1
|
Gunčar G, Kukar M, Smole T, Moškon S, Vovko T, Podnar S, Černelč P, Brvar M, Notar M, Köster M, Jelenc MT, Osterc Ž, Notar M. Differentiating viral and bacterial infections: A machine learning model based on routine blood test values. Heliyon 2024; 10:e29372. [PMID: 38644832 PMCID: PMC11033127 DOI: 10.1016/j.heliyon.2024.e29372] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/18/2023] [Revised: 04/02/2024] [Accepted: 04/07/2024] [Indexed: 04/23/2024] Open
Abstract
The growing threat of antibiotic resistance necessitates accurate differentiation between bacterial and viral infections for proper antibiotic administration. In this study, a Virus vs. Bacteria machine learning model was developed to distinguish between these infection types using 16 routine blood test results, C-reactive protein concentration (CRP), biological sex, and age. With a dataset of 44,120 cases from a single medical center, the model achieved an accuracy of 82.2 %, a sensitivity of 79.7 %, a specificity of 84.5 %, a Brier score of 0.129, and an area under the ROC curve (AUC) of 0.905, outperforming a CRP-based decision rule. Notably, the machine learning model enhanced accuracy within the CRP range of 10-40 mg/L, a range where CRP alone is less informative. These results highlight the advantage of integrating multiple blood parameters in diagnostics. The "Virus vs. Bacteria" model paves the way for advanced diagnostic tools, leveraging machine learning to optimize infection management.
Collapse
Affiliation(s)
- Gregor Gunčar
- Smart Blood Analytics Swiss SA, CH-8008, Zürich, Switzerland
- Faculty of Chemistry and Chemical Technology, University of Ljubljana, Slovenia
| | - Matjaž Kukar
- Smart Blood Analytics Swiss SA, CH-8008, Zürich, Switzerland
- Faculty of Computer and Information Science, University of Ljubljana, Slovenia
| | - Tim Smole
- Smart Blood Analytics Swiss SA, CH-8008, Zürich, Switzerland
| | - Sašo Moškon
- Smart Blood Analytics Swiss SA, CH-8008, Zürich, Switzerland
| | - Tomaž Vovko
- Department of Infectious Diseases, University Medical Centre Ljubljana, Slovenia
| | - Simon Podnar
- Division of Neurology, University Medical Centre Ljubljana, Slovenia
| | - Peter Černelč
- Smart Blood Analytics Swiss SA, CH-8008, Zürich, Switzerland
| | - Miran Brvar
- Centre for Clinical Toxicology and Pharmacology, University Medical Centre Ljubljana, Slovenia
| | - Mateja Notar
- Smart Blood Analytics Swiss SA, CH-8008, Zürich, Switzerland
| | - Manca Köster
- Smart Blood Analytics Swiss SA, CH-8008, Zürich, Switzerland
| | | | - Žiga Osterc
- Smart Blood Analytics Swiss SA, CH-8008, Zürich, Switzerland
| | - Marko Notar
- Smart Blood Analytics Swiss SA, CH-8008, Zürich, Switzerland
| |
Collapse
|
2
|
Pallon J, Sundqvist M, Hedin K. The use and usefulness of point-of-care tests in patients with pharyngotonsillitis - an observational study in primary health care. BMC PRIMARY CARE 2024; 25:15. [PMID: 38184547 PMCID: PMC10770901 DOI: 10.1186/s12875-023-02245-9] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Subscribe] [Scholar Register] [Received: 02/07/2023] [Accepted: 12/13/2023] [Indexed: 01/08/2024]
Abstract
BACKGROUND Rapid antigen detection tests (RADT) for Group A streptococci (GAS) and point-of-care tests for C-reactive protein (CRP) are commonly used in patients with pharyngotonsillitis in Sweden and Denmark although CRP testing is not supported by guidelines. We aimed to describe (1) the proportion of patients tested with RADT and/or CRP, (2) the relation between test results and antibiotic prescribing, and (3) the association between CRP level and microbial aetiology. METHODS We used a post-hoc-analysis of data collected in primary health care in a prospective aetiological study of 220 patients 15-45 years old diagnosed with pharyngotonsillitis. The outcomes of RADTs and CRP tests were related to antibiotic prescribing and microbial aetiology. RESULTS A RADT was used in 94% of the patients. A CRP test was used in 50% of the patients but more commonly in those with a negative RADT (59%) than in those with a positive RADT (38%) (p = 0.005). Most (74%) CRP tests were used in patients with a negative RADT. Antibiotic prescribing differed greatly between patients with a positive RADT (96%) and patients with a negative RADT (17%) (p < 0.001). In patients with a negative RADT, there was a positive association between CRP value and antibiotic prescribing (OR 1.05; 95% CI 1.02-1.07; p < 0.001). Patients with CRP values ≤ 30 mg/l were seldomly prescribed antibiotics. Patients with GAS in culture had the highest median CRP (46 mg/l), which was higher than in patients without GAS (8 mg/l; p < 0.001). However, the positive predictive value for GAS never exceeded 0.60 (95% CI 0.31-0.83) at the investigated CRP levels. CONCLUSIONS The widespread use of tests is a major deviation from national guidelines. Most CRP tests were used in patients with a negative RADT, suggesting a belief in the added value of a CRP test, and the CRP result seemed to influence antibiotic prescribing. However, as an aetiological test, CRP is not useful for predicting GAS.
Collapse
Affiliation(s)
- Jon Pallon
- Department of Clinical Sciences in Malmö, Family Medicine, Lund University, Malmö, Sweden.
- Department of Research and Development, Region Kronoberg, Växjö, Sweden.
| | - Martin Sundqvist
- Department of Laboratory Medicine, Clinical Microbiology, Faculty of Medicine and Health, Örebro University, Örebro, Sweden
| | - Katarina Hedin
- Department of Clinical Sciences in Malmö, Family Medicine, Lund University, Malmö, Sweden
- Department of Health, Medicine and Caring Sciences, Futurum, Region Jönköping County, Linköping University, Linköping, Sweden
| |
Collapse
|
3
|
Calarco S, Fernandez-Carballo BL, Keller T, Weber S, Jakobi M, Marsall P, Schneiderhan-Marra N, Dittrich S. Analytical performance of 17 commercially available point-of-care tests for CRP to support patient management at lower levels of the health system. PLoS One 2023; 18:e0267516. [PMID: 36662693 PMCID: PMC9858008 DOI: 10.1371/journal.pone.0267516] [Citation(s) in RCA: 3] [Impact Index Per Article: 3.0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/08/2022] [Accepted: 01/02/2023] [Indexed: 01/21/2023] Open
Abstract
Accurate and precise point-of-care (POC) testing for C-reactive protein (CRP) can help support healthcare providers in the clinical management of patients. Here, we compared the analytical performance of 17 commercially available POC CRP tests to enable more decentralized use of the tool. The following CRP tests were evaluated. Eight quantitative tests: QuikRead go (Aidian), INCLIX (Sugentech), Spinit (Biosurfit), LS4000 (Lansionbio), GS 1200 (Gensure Biotech), Standard F200 (SD Biosensor), Epithod 616 (DxGen), IFP-3000 (Xincheng Biological); and nine semi-quantitative tests: Actim CRP (ACTIM), NADAL Dipstick (nal von minden), NADAL cassette (nal von minden), ALLTEST Dipstick (Hangzhou Alltest Biotech), ALLTEST Cassette cut-off 10-40-80 (Hangzhou Alltest Biotech), ALLTEST Cassette cut-off 10-30 (Hangzhou Alltest Biotech), Biotest (Hangzhou Biotest Biotech), BTNX Quad Line (BTNX), BTNX Tri Line (BTNX). Stored samples (n = 660) had previously been tested for CRP using Cobas 8000 Modular analyzer (Roche Diagnostics International AG, Rotkreuz, Switzerland (reference standards). CRP values represented the clinically relevant range (10-100 mg/L) and were grouped into four categories (<10 mg/L, 10-40 mg/L or 10-30 mg/L, 40-80 mg/L or 30-80 mg/L, and > 80mg/L) for majority of the semi-quantitative tests. Among the eight quantitative POC tests evaluated, QuikRead go and Spinit exhibited better agreement with the reference method, showing slopes of 0.963 and 0.921, respectively. Semi-quantitative tests with the four categories showed a poor percentage agreement for the intermediate categories and higher percentage agreement for the lower and upper limit categories. Analytical performance varied considerably for the semi-quantitative tests, especially among the different categories of CRP values. Our findings suggest that quantitative tests might represent the best choice for a variety of use cases, as they can be used across a broad range of CRP categories.
Collapse
Affiliation(s)
| | | | | | | | - Meike Jakobi
- NMI Natural and Medical Sciences Institute at the University of Tübingen, Reutlingen, Germany
| | - Patrick Marsall
- NMI Natural and Medical Sciences Institute at the University of Tübingen, Reutlingen, Germany
| | | | | |
Collapse
|
4
|
The Role of CRP POC Testing in the Fight against Antibiotic Overuse in European Primary Care: Recommendations from a European Expert Panel. Diagnostics (Basel) 2023; 13:diagnostics13020320. [PMID: 36673130 PMCID: PMC9857389 DOI: 10.3390/diagnostics13020320] [Citation(s) in RCA: 6] [Impact Index Per Article: 6.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/28/2022] [Revised: 01/04/2023] [Accepted: 01/09/2023] [Indexed: 01/18/2023] Open
Abstract
Tackling antibiotic resistance represents one of the major challenges in modern medicine, and limiting antibiotics' overuse represents the first step in this fight. Most antibiotics are prescribed in primary care settings, and lower respiratory tract infections (LRTIs) are one of the most common indications for their prescription. An expert panel conducted an extensive report on C-reactive protein point-of-care (CRP POC) testing in the evaluation of LRTIs and its usefulness to limit antibiotic prescriptions. The expert panel stated that CRP POC testing is a potentially useful tool to limit antibiotic prescriptions for LRTI in a community setting. CRP POC must be used in conjunction with other strategies such as improved communication skills and the use of other molecular POC testing. Potential barriers to the adoption of CRP POC testing are financial and logistical issues. Moreover, the efficacy in limiting antibiotic prescriptions could be hampered by the fact that, in some countries, patients may gain access to antibiotics even without a prescription. Through the realization of a better reimbursement structure, the inclusion in standardized procedures in local guidelines, and better patient education, CRP point-of-care testing can represent a cornerstone in the fight against antimicrobial resistance.
Collapse
|
5
|
Sydenham RV, Hansen MP, Justesen US, Pedersen LB, Aabenhus RM, Wehberg S, Jarbøl DE. Factors associated with C-reactive protein testing when prescribing antibiotics in general practice: a register-based study. BMC PRIMARY CARE 2022; 23:17. [PMID: 35172735 PMCID: PMC8783519 DOI: 10.1186/s12875-021-01614-6] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.5] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Download PDF] [Subscribe] [Scholar Register] [Received: 01/30/2021] [Accepted: 12/14/2021] [Indexed: 11/10/2022]
Abstract
BACKGROUND The use of C-reactive protein (CRP) tests has been shown to safely reduce antibiotic prescribing for acute respiratory tract infections (RTIs). The aim of this study was to explore patient and clinical factors associated with the use of CRP testing when prescribing antibiotics recommended for RTIs. METHODS A nation-wide retrospective cross-sectional register-based study based on first redeemed antibiotic prescriptions issued to adults in Danish general practice between July 2015 and June 2017. Only antibiotics recommended for treatment of RTIs were included in the analysis (penicillin-V, amoxicillin, co-amoxicillin or roxithromycin/clarithromycin). Logistic regression models were used to estimate odds ratios for patient-related and clinical factors on performing a CRP test in relation to antibiotic prescribing. RESULTS A total of 984,149 patients redeemed at least one antibiotic prescription during the two-year period. About half of these prescriptions (49.6%) had an RTI stated as the indication, and a CRP test was performed in relation to 45.2% of these scripts. Lower odds of having a CRP test performed in relation to an antibiotic prescription was found for patients aged 75 years and above (OR 0.82, 95CI 0.79-0.86), with a Charlson Comorbidity Index of more than one (OR 0.93, 95CI 0.91-0.95), unemployed or on disability pension (OR 0.84, 95CI 0.83-0.85) and immigrants (OR 0.91, 95CI 0.88-0.95) or descendants of immigrants (OR 0.90, 95CI 0.84-0.96). Living with a partner (OR 1.08, 95CI 1.07-1.10), being followed in practice for a chronic condition (OR 1.22, 95CI 1.18-1.26) and having CRP tests performed in the previous year (OR 1.78, 95CI 1.73-1.84) were associated with higher odds of CRP testing in relation to antibiotic prescribing. CONCLUSIONS Differences were observed in the use of CRP tests among subgroups of patients indicating that both sociodemographic factors and comorbidity influence the decision to use a CRP test in relation to antibiotic prescriptions in general practice. Potentially, this means that the use of CRP tests could be optimised to increase diagnostic certainty and further promote rational prescribing of antibiotics. The rationale behind the observed differences could be further explored in future qualitative studies.
Collapse
Affiliation(s)
- Rikke Vognbjerg Sydenham
- Research Unit of General Practice, Institute of Public Health, University of Southern Denmark, JB Winsløws Vej 9A, 5000, Odense C, Denmark.
| | | | - Ulrik Stenz Justesen
- Department of Clinical Microbiology, Odense University Hospital, Odense, Denmark
| | - Line Bjørnskov Pedersen
- Research Unit of General Practice, Institute of Public Health, University of Southern Denmark, JB Winsløws Vej 9A, 5000, Odense C, Denmark
- Danish Centre for Health Economics, Institute of Public Health, University of Southern Denmark, Odense, Denmark
| | - Rune Munck Aabenhus
- Research Unit for General Practice, University of Copenhagen, Copenhagen, Denmark
| | - Sonja Wehberg
- Research Unit of General Practice, Institute of Public Health, University of Southern Denmark, JB Winsløws Vej 9A, 5000, Odense C, Denmark
| | - Dorte Ejg Jarbøl
- Research Unit of General Practice, Institute of Public Health, University of Southern Denmark, JB Winsløws Vej 9A, 5000, Odense C, Denmark
| |
Collapse
|