1
|
Dobrijević D, Vilotijević-Dautović G, Katanić J, Horvat M, Horvat Z, Pastor K. Rapid Triage of Children with Suspected COVID-19 Using Laboratory-Based Machine-Learning Algorithms. Viruses 2023; 15:1522. [PMID: 37515208 PMCID: PMC10383367 DOI: 10.3390/v15071522] [Citation(s) in RCA: 2] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/14/2023] [Revised: 07/05/2023] [Accepted: 07/06/2023] [Indexed: 07/30/2023] Open
Abstract
In order to limit the spread of the novel betacoronavirus (SARS-CoV-2), it is necessary to detect positive cases as soon as possible and isolate them. For this purpose, machine-learning algorithms, as a field of artificial intelligence, have been recognized as a promising tool. The aim of this study was to assess the utility of the most common machine-learning algorithms in the rapid triage of children with suspected COVID-19 using easily accessible and inexpensive laboratory parameters. A cross-sectional study was conducted on 566 children treated for respiratory diseases: 280 children with PCR-confirmed SARS-CoV-2 infection and 286 children with respiratory symptoms who were SARS-CoV-2 PCR-negative (control group). Six machine-learning algorithms, based on the blood laboratory data, were tested: random forest, support vector machine, linear discriminant analysis, artificial neural network, k-nearest neighbors, and decision tree. The training set was validated through stratified cross-validation, while the performance of each algorithm was confirmed by an independent test set. Random forest and support vector machine models demonstrated the highest accuracy of 85% and 82.1%, respectively. The models demonstrated better sensitivity than specificity and better negative predictive value than positive predictive value. The F1 score was higher for the random forest than for the support vector machine model, 85.2% and 82.3%, respectively. This study might have significant clinical applications, helping healthcare providers identify children with COVID-19 in the early stage, prior to PCR and/or antigen testing. Additionally, machine-learning algorithms could improve overall testing efficiency with no extra costs for the healthcare facility.
Collapse
Affiliation(s)
- Dejan Dobrijević
- Faculty of Medicine, University of Novi Sad, 21000 Novi Sad, Serbia
- Institute for Child and Youth Health Care of Vojvodina, 21000 Novi Sad, Serbia
| | - Gordana Vilotijević-Dautović
- Faculty of Medicine, University of Novi Sad, 21000 Novi Sad, Serbia
- Institute for Child and Youth Health Care of Vojvodina, 21000 Novi Sad, Serbia
| | - Jasmina Katanić
- Faculty of Medicine, University of Novi Sad, 21000 Novi Sad, Serbia
- Institute for Child and Youth Health Care of Vojvodina, 21000 Novi Sad, Serbia
| | - Mirjana Horvat
- Faculty of Civil Engineering Subotica, University of Novi Sad, 24000 Subotica, Serbia
| | - Zoltan Horvat
- Faculty of Civil Engineering Subotica, University of Novi Sad, 24000 Subotica, Serbia
| | - Kristian Pastor
- Faculty of Technology, University of Novi Sad, 21000 Novi Sad, Serbia
| |
Collapse
|
2
|
Kumar Y, Dogra A, Kaushik A, Kumar S. Progressive evaluation in spectroscopic sensors for non-invasive blood haemoglobin analysis - a review. Physiol Meas 2021; 43. [PMID: 34883473 DOI: 10.1088/1361-6579/ac41b7] [Citation(s) in RCA: 2] [Impact Index Per Article: 0.7] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/17/2021] [Accepted: 12/09/2021] [Indexed: 11/11/2022]
Abstract
Frequent monitoring of haemoglobin concentration is highly recommended by physicians to diagnose anaemia and polycythemia Vera. Moreover, Some other conditions also demand assessment of haemoglobin, and these conditions are blood loss, before blood donation, during pregnancy, preoperative, perioperative and postoperative conditions. Cyanmethaemoglobin/haemiglobincyanide method, portable haemoglobinometers and haematology analyzers are few standard methods to diagnose mentioned ailments. However, discomfort, delay and risk of infection are typical limitations of traditional measuring solutions. These limitations create the necessity to develop a non-invasive haemoglobin monitoring technique for a better lifestyle. Various methods and products are already developed and popular due to their non-invasiveness; however, invasive solutions are still considered as the reference standard method. Therefore, this review summarizes the attributes of existing non-invasive solutions. These attributes are finalized as brief details, accuracy, optimal benefits, and research challenges for exploring potential gaps, advancements and possibilities to consider as futuristic alternative methodologies. Non-invasive total haemoglobin assessing techniques are mainly based on optical spectroscopy (reflectance/transmittance) or digital photography or spectroscopic imaging in spot check/continuous monitoring mode. In all these techniques, we have noticed that there is a need to consider different light conditions, motion artefacts, melanocytes, other blood constituents, smoking and precise fixing of the sensor from the sensing spot for exact formulation. Moreover, based on careful and critical analysis of outcomes, none of these techniques or products is used independently or intended to replace invasive laboratory testing. Therefore there is a requirement for a more accurate technique that can eliminate the requirement of blood samples and likely end up as a reference standard method.
