Bayesian Latent Class Model Research Articles

Bayesian latent class analysis of the characteristics of diagnostic tests to assess the passive immunity transfer status in neonatal Belgian Blue beef calves

The aim of the current study was to assess the diagnostic characteristics of radial immunodiffusion (RID), capillary electrophoresis (CE) and digital brix refractometry (Bx) for the diagnosis of failure of passive transfer (FPT) of immunity in neonatal Belgian Blue beef calves in the absence of a gold standard using a Bayesian latent class model. Belgian blue beef calves (n = 202) from a large farm in the south of Belgium were blood-sampled at 48–72 h of age and tested for FPT. The true prevalence of FPT in this population of calves was 34.5 % (95 % BCI: 26.1–44.3) using a FPT cut-off point of 10 g IgG/L. This true prevalence increased to 66.3 (95 % BCI: 56.9–74.8) and 88.9 % (95 % BCI: 83.1–93.2) at FPT cut-off points of respectively 18 and 25 g IgG/L serum. The Bland-Altman plot comparing the RID and CE methods, revealed that the serum IgG concentrations obtained by RID were on average 2.25 (95 % CI 1.62–2.88) g/L higher than those measured by CE. Optimal cut-off values for CE, corresponding to the FPT values as measured by RID of 10, 18, and 25 g IgG/L serum, were respectively 10, 15, and 20 g IgG/L. The overall diagnostic accuracy of the three diagnostic tests was comparable at the FPT cut-off point of 10 g IgG/L serum (i.e. 85 %). At higher cut-off points for FPT, the RID and CE assays presumably performed better that the Bx method. In conclusion, we demonstrated that: (1) the CE method is a good alternative for the RID assay, the latter having important constraints when considering its practicality, and (2) the Bx method is a cheap and user-friendly indirect method to evaluate the FPT in new-born Belgian Blue beef calves.

Improving the diagnosis of severe malaria in African children using platelet counts and plasma PfHRP2 concentrations.

Severe malaria caused by Plasmodium falciparum is difficult to diagnose accurately in children in high-transmission settings. Using data from 2649 pediatric and adult patients enrolled in four studies of severe illness in three countries (Bangladesh, Kenya, and Uganda), we fitted Bayesian latent class models using two diagnostic markers: the platelet count and the plasma concentration of P. falciparum histidine-rich protein 2 (PfHRP2). In severely ill patients with clinical features consistent with severe malaria, the combination of a platelet count of ≤150,000/μl and a plasma PfHRP2 concentration of ≥1000 ng/ml had an estimated sensitivity of 74% and specificity of 93% in identifying severe falciparum malaria. Compared with misdiagnosed children, pediatric patients with true severe malaria had higher parasite densities, lower hematocrits, lower rates of invasive bacterial disease, and a lower prevalence of both sickle cell trait and sickle cell anemia. We estimate that one-third of the children enrolled into clinical studies of severe malaria in high-transmission settings in Africa had another cause of their severe illness.

Bayesian estimation of sensitivity and specificity of fecal culture, fecal PCR and serum ELISA for diagnosis of Mycobacterium avium subsp. paratuberculosis infections in sheep

The objective of the present study was to evaluate the diagnostic accuracy of the individual fecal culture (IFC), fecal PCR (FPCR), and serum ELISA for the detection of Mycobacterium avium subsp. paratuberculosis (MAP) infections in sheep from four governorates in Egypt, using a latent class model (LCM) fitted within a Bayesian framework. Furthermore, the within-governorate prevalence of MAP infection in sheep was estimated as a secondary objective. Fecal and blood samples were collected from 370 sheep in four Egyptian governorates. Fecal samples were analyzed by IFC and RT-PCR based on ISMav2 gene, while ELISA was performed on serum samples. The sensitivity (Se) and specificity (Sp) of the three diagnostic tests were estimated using a three-tests-four-populations Bayesian LCM to obtain posterior estimates [medians and 95% Bayesian credible intervals (95% BCI)] for each parameter. The median Se estimates (95% BCI) for IFC, FPCR, and serum ELISA were 31.8% (22.8–41.4), 49.7% (31.8–79.9), and 61.2% (39.8–81.4), respectively. The median Sp estimates (95% BCI) for IFC, FPCR, and serum ELISA were 97.7% (96.1–98.9), 97.7% (95.6–99.5), and 98.4% (96.9–99.3), respectively. The median within-governorate paratuberculosis prevalence (95% BCI) was 5.2% (1.1–13.6), 8.4% (2.9–17.7), 9.4% (3.0–20.7), and 18.2% (10.5–29.5) for the Gharbia, Menoufia, Qalyubia, and Kafr El-Sheikh governorates, respectively. In conclusion, at a ratio of the optical density (OD) sample/OD positive control threshold of > 45%, ELISA showed the highest Se among the three tests and comparable Sp to IFC and FPCR. The test ELISA evaluated in this study is an interesting alternative for detecting MAP in sheep due to its higher Se, lower cost, and shorter turnaround laboratory time compared to IFC and FPCR.

