SHM Converge Scientific Abstract ePoster Gallery-Interprofessional Communication and Heart Failure Inpatients' Length of Stay

Pdf Summary

This study examined whether interprofessional communication (IPC), measured through Epic Secure Chat metadata and social network analysis, is associated with length of stay in hospitalized heart failure patients. The authors analyzed 1,026 CHF hospitalizations at NYU from 2021 and 2023, using 100 variables that included patient factors, discharge factors, and network-based IPC metrics such as density, reciprocity, centrality, and transitivity.<br /><br />Two models were compared: ordinary least squares regression (OLS) and an explainable boosting machine (EBM). Both performed similarly, with mean squared error of 0.10 for OLS and 0.14 for EBM. The models identified several IPC variables that helped explain variance in observed-to-expected length of stay (O:E LOS). In the OLS model, higher network reciprocity and nurse indegree centrality were associated with increased O:E LOS, while more sent messages per day and greater social worker/care manager transitivity were associated with decreased O:E LOS. The EBM model highlighted 14 important IPC-related variables, including network density, nurse betweenness centrality, and sent messages per day.<br /><br />Across both models, the most overlapping important IPC metrics were network density, network reciprocity, network size, and nurse indegree centrality. The authors conclude that EHR-based IPC data can explain variation in hospital length of stay even after adjusting for patient characteristics. They also note that the explainable model provided more detailed insight into how each variable relates to outcomes. Overall, the study suggests that advanced analytics applied to communication data may help identify areas for improving interprofessional collaboration and potentially enhance patient care.

Asset Subtitle

William R. Small

Meta Tag

Author List Jesse Burk-Rafel, Nicholas Genes, Rich Caruana, Stefanie Sebok-Syer, William R. Small

Category Research

Concept Interprofessional Communication

Concept Social Network Analysis

Concept Observed-to-Expected Length of Stay

Concept Electronic Health Record

Concept Secure Message Metadata

Distinguished Non-Finalist

Presenter Organization NYU Langone Health

Presenting Author William R. Small

Track Communication

Keywords

interprofessional communication

Epic Secure Chat

social network analysis

length of stay

heart failure hospitalization

observed-to-expected LOS

ordinary least squares

explainable boosting machine

network reciprocity

nurse centrality

Interprofessional Communication

Social Network Analysis

Observed-to-Expected Length of Stay

Electronic Health Record

Secure Message Metadata