Collapse
Affiliation(s)
- Yogesh Kumar
- Biomedical Instrumentation, CSIR Central Scientific Instruments Organisation, ., Chandigarh, 160030, INDIA
| | | | - Ajeet Kaushik
- Department of Natural Sciences, Florida Polytechnic University, 4700 Research Way, IST#2018, Lakeland, Florida, 33805, UNITED STATES
| | - Sanjeev Kumar
- Biomedical Instrumentation, CSIR Central Scientific Instruments Organisation, ., Chandigarh, 160020, INDIA
| |
Collapse
|
3
|
Niederwanger C, Varga T, Hell T, Stuerzel D, Prem J, Gassner M, Rickmann F, Schoner C, Hainz D, Cortina G, Hetzer B, Treml B, Bachler M. Comparison of pediatric scoring systems for mortality in septic patients and the impact of missing information on their predictive power: a retrospective analysis. PeerJ 2020; 8:e9993. [PMID: 33083117 PMCID: PMC7543722 DOI: 10.7717/peerj.9993] [Citation(s) in RCA: 9] [Impact Index Per Article: 2.3] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/06/2020] [Accepted: 08/28/2020] [Indexed: 01/08/2023] Open
Abstract
Background Scores can assess the severity and course of disease and predict outcome in an objective manner. This information is needed for proper risk assessment and stratification. Furthermore, scoring systems support optimal patient care, resource management and are gaining in importance in terms of artificial intelligence. Objective This study evaluated and compared the prognostic ability of various common pediatric scoring systems (PRISM, PRISM III, PRISM IV, PIM, PIM2, PIM3, PELOD, PELOD 2) in order to determine which is the most applicable score for pediatric sepsis patients in terms of timing of disease survey and insensitivity to missing data. Methods We retrospectively examined data from 398 patients under 18 years of age, who were diagnosed with sepsis. Scores were assessed at ICU admission and re-evaluated on the day of peak C-reactive protein. The scores were compared for their ability to predict mortality in this specific patient population and for their impairment due to missing data. Results PIM (AUC 0.76 (0.68-0.76)), PIM2 (AUC 0.78 (0.72-0.78)) and PIM3 (AUC 0.76 (0.68-0.76)) scores together with PRSIM III (AUC 0.75 (0.68-0.75)) and PELOD 2 (AUC 0.75 (0.66-0.75)) are the most suitable scores for determining patient prognosis at ICU admission. Once sepsis is pronounced, PELOD 2 (AUC 0.84 (0.77-0.91)) and PRISM IV (AUC 0.8 (0.72-0.88)) become significantly better in their performance and count among the best prognostic scores for use at this time together with PRISM III (AUC 0.81 (0.73-0.89)). PELOD 2 is good for monitoring and, like the PIM scores, is also largely insensitive to missing values. Conclusion Overall, PIM scores show comparatively good performance, are stable as far as timing of the disease survey is concerned, and they are also relatively stable in terms of missing parameters. PELOD 2 is best suitable for monitoring clinical course.
