and Restricted Boltzmann Machine (RBM). Then, IV medication regimens of patients with and without FO were compared between clusters to assess for temporal clusters associated with FO using the Wilcoxon rank sum test. Exploratory analyses of the medication cluster most associated with FO for medications frequently appearing and used in the first 24 hours was conducted.

Results: FO occurred in 127/927 (13.7%) of the patients enrolled. Patients received a median (IQR) of 31 (13-65) discrete IV medication administrations over the 72-hour period. Across all 2,229 IV medication administrations, ten unique IV medication clusters were identified; 97 distinct IV medications were common across each cluster. Among the ten clusters, cluster 7 (n=127 unique IV medications) had the greatest association with FO; the mean number of cluster 7 medications was significantly greater in patients who developed FO (vs. those who did not)(25.6 vs.10.9. p< 0.0001). A total of 51 (40.2%) of the cluster 7 medications were administered in >5 separate 3-hour periods. The most common cluster 7 medications included continuous infusions, antibiotics, and sedatives/analgesics. Fifteen cluster 7 medications were administered only in the first 24 hours of ICU admission.

Conclusions: Using ML approaches, we identified a unique IV medication cluster strongly associated with ICU FO. Within this cluster, a unique subset of IV medications is associated with increased FO when given early in the ICU stay. Strategies to mitigate IV medications as a risk for FO are required.