Developing and validating a machine learning prognostic model for alerting to imminent deterioration of hospitalized patients with COVID-19

Abstract

Our study was aimed at developing and validating a new approach, embodied in a machine learning-based model, for sequentially monitoring hospitalized COVID-19 patients and directing professional attention to patients whose deterioration is imminent. Model development employed real-world patient data (598 prediction events for 210 patients), internal validation (315 prediction events for 97 patients), and external validation (1373 prediction events for 307 patients). Results show significant divergence in longitudinal values of eight routinely collected blood parameters appearing several days before deterioration. Our model uses these signals to predict the personal likelihood of transition from non-severe to severe status within well-specified short time windows. Internal validation of the model's prediction accuracy showed ROC AUC of 0.8 and 0.79 for prediction scopes of 48 or 96 h, respectively; external validation showed ROC AUC of 0.7 and 0.73 for the same prediction scopes. Results indicate the feasibility of predicting the forthcoming deterioration of non-severe COVID-19 patients by eight routinely collected blood parameters, including neutrophil, lymphocyte, monocyte, and platelets counts, neutrophil-to-lymphocyte ratio, CRP, LDH, and D-dimer. A prospective clinical study and an impact assessment will allow implementation of this model in the clinic to improve care, streamline resources and ease hospital burden by timely focusing the medical attention on potentially deteriorating patients.

Deciphering the microbial landscape of lower respiratory tract infections: insights from metagenomics and machine learning

“…Machine learning applications in disease diagnosis ( Lai et al., 2024 ), complication prediction ( Pax et al., 2024 ), and forecasting of factors such as bacterial drug resistance and predictive models for bacteriophage therapy of Escherichia coli urinary tract infections have demonstrated promising predictive efficacy ( Hu et al., 2023 ; Dixit et al., 2024 ; Keith et al., 2024 ; Nsubuga et al., 2024 ). Additionally, numerous clinical machine learning prediction models have been developed to predict disease prognosis and survival time by collecting large-scale clinical features ( Kogan et al., 2022 ; Li et al., 2022 ; Tang et al., 2022 ; Li et al., 2023 ), demonstrating excellent predictive performance. However, in the actual treatment of patients with lower respiratory tract infections ( Sethi, 2010 ; Jain et al., 2015 ), the complex relationships between microorganisms must be considered.…”

Multidimensional dynamic prediction model for hospitalized patients with the omicron variant in China

“…Kogan et al. first developed a dynamic machine learning model for alerting of the imminent deterioration of hospitalized COVID-19 patients using longitudinal values of eight routinely collected laboratory test results ( Kogan et al., 2022 ). However, the patient's clinical condition changes could not only be determined by daily laboratory test results but also need to consider multidimensional variables on each day, including viral infection reflected by the disease course, the patient's underlying condition and organ dysfunction reflected by vital signs and laboratory tests, and the timing, manner, and intensity of clinical treatments.…”

“…Although numerous studies have tried to predict the risks borne by COVID-19 patients, there are still unmet challenges. Most previous studies used patient status at some fixed time point (usually at admission) ( Assaf et al., 2020 ; Gao et al., 2020 ) or used only a limited dimension of longitudinal clinical status, e.g., only laboratory tests ( Kogan et al., 2022 ), to predict the risk of deterioration or death. However, a patient outcome is determined by the interaction of viral invasion and replication process, immune response intensity, underlying physical condition, disease progression, state of organ dysfunction, and the timing, manner, and intensity of clinical treatments, which are constantly evolving.…”