Collapse
Affiliation(s)
- Christian Niederwanger
- Department of Pediatrics, Pediatrics I, Medical University of Innsbruck, Innsbruck, Austria
| | - Thomas Varga
- Institute of Anaesthesiology, University of Zurich and University Hospital Zurich, Zurich, Switzerland
| | - Tobias Hell
- Department of Mathematics, Faculty of Mathematics, Computer Science and Physics, University of Innsbruck, Innsbruck, Austria
| | - Daniel Stuerzel
- Department of General and Surgical Critical Care Medicine, Medical University of Innsbruck, Innsbruck, Austria
| | - Jennifer Prem
- Department of General and Surgical Critical Care Medicine, Medical University of Innsbruck, Innsbruck, Austria
| | - Magdalena Gassner
- Department of General and Surgical Critical Care Medicine, Medical University of Innsbruck, Innsbruck, Austria
| | - Franziska Rickmann
- Department of General and Surgical Critical Care Medicine, Medical University of Innsbruck, Innsbruck, Austria
| | - Christina Schoner
- Department of General and Surgical Critical Care Medicine, Medical University of Innsbruck, Innsbruck, Austria
| | - Daniela Hainz
- Department of General and Surgical Critical Care Medicine, Medical University of Innsbruck, Innsbruck, Austria
| | - Gerard Cortina
- Department of Pediatrics, Pediatrics I, Medical University of Innsbruck, Innsbruck, Austria
| | - Benjamin Hetzer
- Department of Pediatrics, Pediatrics I, Medical University of Innsbruck, Innsbruck, Austria
| | - Benedikt Treml
- Department of General and Surgical Critical Care Medicine, Medical University of Innsbruck, Innsbruck, Austria
| | - Mirjam Bachler
- Department of General and Surgical Critical Care Medicine, Medical University of Innsbruck, Innsbruck, Austria.,Department of Sports Medicine, Alpine Medicine and Health Tourism, UMIT - University for Health Sciences, Medical Informatics and Technology, Hall in Tyrol, Austria
| |
Collapse
|
4
|
Noninvasive Measurement of Hemoglobin Using Spectrophotometry: Is it Useful for the Critically Ill Child? J Pediatr Hematol Oncol 2018; 40:e19-e22. [PMID: 29200161 DOI: 10.1097/mph.0000000000001038] [Citation(s) in RCA: 6] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Indexed: 11/27/2022]
Abstract
This study compared the accuracy of noninvasively measuring hemoglobin using spectrophotometry (SpHb) with a pulse CO-oximeter and laboratory hemoglobin (Hb) measurements. A total of 345 critically ill children were included prospectively. Age, sex, and factors influencing the reliabilityof SpHb such as SpO2, heart rate, perfusion index (PI), and vasoactive inotropic score were recorded. SpHb measurements were recorded during the blood draw and compared with the Hb measurement. Thirteen patients (low PI in 9 patients and no available Hb in 4 patients) were excluded and 332 children were eligible for final analysis. The mean Hb was 8.71±1.49 g/dL (range, 5.9 to 12 g/dL) and the mean SpHb level was 9.55±1.53 g/dL (range, 6 to 14.2 g/dL). The SpHb bias was 0.84±0.86,with the limits of agreement ranging from -2.5 to 0.9 g/dL. The difference between Hb and SpHb was >1.5 g/dL for only 47 patients. Of these, 24 patients had laboratory Hb levels <7 g/dL. There was a weak positive correlation between differences and PI (r=0.349; P= 0.032). The pulse CO-oximeter is a promising tool for measuring SpHb and monitoring critically ill children. However, PI may affect these results. Additional studies investigating the reliability of the trend of continuous SpHb values compared with simultaneously measured laboratory Hb values in the same patient are warranted.
Collapse
|
5
|
A Novel Needle-Free Blood Draw Device for Sample Collection From Short Peripheral Catheters. JOURNAL OF INFUSION NURSING 2017; 40:156-162. [PMID: 28419012 PMCID: PMC5400409 DOI: 10.1097/nan.0000000000000222] [Citation(s) in RCA: 9] [Impact Index Per Article: 1.3] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 12/02/2022]
Abstract
A new US Food and Drug Administration-cleared needleless blood collection device (PIVO; Velano Vascular, San Francisco, CA) for short peripheral catheters was compared with conventional venipuncture for collecting blood samples for routine laboratory analysis from adult healthy volunteers. The PIVO device was comparable with venipuncture in terms of providing high-integrity samples (no hemolysis or clotting), equivalent laboratory values, and better patient experience as assessed by pain scores. Further studies to assess the overall utility of the PIVO device are warranted.
