Key result
Recurrent neural network using 48-hour data predicts ICU readmission with 0.82 AUC, beating full-stay models.
Why the study?
Unexpected ICU readmission endangers patients' lives, which could be mitigated by stratifying readmission risk at discharge using machine learning methods.
Does an RNN model using the last 48 hours of clinical time series data improve prediction of cardiovascular ICU readmission compared to using all available data?
Observational
Yes
Does an RNN model using the last 48 hours of clinical time series data improve prediction of cardiovascular ICU readmission compared to using all available data?
Effect estimate: AUC-ROC 0.82
An LSTM-based recurrent neural network using the last 48 hours of ICU clinical data shows promise for predicting cardiovascular ICU readmissions with good discrimination (AUC 0.82).
May support recent-data models for ICU readmission risk; leaves open prospective validation before cardiovascular use.
Unexpected readmission to intensive care units (ICUs) endangers patients' lives due to premature patient transfers or prolonged stays at the care units. This can be mitigated by stratification of the readmission risk at discharge times using state-of-the-art machine learning (ML) methods. We fitted two alternative recurrent neural network (RNN) models based on long short-term memory (LSTM) on the Medical Information Mart for Intensive Care (MIMIC-III) dataset and evaluated them with an independent cohort from our hospital's ICU (UKD). The first model processed all the available time series data from each patient's ICU stay, whereas the second model focused on the data from the last 48 hours of the ICU stay prior to transfer. Our readmission prediction on MIMIC data reached an area under the curve of receiver operating characteristic (AUC-ROC) of 0.82. Furthermore, the model with the 48 h time frame outperformed the other model, as both models were applied to the independent test cohort. The results suggest that the RNN model for time series forecasting holds promise for future use as a clinical decision support tool, although follow-up studies with larger cohorts as well as user studies should be conducted to assess the generalizability and usability of the methods, respectively.
No takes yet. Share an insight, caveat, or question.
Moazemi et al. (2022) conducted an observational in Cardiovascular ICU readmission. Recurrent neural network (RNN) model (last 48 hours of data) vs. RNN model (all available ICU stay data) was evaluated on Prediction of ICU readmission (AUC-ROC 0.82). A recurrent neural network model using the last 48 hours of clinical time series data predicted ICU readmission with an AUC-ROC of 0.82, outperforming a model using all available ICU stay data.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: