Diagnostic accuracy estimates for COVID-19 RT-PCR and Lateral flow immunoassay tests with Bayesian latent class models

Abstract The objective of this work was to estimate the diagnostic accuracy of RT-PCR and Lateral flow immunoassay tests (LFIA) for COVID-19, depending on the time post symptom onset. Based on the cross-classified results of RT-PCR and LFIA, we used Bayesian latent class models (BLCMs), which do not require a gold standard for the evaluation of diagnostics. Data were extracted from studies that evaluated LFIA (IgG and/or IgM) assays using RT-PCR as the reference method. The cross-classified results of LFIA and RT-PCR were analysed separately for the first, second and third week post symptom onset. The SeRT-PCR was 0.695 (95% probability intervals: 0.563; 0.837) for the first week and remained similar for the second and the third week. The SeIgG/M was 0.318 (0.229; 0.416) for the first week and increased steadily. It was 0.755 (0.673; 0.829) and 0.927 (0.881; 0.965) for the second and third week, respectively. Both tests had a high to absolute Sp, with point median estimates for SpRT-PCR being consistently higher. SpRT-PCR was 0.990 (0.980; 0.998) for the first week. The corresponding value for SpIgG/M was 0.962 (0.905; 0.998). Further, Sp estimates for each test did not differ between weeks. BLCMs provide a valid and efficient alternative for evaluating the rapidly evolving diagnostics for COVID-19, under various clinical settings and for different risk profiles.

Download Full-text

STARD-BLCM: Standards for the Reporting of Diagnostic accuracy studies that use Bayesian Latent Class Models

Preventive Veterinary Medicine ◽

10.1016/j.prevetmed.2017.01.006 ◽

2017 ◽

Vol 138 ◽

pp. 37-47 ◽

Cited By ~ 65

Author(s):

Polychronis Kostoulas ◽

Søren S. Nielsen ◽

Adam J. Branscum ◽

Wesley O. Johnson ◽

Nandini Dendukuri ◽

...

Keyword(s):

Diagnostic Accuracy ◽

Latent Class ◽

Latent Class Models ◽

Class Models

Download Full-text

Point-Of-Care Clinical Evaluation of the Clungene® SARS-CoV-2 Virus IgG/IgM 15-minute rapid test cassette with the Cobas® Roche RT-PCR platform in patients with or without Covid-19

LymphoSign Journal ◽

10.14785/lymphosign-2021-0017 ◽

2021 ◽

Author(s):

Fadi Haddad ◽

Christopher C Lamb ◽

Ravina Kullar ◽

George Sakoulas

Keyword(s):

Symptom Onset ◽

Serological Test ◽

Point Of Care ◽

Diagnostic Testing ◽

Rapid Test ◽

Lateral Flow ◽

Added Value ◽

Lateral Flow Immunoassay ◽

Rt Pcr ◽

Past Infection

Background: Covid-19 remains a pandemic with multiple challenges to confirm patient infectivity: lack of sufficient tests, accurate results, validated quality, and timeliness of results. We hypothesize that a rapid 15-minute Point-Of-Care serological test to evaluate past infection complements diagnostic testing for Covid-19 and significantly enhances testing availability. Method: A three arm observational study at Sharp Healthcare, San Diego, California was conducted using the Clungene® lateral flow immunoassay (LFI) and compared with the Cobas® Roche RT PCR results. Arm 1: Thirty-five (35) subjects with confirmed Covid-19 using RT-PCR were tested twice: prior to 14 days following symptom onset and once between 12 and 70 days. Arm 2: Thirty (30) subjects with confirmed Covid-19 using RT-PCR were tested 12-70 days post symptom onset. Arm 3: Thirty (30) subjects with a negative RT-PCR for Covid-19 were tested 1-10 days following the RT-PCR test date. Results: Specificity of confirmed negative Covid-19 by RT-PCR was 100% (95% CI, 88.4%-100.0%); meaning there was 100% negative positive agreement between the RT-PCR and the Clungene® serological test results. Covid-19 subjects tested prior to day 7 symptom onset were antibody negative. In subjects 7-12 days following symptom onset with a confirmed positive Covid-19 by RT-PCR, the combined sensitivity of IgM and IgG was 58.6% (95% CI, 38.9%-76.5%). In subjects 13-70 days following symptom onset with a confirmed positive Covid-19 by RT-PCR the combined sensitivity of IgM and IgG was 90.5% (95% CI, 80.4%-96.4%). Conclusion: The Clungene® lateral flow immunoassay (LFI) is a useful tool to confirm individuals with an adaptive immune response to SARS-CoV-2 indicating past infection. Providing Point-Of-Care results within 15 minutes without any laboratory instrumentation or specialized software has an added value of increasing test availability to patients who have been symptomatic for more than one week to confirm past infection. Performance characteristics are optimal after 13 days with a sensitivity and specificity of 90% and 100%, respectively.

Download Full-text

Different latent class models were used and evaluated for assessing the accuracy of campylobacter diagnostic tests: overcoming imperfect reference standards?

Epidemiology and Infection ◽

10.1017/s0950268818001723 ◽

2018 ◽

Vol 146 (12) ◽

pp. 1556-1564 ◽

Cited By ~ 2

Author(s):

J. Asselineau ◽

A. Paye ◽

E. Bessède ◽

P. Perez ◽

C. Proust-Lima

Keyword(s):

Diagnostic Accuracy ◽

Diagnostic Tests ◽

Goodness Of Fit ◽

Latent Class ◽

Latent Class Models ◽

Reference Standard ◽

Promising Alternative ◽

Data Sparseness ◽

Campylobacter Infection ◽

Class Models

AbstractIn the absence of perfect reference standard, classical techniques result in biased diagnostic accuracy and prevalence estimates. By statistically defining the true disease status, latent class models (LCM) constitute a promising alternative. However, LCM is a complex method which relies on parametric assumptions, including usually a conditional independence between tests and might suffer from data sparseness. We carefully applied LCMs to assess new campylobacter infection detection tests for which bacteriological culture is an imperfect reference standard. Five diagnostic tests (culture, polymerase chain reaction and three immunoenzymatic tests) of campylobacter infection were collected in 623 patients from Bordeaux and Lyon Hospitals, France. Their diagnostic accuracy were estimated with standard and extended LCMs with a thorough examination of models goodness-of-fit. The model including a residual dependence specific to the immunoenzymatic tests best complied with LCM assumptions. Asymptotic results of goodness-of-fit statistics were substantially impaired by data sparseness and empirical distributions were preferred. Results confirmed moderate sensitivity of the culture and high performances of immunoenzymatic tests. LCMs can be used to estimate diagnostic tests accuracy in the absence of perfect reference standard. However, their implementation and assessment require specific attention due to data sparseness and limitations of existing software.

Download Full-text

Continued controversy in using latent class models for estimating diagnostic accuracy without a gold standard

Statistics in Medicine ◽

10.1002/sim.9085 ◽

2021 ◽

Vol 40 (22) ◽

pp. 4764-4765

Author(s):

Paul S. Albert

Keyword(s):

Diagnostic Accuracy ◽

Gold Standard ◽

Latent Class ◽

Latent Class Models ◽

Class Models ◽

Without A Gold Standard

Download Full-text

Better than we thought? The diagnostic performance of an influenza point-of-care test in children, a Bayesian re-analysis

F1000Research ◽

10.12688/f1000research.10068.1 ◽

2017 ◽

Vol 6 ◽

pp. 53

Author(s):

Joseph Lee

Keyword(s):

Diagnostic Accuracy ◽

Influenza A ◽

Diagnostic Performance ◽

Latent Class ◽

Point Of Care ◽

Credible Interval ◽

Latent Class Models ◽

Rt Pcr ◽

Point Of Care Test ◽

Low Sensitivity

Background: Point-of-care tests (POCTs) for influenza have been criticised for their diagnostic accuracy, with clinical use limited by low sensitivity. These criticisms are based on diagnostic-accuracy studies that often use the questionable assumption of an infallible gold standard. Bayesian latent class modelling can estimate diagnostic performance without this assumption. Methods: Data extracted from published diagnostic-accuracy studies comparing the QuickVue® influenza A+B influenza POCT to reverse-transcriptase polymerase chain reaction (RT-PCR) in two different populations were re-analysed. Classical and Bayesian latent class methods were applied using the Modelling for Infectious diseases CEntre (MICE) web-based application. Results: Under classical analyses the estimated sensitivity and specificity of the QuickVue® were 66.9% (95% confidence interval (CI) 61.4-71.9) and 97.8% (95% CI 95.7-98.9), respectively. Bayesian latent class models estimated sensitivity of 97.8% (95% credible interval (CrI) 82.1-100) and specificity of 98.5% (95% CrI 96.5-100). Conclusions: Data from studies comparing the QuickVue® point-of-care test to RT-PCR are compatible with better diagnostic performance than previously reported.

Download Full-text

Reporting guidelines for diagnostic accuracy studies that use Bayesian latent class models (STARD-BLCM)

Statistics in Medicine ◽

10.1002/sim.7316 ◽

2017 ◽

Vol 36 (23) ◽

pp. 3603-3604 ◽

Cited By ~ 2

Author(s):

Polychronis Kostoulas ◽

Søren S. Nielsen ◽

Adam J. Branscum ◽

Wesley O. Johnson ◽

Nandini Dendukuri ◽

...

Keyword(s):

Diagnostic Accuracy ◽

Latent Class ◽

Latent Class Models ◽

Reporting Guidelines ◽

Class Models

Download Full-text

Modeling the Accuracy of Three Detection Methods of Grapevine leafroll-associated virus 3 During the Dormant Period Using a Bayesian Approach

Phytopathology ◽

10.1094/phyto-10-15-0246-r ◽

2016 ◽

Vol 106 (5) ◽

pp. 510-518 ◽

Cited By ~ 4

Author(s):

Antonio Olmos ◽

Edson Bertolini ◽

Ana B. Ruiz-García ◽

Carmen Martínez ◽

Rosa Peiró ◽

...

Keyword(s):

Nucleic Acid ◽

Real Time ◽

Latent Class ◽

Latent Class Models ◽

Rt Pcr ◽

Likelihood Ratios ◽

Dormant Period ◽

Degree Of Confidence ◽

Class Models ◽

Das Elisa

Grapevine leafroll-associated virus 3 (GLRaV-3) has a worldwide distribution and is the most economically important virus that causes grapevine leafroll disease. Reliable, sensitive, and specific methods are required for the detection of the pathogen in order to assure the production of healthy plant material and control of the disease. Although different serological and nucleic acid-based methods have been developed for the detection of GLRaV-3, diagnostic parameters have not been established, and there is no gold standard method. Therefore, the main aim of this work was to determine the sensitivity, specificity, and likelihood ratios of three commonly used methods, including one serological test (double-antibody sandwich enzyme-linked immunosorbent assay [DAS-ELISA]) and two nucleic acid-based techniques (spot and conventional real-time reverse transcription-polymerase chain reaction [RT-PCR]). Latent class models using a Bayesian approach have been applied to determine diagnostic test parameters and to facilitate decision-making regarding diagnostic test selection. Statistical analysis has been based on the results of a total of 281 samples, which were collected during the dormant period from three different populations. The best-fit model out of the 49 implemented models revealed that DAS-ELISA was the most specific method (value = 0.99) and provided the highest degree of confidence in positive results. Conversely, conventional real-time RT-PCR was the most sensitive method (value = 0.98) and produced the highest degree of confidence in negative results. Furthermore, the estimation of likelihood ratios showed that in populations with low GLRaV-3 prevalence the most appropriate method could be DAS-ELISA, while conventional real-time RT-PCR could be the most appropriate method in medium or high prevalence populations. Combining both techniques significantly increases detection accuracy. The flexibility and power of Bayesian latent class models open new possibilities for the evaluation of diagnostic tests for plant viruses.

Download Full-text