Collapse
|
6
|
Phillips MR, Khoury AL, Bortsov AV, Marzinsky A, Short KA, Cairns BA, Charles AG, Joyner BL, McLean SE. A noninvasive hemoglobin monitor in the pediatric intensive care unit. J Surg Res 2015; 195:257-62. [PMID: 25724765 DOI: 10.1016/j.jss.2014.12.051] [Citation(s) in RCA: 17] [Impact Index Per Article: 1.9] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/24/2014] [Revised: 12/08/2014] [Accepted: 12/31/2014] [Indexed: 10/24/2022]
Abstract
BACKGROUND Critically ill pediatric patients frequently require hemoglobin monitoring. Accurate noninvasive Hb (SpHb) would allow practitioners to decrease anemia from repeated blood draws, traumatic blood draws, and a decreased number of laboratory Hb (LabHb) medical tests. The Food and Drug Administration has approved the Masimo Pronto SpHb and associated Rainbow probes; however, its use in the pediatric intensive care unit (PICU) is controversial. In this study, we define the degree of agreement between LabHb and SpHb using the Masimo Pronto SpHb Monitor and identify clinical and demographic conditions associated with decreased accuracy. MATERIALS AND METHODS We performed a prospective, observational study in a large PICU at an academic medical center. Fifty-three pediatric patients (30-d and 18-y-old), weighing >3 kg, admitted to the PICU from January-April 2013 were examined. SpHb levels measured at the time of LabHb blood draw were compared and analyzed. RESULTS Only 83 SpHb readings were obtained in 118 attempts (70.3%) and 35 readings provided a result of "unable to obtain." The mean LabHb and SpHb were 11.1 g/dL and 11.2 g/dL, respectively. Bland-Altman analysis showed a mean difference of 0.07 g/dL with a standard deviation of ±2.59 g/dL. Pearson correlation is 0.55, with a 95% confidence interval between 0.38 and 0.68. Logistic regression showed that extreme LabHb values, increasing skin pigmentation, and increasing body mass index were predictors of poor agreement between SpHb and LabHb (P < 0.05). Separately, increasing body mass index, hypoxia, and hypothermia were predictors for undetectable readings (P < 0.05). CONCLUSIONS The Masimo Pronto SpHb Monitor provides adequate agreement for the trending of hemoglobin levels in critically ill pediatric patients. However, the degree of agreement is insufficient to be used as the sole indicator for transfusion decisions and should be used in context of other clinical parameters to determine the need for LabHb in critically ill pediatric patients.
Collapse
Affiliation(s)
- Michael R Phillips
- Division of Pediatric Surgery, Department of Surgery, North Carolina Children's Hospital, University of North Carolina, Chapel Hill, North Carolina
| | - Amal L Khoury
- Division of Pediatric Surgery, Department of Surgery, North Carolina Children's Hospital, University of North Carolina, Chapel Hill, North Carolina
| | - Andrey V Bortsov
- Department of Anesthesiology, University of North Carolina, School of Medicine, Chapel Hill, North Carolina
| | - Amy Marzinsky
- Division of Pediatric Surgery, Department of Surgery, North Carolina Children's Hospital, University of North Carolina, Chapel Hill, North Carolina
| | - Kathy A Short
- Department of Respiratory Care and Pulmonary Diagnostics, University of North Carolina, Chapel Hill, North Carolina
| | - Bruce A Cairns
- Department of Surgery, University of North Carolina, School of Medicine, Chapel Hill, North Carolina
| | - Anthony G Charles
- Department of Surgery, University of North Carolina, School of Medicine, Chapel Hill, North Carolina
| | - Benny L Joyner
- Department of Pediatrics, Division of Critical Care Medicine, University of North Carolina, School of Medicine, Chapel Hill, North Carolina
| | - Sean E McLean
- Division of Pediatric Surgery, Department of Surgery, North Carolina Children's Hospital, University of North Carolina, Chapel Hill, North Carolina.
| |
Collapse
|