A real-time integrated framework to support clinical decision making for covid-19 patients.

BACKGROUND: The COVID-19 pandemic affected healthcare systems worldwide. Predictive models developed by Artificial Intelligence (AI) and based on timely, centralized and standardized real world patient data could improve management of COVID-19 to achieve better clinical outcomes. The objectives of this manuscript are to describe the structure and technologies used to construct a COVID-19 Data Mart architecture and to present how a large hospital has tackled the challenge of supporting daily management of COVID-19 pandemic emergency, by creating a strong retrospective knowledge base, a real time environment and integrated information dashboard for daily practice and early identification of critical condition at patient level. This framework is also used as an informative, continuously enriched data lake, which is a base for several on-going predictive studies. METHODS: The information technology framework for clinical practice and research was described. It was developed using SAS Institute software analytics tool and SAS® Vyia® environment and Open-Source environment R ® and Python ® for fast prototyping and modeling. The included variables and the source extraction procedures were presented. RESULTS: The Data Mart covers a retrospective cohort of 5528 patients with SARS-CoV-2 infection. People who died were older, had more comorbidities, reported more frequently dyspnea at onset, had higher d-dimer, C-reactive protein and urea nitrogen. The dashboard was developed to support the management of COVID-19 patients at three levels: hospital, single ward and individual care level. INTERPRETATION: The COVID-19 Data Mart based on integration of a large collection of clinical data and an AI-based integrated framework has been developed, based on a set of automated procedures for data mining and retrieval, transformation and integration, and has been embedded in the clinical practice to help managing daily care. Benefits from the availability of a Data Mart include the opportunity to build predictive models with a machine learning approach to identify undescribed clinical phenotypes and to foster hospital networks. A real-time updated dashboard built from the Data Mart may represent a valid tool for a better knowledge of epidemiological and clinical features of COVID-19, especially when multiple waves are observed, as well as for epidemic and pandemic events of the same nature (e. g. with critical clinical conditions leading to severe pulmonary inflammation). Therefore, we believe the approach presented in this paper may find several applications in comparable situations even at region or state levels. Finally, models predicting the course of future waves or new pandemics could largely benefit from network of DataMarts.

A machine-learning parsimonious multivariable predictive model of mortality risk in patients with Covid-19.

The COVID-19 pandemic is impressively challenging the healthcare system. Several prognostic models have been validated but few of them are implemented in daily practice. The objective of the study was to validate a machine-learning risk prediction model using easy-to-obtain parameters to help to identify patients with COVID-19 who are at higher risk of death. The training cohort included all patients admitted to Fondazione Policlinico Gemelli with COVID-19 from March 5, 2020, to November 5, 2020. Afterward, the model was tested on all patients admitted to the same hospital with COVID-19 from November 6, 2020, to February 5, 2021. The primary outcome was in-hospital case-fatality risk. The out-of-sample performance of the model was estimated from the training set in terms of Area under the Receiving Operator Curve (AUROC) and classification matrix statistics by averaging the results of fivefold cross validation repeated 3-times and comparing the results with those obtained on the test set. An explanation analysis of the model, based on the SHapley Additive exPlanations (SHAP), is also presented. To assess the subsequent time evolution, the change in paO2/FiO2 (P/F) at 48 h after the baseline measurement was plotted against its baseline value. Among the 921 patients included in the training cohort, 120 died (13%). Variables selected for the model were age, platelet count, SpO2, blood urea nitrogen (BUN), hemoglobin, C-reactive protein, neutrophil count, and sodium. The results of the fivefold cross-validation repeated 3-times gave AUROC of 0.87, and statistics of the classification matrix to the Youden index as follows: sensitivity 0.840, specificity 0.774, negative predictive value 0.971. Then, the model was tested on a new population (n = 1463) in which the case-fatality rate was 22.6%. The test model showed AUROC 0.818, sensitivity 0.813, specificity 0.650, negative predictive value 0.922. Considering the first quartile of the predicted risk score (low-risk score group), the case-fatality rate was 1.6%, 17.8% in the second and third quartile (high-risk score group) and 53.5% in the fourth quartile (very high-risk score group). The three risk score groups showed good discrimination for the P/F value at admission, and a positive correlation was found for the low-risk class to P/F at 48 h after admission (adjusted R-squared = 0.48). We developed a predictive model of death for people with SARS-CoV-2 infection by including only easy-to-obtain variables (abnormal blood count, BUN, C-reactive protein, sodium and lower SpO2). It demonstrated good accuracy and high power of discrimination. The simplicity of the model makes the risk prediction applicable for patients in the Emergency Department, or during hospitalization. Although it is reasonable to assume that the model is also applicable in not-hospitalized persons, only appropriate studies can assess the accuracy of the model also for persons at home.