Supervised Machine Learning Approach to Identify Early Predictors of Poor Outcome in Patients with COVID-19
Jason Zucker1, Angela Gomez-Simmonds1, Lawrence J Purpura1
1Division of Infectious Diseases, Columbia University Irving Medical Center, New York, NY 10032, USA.
Insights
Early COVID-19 symptom duration and initial markers predict poor outcomes. Identifying at-risk patients requires considering symptom length alongside clinical and lab data for better management.
Area of Science:
- Infectious Diseases
- Critical Care Medicine
- Medical Informatics
Background:
- Clinical manifestations of coronavirus disease 2019 (COVID-19) vary significantly.
- Understanding symptom duration at hospital presentation is crucial for predictive algorithms.
- Timely identification of high-risk COVID-19 patients is essential for effective resource allocation and treatment.
Purpose of the Study:
- To identify predictors of poor outcomes in hospitalized COVID-19 patients.
- To evaluate the role of symptom duration in conjunction with clinical and laboratory markers.
- To develop a more accurate risk stratification model for COVID-19 severity.
Main Methods:
- Nested case-control analysis of 4103 adult COVID-19 patients with at least 28 days of follow-up.
- Utilized multivariable logistic regression and classification and regression tree (CART) analysis.
- Identified predictors of decompensation and poor outcomes based on symptom duration and clinical variables.
Main Results:
- Patients presenting earlier (<4 days) were older, had more comorbidities, and higher decompensation rates (41%).
- Oxygen delivery method was a key predictor of decompensation.
- Specific combinations of symptom duration, age, neutrophil/lymphocyte ratio, IL-6, D-dimer, and CRP levels independently predicted poor outcomes.
Conclusions:
- Symptom duration is a critical factor in assessing COVID-19 patient risk.
- Combining symptom duration with initial clinical and laboratory markers enhances risk prediction.
- This approach can identify COVID-19 patients at increased risk for adverse outcomes, informing clinical decision-making.
Background:
The progression of clinical manifestations in patients with coronavirus disease 2019 (COVID-19) highlights the need to account for symptom duration at the time of hospital presentation in decision-making algorithms.
Methods:
We performed a nested case-control analysis of 4103 adult patients with COVID-19 and at least 28 days of follow-up who presented to a New York City medical center. Multivariable logistic regression and classification and regression tree (CART) analysis were used to identify predictors of poor outcome.
Results:
Patients presenting to the hospital earlier in their disease course were older, had more comorbidities, and a greater proportion decompensated (<4 days, 41%; 4-8 days, 31%; >8 days, 26%). The first recorded oxygen delivery method was the most important predictor of decompensation overall in CART analysis. In patients with symptoms for <4, 4-8, and >8 days, requiring at least non-rebreather, age ≥ 63 years, and neutrophil/lymphocyte ratio ≥ 5.1; requiring at least non-rebreather, IL-6 ≥ 24.7 pg/mL, and D-dimer ≥ 2.4 µg/mL; and IL-6 ≥ 64.3 pg/mL, requiring non-rebreather, and CRP ≥ 152.5 mg/mL in predictive models were independently associated with poor outcome, respectively.
Conclusion:
Symptom duration in tandem with initial clinical and laboratory markers can be used to identify patients with COVID-19 at increased risk for poor outcomes.
More Related Videos
07:31Implementation of a Real-Time Psychosis Risk Detection and Alerting System Based on Electronic Health Records using CogStack
Published on: May 15, 2020
08:05Lung CT Segmentation to Identify Consolidations and Ground Glass Areas for Quantitative Assesment of SARS-CoV Pneumonia
Published on: December 19, 2020
Related Concept Videos
01:18Steps in Outbreak Investigation
01:24Kaplan-Meier Approach