Learning From Past Respiratory Infections to Predict COVID-19 Outcomes: Retrospective Study

Abstract

Background For the clinical care of patients with well-established diseases, randomized trials, literature, and research are supplemented with clinical judgment to understand disease prognosis and inform treatment choices. In the void created by a lack of clinical experience with COVID-19, artificial intelligence (AI) may be an important tool to bolster clinical judgment and decision making. However, a lack of clinical data restricts the design and development of such AI tools, particularly in preparation for an impending crisis or pandemic. Objective This study aimed to develop and test the feasibility of a “patients-like-me” framework to predict the deterioration of patients with COVID-19 using a retrospective cohort of patients with similar respiratory diseases. Methods Our framework used COVID-19–like cohorts to design and train AI models that were then validated on the COVID-19 population. The COVID-19–like cohorts included patients diagnosed with bacterial pneumonia, viral pneumonia, unspecified pneumonia, influenza, and acute respiratory distress syndrome (ARDS) at an academic medical center from 2008 to 2019. In total, 15 training cohorts were created using different combinations of the COVID-19–like cohorts with the ARDS cohort for exploratory purposes. In this study, two machine learning models were developed: one to predict invasive mechanical ventilation (IMV) within 48 hours for each hospitalized day, and one to predict all-cause mortality at the time of admission. Model performance was assessed using the area under the receiver operating characteristic curve (AUROC), sensitivity, specificity, positive predictive value, and negative predictive value. We established model interpretability by calculating SHapley Additive exPlanations (SHAP) scores to identify important features. Results Compared to the COVID-19–like cohorts (n=16,509), the patients hospitalized with COVID-19 (n=159) were significantly younger, with a higher proportion of patients of Hispanic ethnicity, a lower proportion of patients with smoking history, and fewer patients with comorbidities (P<.001). Patients with COVID-19 had a lower IMV rate (15.1 versus 23.2, P=.02) and shorter time to IMV (2.9 versus 4.1 days, P<.001) compared to the COVID-19–like patients. In the COVID-19–like training data, the top models achieved excellent performance (AUROC>0.90). Validating in the COVID-19 cohort, the top-performing model for predicting IMV was the XGBoost model (AUROC=0.826) trained on the viral pneumonia cohort. Similarly, the XGBoost model trained on all 4 COVID-19–like cohorts without ARDS achieved the best performance (AUROC=0.928) in predicting mortality. Important predictors included demographic information (age), vital signs (oxygen saturation), and laboratory values (white blood cell count, cardiac troponin, albumin, etc). Our models had class imbalance, which resulted in high negative predictive values and low positive predictive values. Conclusions We provided a feasible framework for modeling patient deterioration using existing data and AI technology to address data limitations during the onset of a novel, rapidly changing pandemic.

Consolidated Reporting Guidelines for Prognostic and Diagnostic Machine Learning Models (CREMLS)

“…Describes how missing values were replaced [20] Strategies for handling missing data 3.1 Describes the approach of using SMOTE c to adjust class ratios to address imbalance [23] Strategies for addressing class imbalance 3.2 Describes the vectorization of a dimension of 100 into a 2D space using an established algorithm [22] Strategies for reducing dimensionality of data 3.3…”

“…Describes the institution as an academic (teaching) community hospital where the data were collected [23] The medical institutional settings 1.6…”

“…Defined in Figure 1 in the paper by Kendale et al [5] Inclusion or exclusion criteria for the patient cohort 2.1 Describes sources and methods of data collection, what type of data were used, and potential implied bias in interpretation [23] Methods of data collection 2.2 Discusses potential bias in data collection and outcome definition [23] Bias introduced due to the method of data collection used 2.3…”

“…Uses descriptive statistics to show data characteristics for different types of data (demographics and clinical measurements) [23]…”

Predicting the Next-Day Perceived and Physiological Stress of Pregnant Women by Using Machine Learning and Explainability: Algorithm Development and Validation

“…We tested 6 widely used machine learning models using the scikit-learn Python package to evaluate the importance of input variables on next-day prediction of perceived and physiological stress [47][48][49]. In baseline models, we included the following models with default hyperparameters: gradient boost machine (min_samples_split: 5; min_samples_leaf: 2; max_depth: 3), SVM (kernel: rbf; C: 1.0; gamma: "scale"), adaptive boosting (n_estimators: 50), naïve Bayes, decision tree (min_samples_split: 5; min_samples_leaf: 2; max_depth: 3), and random forest (n_estimators: 10; min_samples_split: 5; min_samples_leaf: 2; max_depth: 3).…”