Performance of Malnutrition Universal Screening Tool and Patient-Generated Global Subjective Assessment in screening for cancer-related malnutrition in Nairobi, Kenya

Background: Malnutrition is a common feature among oncology patients. It is responsible for poor response and tolerance to anticancer therapy, increased morbidity, and mortality. More than half of malnourished cancer patients remain undetected owing to lack of effective screening. Body mass index is the main indicator for assessing malnutrition in Kenyan public hospitals. However, it underestimates weight loss in patients with chronic illnesses. The Malnutrition Universal Screening Tool and Patient-Generated Subjective Global Assessment have been widely used in research and clinical practice and have both reported good validity and reliability. However, their diagnostic evaluation has not been performed in Kenya. Methods: A cross-sectional study was conducted among 138 and 76 cancer outpatients from Kenyatta National Hospital and Texas cancer treatment centres, respectively. Participants had a confirmed disease, stage 1-4 cancer, and aged 18 years and above. They were screened for malnutrition using both Malnutrition Universal Screening Tool and Patient Generated-Subjective Global Assessment. A separate study questionnaire was utilized to gather participant’s socio-demographic and clinical characteristics. A Bayesian latent class modelling framework was employed to infer the tests’ estimates based on participants ‘cumulative scores from the two tests. Results: The cut-off value of ≥ 1 and ≥ 4 gave the best combination of sensitivity and specificity of Malnutrition Universal Screening Tool and Patient Generated-Subjective Global Assessment. Both tests yielded statistically similar sensitivities and specificities. Predictive values were comparable between the tests and across the two populations. The posterior median true prevalences of malnutrition were high (˃ 54%) and numerically similar between the studied populations. Conclusions: The performance of both tests among patients with cancer is similar. Healthcare workers are therefore at liberty to use either of them to inform treatment. Given the high true prevalence of cancer-related malnutrition, routine screening is critical and should be included as part of cancer care.

Bayesian estimation of sensitivity and specificity of a milk pregnancy-associated glycoprotein ELISA test for pregnancy diagnosis between 23 and 27 days after insemination in Holstein dairy cows

Pregnancy diagnosis using pregnancy-associated glycoprotein (PAG) ELISA technology in blood or milk samples is validated from 28 d after insemination in dairy cows. The objective of this study was to estimate the sensitivity (Se) and specificity (Sp) of a commercial milk PAG-based ELISA in Holstein dairy cows between 23 and 27 d after insemination. Milk samples (n = 268) from 257 Holstein dairy cows 23 to 27 d after AI were submitted for PAG ELISA testing. Pregnancy status was confirmed by either a second milk PAG ELISA test conducted between 28 and 50 d after insemination (n = 200) or transrectal ultrasonography performed between 28 and 59 d after insemination (n = 68). A Bayesian latent class model was used to compare the paired results from the test at 23 to 27 d after AI test to the reference test. The latent class model typically used for comparing 2 or more imperfect tests was extended to include the possibility of pregnancy loss between the 23 to 27 d test and the reference test. Informative priors for the probability of pregnancy loss, and for the Se and Sp of the PAG and ultrasonography reference tests were obtained from the scientific literature. Estimated median Se and Sp of the PAG ELISA test conducted between 23 and 27 d after AI were 0.98 (95% credible interval 0.93 to 1.0) and 0.98 (0.89 to 1.0), respectively, when using a standardized corrected optical density threshold of 0.15. Although the accuracy of the test under investigation was excellent, more data will be needed to confirm the optimal diagnostic cut point for PAG in milk for early pregnancy diagnosis in this time window. The optimal timing of pregnancy diagnosis will depend on herd-specific logistics and the action to be taken to re-inseminate nonpregnant cows.

Meta-analysis of dichotomous and ordinal tests with an imperfect gold standard.

