Hemogram data as a tool for decision-making in COVID-19 management: applications to resource scarcity scenarios

Abstract

<jats:sec> <jats:title>Background</jats:title> <jats:p>COVID-19 pandemics has challenged emergency response systems worldwide, with widespread reports of essential services breakdown and collapse of health care structure. A critical element involves essential workforce management since current protocols recommend release from duty for symptomatic individuals, including essential personnel. Testing capacity is also problematic in several countries, where diagnosis demand outnumbers available local testing capacity.</jats:p> </jats:sec> <jats:sec> <jats:title>Purpose</jats:title> <jats:p>This work describes a machine learning model derived from hemogram exam data performed in symptomatic patients and how they can be used to predict qRT-PCR test results.</jats:p> </jats:sec> <jats:sec> <jats:title>Methods</jats:title> <jats:p>Hemogram exams data from 510 symptomatic patients (73 positives and 437 negatives) were used to model and predict qRT-PCR results through Naïve-Bayes algorithms. Different scarcity scenarios were simulated, including symptomatic essential workforce management and absence of diagnostic tests. Adjusts in assumed <jats:italic>prior</jats:italic> probabilities allow fine-tuning of the model, according to actual prediction context.</jats:p> </jats:sec> <jats:sec> <jats:title>Results</jats:title> <jats:p>Proposed models can predict COVID-19 qRT-PCR results in symptomatic individuals with high accuracy, sensitivity and specificity, yielding a 100% sensitivity and 22.6% specificity with a <jats:italic>prior</jats:italic> of 0.9999; 76.7% for both sensitivity and specificity with a <jats:italic>prior</jats:italic> of 0.2933; and 0% sensitivity and 100% specificity with a <jats:italic>prior</jats:italic> of 0.001. Regarding background scarcity context, resources allocation can be significantly improved when model-based patient selection is observed, compared to random choice.</jats:p> </jats:sec> <jats:sec> <jats:title>Conclusions</jats:title> <jats:p>Machine learning models can be derived from widely available, quick, and inexpensive exam data in order to predict qRT-PCR results used in COVID-19 diagnosis. These models can be used to assist strategic decision-making in resource scarcity scenarios, including personnel shortage, lack of medical resources, and testing insufficiency.</jats:p> </jats:sec>

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

“…Further, this work presents some major limitations affecting replicability and generalizability, as the authors do not provide any information regarding how the values of the considered features were measured (analytical instruments, analytical principle, and units of measurement). Avila et al [33] used the same dataset considered in [32] to develop a Bayesian model, reporting 76.7% sensitivity and specificity. Notably, the authors report a number of complete instances (510) which is different from that reported in [32].…”

Artificial intelligence in the diagnosis of COVID-19: challenges and perspectives

“…The potential applications of ML for COVID-19 have been previously described [14,[16][17][18][19][20][21][22][23][24][25][26]. The details are summarized in Table 1.…”

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

“…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.…”