How machine learning could be used in clinical practice during an epidemic

Abstract

The COVID-19 epidemic is the cause of a crisis that is confronting the healthcare community with an unprecedented situation: emergency and intensive care units (ICU) are saturated, compelling physicians to make extremely hard decisions (triage). In such a resourceconstrained situation, physicians need decision support systems that could help them to optimally stratify patient risk. In the present paper, we explain how machine learning (ML) could help clinical practitioners during the epidemic (Fig. 1), and we describe some challenges and perspectives that could direct future efforts in the field. Why might machine learning be of particular interest?The first point to note is that ML encompasses a wide range of data analysis algorithms that run along a continuum from partially to fully machine-guided [1, 2]. For instance, deep learning is a fully machine-guided approach that uses large combinations of mathematical functions called "artificial neurons". Although recent contributions are promising [3], deep learning is not seen as particularly useful by some of the healthcare community because it produces black box models that are not readily interpretable by physicians [4]. Therefore, efforts to develop ML-based decision support systems for clinical practice should focus on models that incorporate as much prior medical knowledge as possible.A controversial step in traditional decision-making algorithms is the selection of variables (often termed "features") that are included in the predictive model.

Risk assessment in COVID-19 patients: A multiclass classification approach

“…Because these symptoms also occur in common colds, it is mandatory to know others that can occur, such as nasal congestion (rhinorrhea), diarrhea and rashes, or color changes in fingers or toes. Based on our information, the major symptoms among infected patients were: fever, headache, cough, and dyspnea (78.51%, 76.14%, 74.28%, and 38.54%), which have been consistent in several studies with the highest reported percentages [6] , [10] . We can point to dyspnea as the comorbidity that can distinguish more patients in a risk scenario because its prevalence percentage is 86.29% of cases.…”

“…Indeed, one aspect that we can highlight from the various studies on the COVID-19 pandemic is that it has driven unprecedented technological developments. Many of these relate to Artificial Intelligence and Machine Learning, and their interaction with diverse areas of knowledge, whether medical, social or economic; including the sciences of massive data and computational analysis [10] , [13] . Amidst this scenario, the health sector will need to incorporate these resources into its analysis and diagnosis support systems, not only of infectious diseases but of any other nature.…”

“…Medical practice is singular and represents a large-scale challenge. Dealing with sensitive information using relevant clinical data leads us to re-formulate how ML could help solve these problems [10] .…”

Non-contact screening system based for COVID-19 on XGBoost and logistic regression

“…Infected patients experience symptoms such as fever, fatigue, dry cough, and dyspnea [ 8 ]. Some researchers have attempted to detect these symptoms using non-contact devices and machine learning algorithms [ 9 ]. They focused on two main issues: the features extracted from the subject and the algorithms used to recognize infected subjects.…”

Economic impact of a machine learning-based strategy for preparation of blood products in brain tumor surgery

“…Nowadays, machine learning (ML) has been studied in clinical research to support decision-making [13,14]. However, how machine learning is used in general practice remains a challenge [15].…”