Standard methods for the meta‐analysis of medical tests, without assuming a gold standard, are limited to dichotomous data. Multivariate probit models are used to analyse correlated dichotomous data, and can be extended to model ordinal data. Within the context of an imperfect gold standard, they have previously been used for the analysis of dichotomous and ordinal test data from a single study, and for the meta‐analysis of dichotomous tests. However, they have not previously been used for the meta‐analysis of ordinal tests. In this article, we developed a Bayesian multivariate probit latent class model for the simultaneous meta‐analysis of ordinal and dichotomous tests without assuming a gold standard, which also allows one to obtain summary estimates of joint test accuracy. We fitted the models using the software Stan, which uses a state‐of‐the‐art Hamiltonian Monte Carlo algorithm, and we applied the models to a dataset in which studies evaluated the accuracy of tests, and test combinations, for deep vein thrombosis. We demonstrate the issues with dichotomising ordinal test accuracy data in the presence of an imperfect gold standard, before applying and comparing several variations of our proposed model which do not require the data to be dichotomised. The models proposed will allow researchers to more appropriately meta‐analyse ordinal and dichotomous tests without a gold standard, potentially leading to less biased estimates of test accuracy. This may lead to a better understanding of which tests, and test combinations, should be used for any given medical condition.

Estimation of the accuracy of an ELISA test applied to bulk tank milk for predicting herd-level status for Salmonella Dublin in dairy herds using Bayesian Latent Class Models

Enzyme-Linked Immunosorbent Assay (ELISA) test is commonly used for detection of antibodies to Salmonella Dublin in individual bovine milk samples. However, little is known about its accuracy when used on bulk tank milk for determining herd-level S. Dublin status and when evaluated without assuming a perfect reference test.The objectives of this study were: i) to estimate the herd prevalence of S. Dublin among dairy cattle herds in Québec, Canada; ii) to estimate the herd sensitivity and specificity of a commercially available ELISA test when used on bulk milk; iii) to examine how the diagnostic test accuracy varies with different bulk milk ELISA cut-offs; and (iv) to assess the added value of combining ELISA screening of bulk milk and individual serum of 10 animals for determining S. Dublin herd status.A cohort of 302 dairy herds selected in three regions (population 1) and 58 herds that have already tested positive to S. Dublin (population 2) were recruited. A total of 715 bulk milk samples and 7150 individual blood samples from cattle over 3 months old (10 animals per herd) sampled on two occasions were collected. Testing was conducted using PrioCHECK™ Salmonella Ab bovine Dublin ELISA test for milk (Bmilk ELISA: test under investigation) and for serum of 10 individual animals (Serum10 ELISA: imperfect reference test) to determine the herd-level S. Dublin status. A latent class model for two populations, two tests, allowing for conditional dependence between tests was fit within a Bayesian framework. At cut-off PP % ≥ 15 for a Bmilk ELISA, which is used by provincial authorities, the herd prevalence of S. Dublin estimated using informative prior was 6.8 % (4.3–9.9) in population 1. The herd sensitivity and specificity estimates (95 % Bayesian Credibility Intervals) for Bmilk ELISA were 40.6 % (15.6–88.8) and 91.9 % (88.3–95.8), respectively. Positive and negative predictive values of Bmilk ELISA applied in population 1 were 26.4 % (8.5–60.2) and 95.8 % (92.1–99.2), respectively. Increasing Bmilk ELISA cut-offs had little influence on predictive values. The combination of both ELISA tests did not improve the diagnostic accuracy of S. Dublin. Our study shows that a test-positive herd based on a single bulk milk sample would require complementary tests for status confirmation. However, a test-negative herd could be classified as true negative with a high certainty.

Leishmania exposure in dogs from two endemic countries from New and Old Worlds (Brazil and Portugal): evaluation of three serological tests using Bayesian Latent Class Models

