Predicting mortality in SARS-COV-2 (COVID-19) positive patients in the inpatient setting using a novel deep neural network

Predicting COVID-19 prognosis in hospitalized patients based on early status

IMPORTANCE

COVID-19 remains the fourth leading cause of death in the United States. Predicting COVID-19 patient prognosis is essential to help efficiently allocate resources, including ventilators and intensive care unit beds, particularly when hospital systems are strained. Our PLABAC and PRABLE models are unique because they accurately assess a COVID-19 patient's risk of death from only age and five commonly ordered laboratory tests. This simple design is important because it allows these models to be used by clinicians to rapidly assess a patient's risk of decompensation and serve as a real-time aid when discussing difficult, life-altering decisions for patients. Our models have also shown generalizability to external populations across the United States. In short, these models are practical, efficient tools to assess and communicate COVID-19 prognosis.

The accuracy of artificial intelligence in predicting COVID-19 patient mortality: a systematic review and meta-analysis

BACKGROUND

The purpose of this paper was to systematically evaluate the application value of artificial intelligence in predicting mortality among COVID-19 patients.

METHODS

The PubMed, Embase, Web of Science, CNKI, Wanfang, China Biomedical Literature, and VIP databases were systematically searched from inception to October 2022 to identify studies that evaluated the predictive effects of artificial intelligence on mortality among COVID-19 patients. The retrieved literature was screened according to the inclusion and exclusion criteria. The quality of the included studies was assessed using the QUADAS-2 tools. Statistical analysis of the included studies was performed using Review Manager 5.3, Stata 16.0, and Meta-DiSc 1.4 statistical software. This meta-analysis was registered in PROSPERO (CRD42022315158).

FINDINGS

Of 2193 studies, 23 studies involving a total of 25 AI models met the inclusion criteria. Among them, 18 studies explicitly mentioned training and test sets, and 5 studies did not explicitly mention grouping. In the training set, the pooled sensitivity was 0.93 [0.87, 0.96], the pooled specificity was 0.94 [0.87, 0.97], and the area under the ROC curve was 0.98 [0.96, 0.99]. In the validation set, the pooled sensitivity was 0.84 [0.78, 0.88], the pooled specificity was 0.89 [0.85, 0.92], and the area under the ROC curve was 0.93 [1.00, 0.00]. In the subgroup analysis, the areas under the summary receiver operating characteristic (SROC) curves of the artificial intelligence models KNN, SVM, ANN, RF and XGBoost were 0.98, 0.98, 0.94, 0.92, and 0.91, respectively. The Deeks funnel plot indicated that there was no significant publication bias in this study (P > 0.05).

INTERPRETATION

Artificial intelligence models have high accuracy in predicting mortality among COVID-19 patients and have high prognostic value. Among them, the KNN, SVM, ANN, RF, XGBoost, and other models have the highest levels of accuracy.

Coronavirus pandemic in the South Asia region: Health policy and economy trade-off

Background

The South Asian Association for Regional Cooperation (SAARC) covers Afghanistan, Bangladesh, Bhutan, India, Maldives, Nepal, Pakistan, and Sri Lanka. We conducted a comparative analysis of the trade-off between the health policies for the prevention of COVID-19 spread and the impact of these policies on the economies and livelihoods of the South Asia populations.

Methods

We analyzed COVID-19 data on epidemiology, public health and health policy, health system capacity, and macroeconomic indicators from January 2020 to March 2021 to determine temporal trends by conducting joinpoint regression analysis using average weekly percent change (AWPC).

Results

Bangladesh had the highest statistically significant AWPC for new COVID-19 cases (17.0; 95% CI = 7.7-27.1, P < 0.001), followed by the Maldives (12.9; 95% CI = 5.3-21.0, P < 0.001) and India (10.0; 95% CI = 8.4-11.5, P < 0.001). The AWPC for COVID-19 deaths was significant for India (6.5; 95% CI = 4.3-8.9, P < 0.001) and Bangladesh (6.1; 95% CI = 3.7-8.5, P < 0.001). Nepal (55.79%), and India (34.91%) had the second- and third-highest increase in unemployment, while Afghanistan (6.83%) and Pakistan (16.83%) had the lowest. The rate of change of real GDP had the highest decrease for Maldives (557.51%), and India (297.03%); Pakistan (46.46%) and Bangladesh (70.80%), however, had the lowest decrease. The government response stringency index for Pakistan had a see-saw pattern with a sharp decline followed by an increase in the government health policy restrictions that approximated the test-positivity trend.

Conclusions

Unlike developed economies, the South Asian developing countries experienced a trade-off between health policy and their economies during the COVID-19 pandemic. South Asian countries (Nepal and India), with extended periods of lockdowns and a mismatch between temporal trends of government response stringency index and the test-positivity or disease incidence, had higher adverse economic effects, unemployment, and burden of COVID-19. Pakistan demonstrated targeted lockdowns with a rapid see-saw pattern of government health policy response that approximated the test-positivity trend and resulted in lesser adverse economic effects, unemployment, and burden of COVID-19.

Application of a Decision Tree Model to Predict the Outcome of Non-Intensive Inpatients Hospitalized for COVID-19

Many studies have identified predictors of outcomes for inpatients with coronavirus disease 2019 (COVID-19), especially in intensive care units. However, most retrospective studies applied regression methods to evaluate the risk of death or worsening health. Recently, new studies have based their conclusions on retrospective studies by applying machine learning methods. This study applied a machine learning method based on decision tree methods to define predictors of outcomes in an internal medicine unit with a prospective study design. The main result was that the first variable to evaluate prediction was the international normalized ratio, a measure related to prothrombin time, followed by immunoglobulin M response. The model allowed the threshold determination for each continuous blood or haematological parameter and drew a path toward the outcome. The model's performance (accuracy, 75.93%; sensitivity, 99.61%; and specificity, 23.43%) was validated with a k-fold repeated cross-validation. The results suggest that a machine learning approach could help clinicians to obtain information that could be useful as an alert for disease progression in patients with COVID-19. Further research should explore the acceptability of these results to physicians in current practice and analyze the impact of machine learning-guided decisions on patient outcomes.

Design of an artificial neural network to predict mortality among COVID-19 patients

Introduction

The fast pandemic of coronavirus disease 2019 (COVID-19) has challenged clinicians with many uncertainties and ambiguities regarding disease outcomes and complications. To deal with these uncertainties, our study aimed to develop and evaluate several artificial neural networks (ANNs) to predict the mortality risk in hospitalized COVID-19 patients.

Material and methods

The data of 1710 hospitalized COVID-19 patients were used in this retrospective and developmental study. First, a Chi-square test (P < 0.05), Eta coefficient (η > 0.4), and binary logistics regression (BLR) analysis were performed to determine the factors affecting COVID-19 mortality. Then, using the selected variables, two types of feed-forward (FF) models, including the back-propagation (BP) and distributed time delay (DTD) were trained. The models' performance was assessed using mean squared error (MSE), error histogram (EH), and area under the ROC curve (AUC-ROC) metrics.

Results

After applying the univariate and multivariate analysis, 13 variables were selected as important features in predicting COVID-19 mortality at P < 0.05. A comparison of the two ANN architectures using the MSE showed that the BP-ANN (validation error: 0.067, most of the classified samples having 0.049 and 0.05 error rates, and AUC-ROC: 0.888) was the best model.

Conclusions

Our findings show the acceptable performance of ANN for predicting the risk of mortality in hospitalized COVID-19 patients. Application of the developed ANN-based CDSS in a real clinical environment will improve patient safety and reduce disease severity and mortality.