Development, evaluation, and validation of machine learning models for COVID-19 detection based on routine blood tests

Abstract

ObjectivesThe rRT-PCR test, the current gold standard for the detection of coronavirus disease (COVID-19), presents with known shortcomings, such as long turnaround time, potential shortage of reagents, false-negative rates around 15–20%, and expensive equipment. The hematochemical values of routine blood exams could represent a faster and less expensive alternative.MethodsThree different training data set of hematochemical values from 1,624 patients (52% COVID-19 positive), admitted at San Raphael Hospital (OSR) from February to May 2020, were used for developing machine learning (ML) models: the complete OSR dataset (72 features: complete blood count (CBC), biochemical, coagulation, hemogasanalysis and CO-Oxymetry values, age, sex and specific symptoms at triage) and two sub-datasets (COVID-specific and CBC dataset, 32 and 21 features respectively). 58 cases (50% COVID-19 positive) from another hospital, and 54 negative patients collected in 2018 at OSR, were used for internal-external and external validation.ResultsWe developed five ML models: for the complete OSR dataset, the area under the receiver operating characteristic curve (AUC) for the algorithms ranged from 0.83 to 0.90; for the COVID-specific dataset from 0.83 to 0.87; and for the CBC dataset from 0.74 to 0.86. The validations also achieved good results: respectively, AUC from 0.75 to 0.78; and specificity from 0.92 to 0.96.ConclusionsML can be applied to blood tests as both an adjunct and alternative method to rRT-PCR for the fast and cost-effective identification of COVID-19-positive patients. This is especially useful in developing countries, or in countries facing an increase in contagions.

Accurate detection of Covid-19 patients based on Feature Correlated Naïve Bayes (FCNB) classification strategy

“…2. These three methodologies are (1) Real-Time reverse transcriptase-Polymerase Chain Reaction (RT-PCR) (Tahamtan and Ardebili 2020;Waller et al 2020;Li et al 2020a, d), (2) chest CT imaging scan (Mishra et al 2020;Li et al 2020a, e;Kovács et al 2020), and (3) numerical laboratory tests (Brinati et al 2020;Kukar et al 2020;Cabitza et al 2020;Qiu et al 2020). RT-PCR tests are fairly quick, sensitive, and reliable.…”

Machine Learning based COVID-19 Diagnosis from Blood Tests with Robustness to Domain Shifts

“…In this large-scale study, we train and evaluate our models with more samples than most studies [18][19][20][21][22] . Besides our large number of tested subjects, we also exploit pre-pandemic negative samples, which vastly increases our data set size.…”

“…We evaluate and compare our proposed approach of lifelong learning and assessment against standard ML approaches on a large-scale data set. This data set comprises 127,115 samples after pre-processing and merging, which exceeds the data set size of many small scale studies [18][19][20][21][22]32 by far. Our data set comprises pre-pandemic negative samples and pandemic negative and positive samples spanning over multiple different departments of the Kepler University Hospital, Linz.…”

“…COVID-19 and the patient's prognosis can be predicted from chest CT-scans, X-rays [11][12][13][14] or sound recordings of coughs or breathing [15][16][17] . Furthermore, it has been shown that ML models based on blood tests are capable of detecting COVID-19 infection [18][19][20][21][22][23][24][25][26][27][28][29][30][31][32] and predicting other outcomes, such as survival or admission to an intensive care unit [33][34][35][36][37][38][39][40][41] .…”

“…Such an unexpected behavior could even lead to unfavorable consequences, like a major disease outbreak in a hospital. Most of the previous ML studies evaluated the predictive performance of the learned models by cross-validation, bootstrapping or fixed splits on randomly drawn samples [18][19][20][21][22][26][27][28][29][30][31][32] . However, the theoretical justification of these evaluation methods is heavily founded on the central ML paradigm: that the distributions remain constant over time.…”