BackgroundZoonotic leishmaniosis caused by Leishmania infantum is endemic in several countries of the Mediterranean Basin, Latin America, and Asia. Dogs are the main hosts and reservoirs of human infection. Thus, from a One Health perspective, early diagnosis of Leishmania infection in dogs is essential to control the dissemination of the parasite among other dogs and to humans. The aim of this study was to estimate the diagnosis accuracy of three serological tests to detect antibodies to Leishmania in dogs from two endemic settings using Bayesian latent class models (BLCM).MethodsA total of 378 dogs from two Portuguese and Brazilian endemic areas of leishmaniosis (194 animals from Portugal and 184 from Brazil) were screened. Detection of anti-Leishmania antibodies was performed using two commercial ELISA (L. infantum IgG-ELISA® and EIE-LVC®) and a rapid immunochromatographic test (DPP-LVC®). Bayesian latent class models were used to estimate Leishmania infection prevalence, together with sensitivities and specificities of the three diagnostic tests, in the two dog populations simultaneously. Predictive values were also calculated. Credibility intervals (CI) were obtained, considering different types of prior information.ResultsA posterior median Leishmania seroprevalence of 13.4% (95% CI 9.0–18.7) and of 21.6% (15.0–28.3) was estimated to the Portuguese and Brazilian dog subpopulations, respectively. The Bayesian analysis indicated that all tests were highly specific (specificity above 90%), and that the DPP-LVC® was more sensitive (96.6%; 83.1–99.9) than both ELISAs in the Portuguese subpopulation, while in the Brazilian subpopulation, EIE-LVC® and L. infantum IgG-ELISA®, had the highest sensitivity (88.2%; 73.7–97.0) and specificity (98.7%; 95.1–99.9), respectively.ConclusionsIn general, the levels of diagnosis accuracy of the three serological tests to detect Leishmania antibodies assessed by BLCM indicate their utility in canine epidemiological studies. The same approach should be used to assess the performance of these techniques in the clinical management of infected and sick dogs using representative samples from the wide spectrum of clinical situations, namely from subclinical infection to manifest disease. The low positive predictive value of the serological tests used in the current protocol of the Brazilian Ministry of Health suggests that they should not be used individually and may not be sufficient to target reservoir-based control interventions.Graphical

Diagnosis of Human Leptospirosis: Comparison of Microscopic Agglutination Test with Recombinant LigA/B Antigen-Based In-House IgM Dot ELISA Dipstick Test and Latex Agglutination Test Using Bayesian Latent Class Model and MAT as Gold Standard.

Leptospirosis is a spirochaetal infection that possesses a broad host range affecting almost all mammals. In the present study, the microscopic agglutination test (MAT) was compared with recombinant LigA/B antigen-based point-of-care diagnostics such as the in-house IgM dot ELISA dipstick test (IgM- DEDT) and the latex agglutination test (LAT) for the serodiagnosis of human leptospirosis. The comparison of the MAT with these two point–of-care diagnostics was performed using the MAT as the gold standard test and using Bayesian latent class modelling (BLCM), which considers all diagnostic tests as imperfect. The N-terminal conserved region of the LigA/B protein spanning the first to fifth big tandem repeat domains (rLigA/BCon1-5) was employed as a serodiagnostic marker in both of the bedside assays. A total of 340 serum samples collected from humans involved in high risk occupations were screened using the MAT, IgM DEDT and LAT. During the early phase of leptospirosis, BLCM analysis showed that the IgM DEDT and LAT had similar sensitivities (99.6 (96.0–100)) and (99.5 (95.2–100)), respectively, while the single acute phase MAT had the lowest sensitivity (83.3 (72.8–91.3)). Both the IgM DEDT and the LAT may be superior to the single acute phase MAT in terms of sensitivity during the early phase of infection and may be suitable for the early diagnosis of leptospirosis. However, BLCM analysis revealed that the use of both acute and convalescent samples substantially increased the sensitivity of the final MAT (98.2% (93.0–99.8%)) as a test to diagnose human leptospirosis. Both the IgM DEDT and LAT can be employed as bedside spot tests in remote locations where the MAT is not easily accessible.

Detection of Low MAP Shedder Prevalence in Large Free-Stall Dairy Herds by Repeated Testing of Environmental Samples and Pooled Milk Samples.

