A machine learning algorithm to increase COVID-19 inpatient diagnostic capacity

Goodman‐Meza, David; Rudas, Ákos; Chiang, Jeffrey N.; Adamson, Paul C; Ebinger, Joseph; Sun, Nancy; Botting, Patrick; Fulcher, Jennifer A.; Saab, Faysal; Brook, Rachel; Eskin, Eleazar; An, Ulzee; Kordi, Misagh; Jew, Brandon; Balliu, Brunilda; Chen, Zeyuan; Hill, Brian; Rahmani, Elior; Halperin, Eran; Manuel, Vladimir

doi:10.1371/journal.pone.0239474

Cited by 63 publications

(57 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, few studies used laboratory, clinical, and imaging data together for COVID-19 diagnosis. Our results are in line with studies that used machine learning models based on clinico-biological variables for COVID-19 diagnosis 55 – 57 . The performances of these models were low, excepted for the model described by Plante and al.…”

Section: Discussionsupporting

confidence: 91%

Machine learning is the key to diagnose COVID-19: a proof-of-concept study

Gangloff

Rafi

Bouzillé

et al. 2021

Sci Rep

View full text Add to dashboard Cite

The reverse transcription-polymerase chain reaction (RT-PCR) assay is the accepted standard for coronavirus disease 2019 (COVID-19) diagnosis. As any test, RT-PCR provides false negative results that can be rectified by clinicians by confronting clinical, biological and imaging data. The combination of RT-PCR and chest-CT could improve diagnosis performance, but this would requires considerable resources for its rapid use in all patients with suspected COVID-19. The potential contribution of machine learning in this situation has not been fully evaluated. The objective of this study was to develop and evaluate machine learning models using routine clinical and laboratory data to improve the performance of RT-PCR and chest-CT for COVID-19 diagnosis among post-emergency hospitalized patients. All adults admitted to the ED for suspected COVID-19, and then hospitalized at Rennes academic hospital, France, between March 20, 2020 and May 5, 2020 were included in the study. Three model types were created: logistic regression, random forest, and neural network. Each model was trained to diagnose COVID-19 using different sets of variables. Area under the receiving operator characteristics curve (AUC) was the primary outcome to evaluate model’s performances. 536 patients were included in the study: 106 in the COVID group, 430 in the NOT-COVID group. The AUC values of chest-CT and RT-PCR increased from 0.778 to 0.892 and from 0.852 to 0.930, respectively, with the contribution of machine learning. After generalization, machine learning models will allow increasing chest-CT and RT-PCR performances for COVID-19 diagnosis.

show abstract

Section: Discussionsupporting

confidence: 91%

Machine learning is the key to diagnose COVID-19: a proof-of-concept study

Gangloff

Rafi

Bouzillé

et al. 2021

Sci Rep

View full text Add to dashboard Cite

show abstract

“…Recently, an algorithm based on laboratory and demographic features was proposed by Goodman-Meza D. et al [ 48 ] to serve as a screening method in hospitals where testing is limited or inaccessible. The methodology used data obtained retrospectively from the UCLA Health System in Los Angeles, California.…”

Section: Related Workmentioning

confidence: 99%

SOM-LWL method for identification of COVID-19 on chest X-rays

et al. 2021

View full text Add to dashboard Cite

The outbreak of coronavirus disease 2019 (COVID-19) has had an immense impact on world health and daily life in many countries. Sturdy observing of the initial site of infection in patients is crucial to gain control in the struggle with COVID-19. The early automated detection of the recent coronavirus disease (COVID-19) will help to limit its dissemination worldwide. Many initial studies have focused on the identification of the genetic material of coronavirus and have a poor detection rate for long-term surgery. The first imaging procedure that played an important role in COVID-19 treatment was the chest X-ray. Radiological imaging is often used as a method that emphasizes the performance of chest X-rays. Recent findings indicate the presence of COVID-19 in patients with irregular findings on chest X-rays. There are many reports on this topic that include machine learning strategies for the identification of COVID-19 using chest X-rays. Other current studies have used non-public datasets and complex artificial intelligence (AI) systems. In our research, we suggested a new COVID-19 identification technique based on the locality-weighted learning and self-organization map (LWL-SOM) strategy for detecting and capturing COVID-19 cases. We first grouped images from chest X-ray datasets based on their similar features in different clusters using the SOM strategy in order to discriminate between the COVID-19 and non-COVID-19 cases. Then, we built our intelligent learning model based on the LWL algorithm to diagnose and detect COVID-19 cases. The proposed SOM-LWL model improved the correlation coefficient performance results between the Covid19, no-finding, and pneumonia cases; pneumonia and no-finding cases; Covid19 and pneumonia cases; and Covid19 and no-finding cases from 0.9613 to 0.9788, 0.6113 to 1 0.8783 to 0.9999, and 0.8894 to 1, respectively. The proposed LWL-SOM had better results for discriminating COVID-19 and non-COVID-19 patients than the current machine learning-based solutions using AI evaluation measures.

show abstract

“…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.…”

Section: Degrading Of Predictive Performance Over Timementioning

confidence: 99%

“…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] .…”

Section: Introductionmentioning

confidence: 99%

“…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.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

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

Roland

Böck

Tschoellitsch

et al. 2021

Preprint

View full text Add to dashboard Cite

We investigate machine learning models that identify COVID-19 positive patients and estimate the mortality risk based on routinely acquired blood tests in a hospital setting. However, during pandemics or new outbreaks, disease and testing characteristics change, thus we face domain shifts. Domain shifts can be caused, e.g., by changes in the disease prevalence (spreading or tested population), by refined RT-PCR testing procedures (taking samples, laboratory), or by virus mutations. Therefore, machine learning models for diagnosing COVID-19 or other diseases may not be reliable and degrade in performance over time. To countermand this effect, we propose methods that first identify domain shifts and then reverse their negative effects on the model performance. Frequent re-training and re-assessment, as well as stronger weighting of more recent samples, keeps model performance and credibility at a high level over time. Our diagnosis models are constructed and tested on large-scale data sets, steadily adapt to observed domain shifts, and maintain high ROC AUC values along pandemics.

show abstract

A machine learning algorithm to increase COVID-19 inpatient diagnostic capacity

Cited by 63 publications

References 15 publications

Machine learning is the key to diagnose COVID-19: a proof-of-concept study

Machine learning is the key to diagnose COVID-19: a proof-of-concept study

SOM-LWL method for identification of COVID-19 on chest X-rays

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

Contact Info

Product

Resources

About