Simple SummaryParatuberculosis is a disease which affects ruminants worldwide. Many countries have implemented certification and monitoring systems to control the disease, particularly in dairy herds. Monitoring herds certified as paratuberculosis non-suspect is an important component of paratuberculosis herd certification programs. The challenge is to detect the introduction or reintroduction of the infectious agent as early as possible with reasonable efforts but high certainty. In our study, we evaluated different low-cost testing schemes in herds where the share of infected animals was low, resulting in a low within-herd prevalence of animals shedding the bacteria that causes paratuberculosis in their feces. The test methods used were repeated pooled milk samples and fecal samples from the barn environment. Our study showed that numerous repetitions of different samples are necessary to monitor such herds with sufficiently high certainty. In the case of herds with a very low prevalence, our study showed that a combination of different sampling approaches is required.An easy-to-use and affordable surveillance system is crucial for paratuberculosis control. The use of environmental samples and milk pools has been proven to be effective for the detection of Mycobacterium avium subsp. paratuberculosis (MAP)-infected herds, but not for monitoring dairy herds certified as MAP non-suspect. We aimed to evaluate methods for the repeated testing of large dairy herds with a very low prevalence of MAP shedders, using different sets of environmental samples or pooled milk samples, collected monthly over a period of one year in 36 herds with known MAP shedder prevalence. Environmental samples were analyzed by bacterial culture and fecal PCR, and pools of 25 and 50 individual milk samples were analyzed by ELISA for MAP-specific antibodies. We estimated the cumulative sensitivity and specificity for up to twelve sampling events by adapting a Bayesian latent class model and taking into account the between- and within-test correlation. Our study revealed that at least seven repeated samplings of feces from the barn environment are necessary to achieve a sensitivity of 95% in herds with a within-herd shedder prevalence of at least 2%. The detection of herds with a prevalence of less than 2% is more challenging and, in addition to numerous repetitions, requires a combination of different samples.

A Bayesian latent class model for predicting gestational age in health administrative data.

Health administrative data are oftentimes of limited use in epidemiological study on drug safety in pregnancy, due to lacking information on gestational age at birth (GAB). Although several studies have proposed algorithms to estimate GAB using claims database, failing to incorporate the unique distributional shape of GAB, can introduce bias in estimates and subsequent modeling. Hence, we develop a Bayesian latent class model to predict GAB. The model employs a mixture of Gaussian distributions with linear covariates within each class. This approach allows modeling heterogeneity in the population by identifying latent subgroups and estimating class-specific regression coefficients. We fit this model in a Bayesian framework conducting posterior computation with Markov Chain Monte Carlo methods. The method is illustrated with a dataset of 10,043 Rhode Island Medicaid mother-child pairs. We found that the three-class and six-class mixture specifications maximized prediction accuracy. Based on our results, Medicaid women were partitioned into three classes, featured by extreme preterm or preterm birth, preterm or" early" term birth, and" late" term birth. Obstetrical complications appeared to pose a significant influence on class-membership. Altogether, compared to traditional linear models our approach shows an advantage in predictive accuracy, because of superior flexibility in modeling a skewed response and population heterogeneity.

Bayesian estimation of sensitivity and specificity of a rapid mastitis test kit, bacterial culture, and PCR for detection of Staphylococcus aureus, Streptococcus species, and coliforms in bovine milk samples

Our objectives were to evaluate the diagnostic accuracy of a rapid and novel immunochromatography-based mastitis kit that includes 3 independent tests to detect coliforms (Escherichia coli or Klebsiella pneumoniae), Streptococcus spp., and Staphylococcus aureus. The kit was developed to facilitate diagnostic-based mastitis treatment. Validation of the kit was based on 154 aseptically collected mastitis samples from 2 clinical herds (clinical population) and 120 milk samples from 3 nonclinical herds (nonclinical population) without clinical cases at the time of enrollment. One herd sampled at different times was common to both populations. A 3-test in 2-population Bayesian latent class model with uniform priors for all test parameters except specificity of culture, which was modeled informatively, was used to estimate sensitivity (Se) and specificity (Sp) of the test kit, culture, and PCR at the cow level. The mastitis test kit's 96.9% Sp for Streptococcus spp. had a low false positive percentage (3.1%), which, together with the kit's rapid turnaround time for results, makes it a suitable initial screening test that producers can use to identify clinical cows to treat based on Streptococcus spp. mastitis in kit-positive results. Due to the 60.4% kit Se, producers should follow up on Streptococcus spp. kit-negative cows using a confirmatory test such as PCR (Sp of 98.4%) or culture (Sp of 99.6%). In contrast, aerobic culture had Se of 76.5% and Sp of 99.6% for Streptococcus spp. Similarly, the Sp of the kit (98.2%) and culture (99.8%) for Staph. aureus were particularly high, and even though the kit's Se (61.0%) was lower than culture (88.4%; posterior probability of difference 98%), the kit could be beneficial before use of a confirmatory test for kit-negative samples due to its ease and rapid turnaround time. Mostly, quantitative real-time (q)PCR outperformed the kit's Se (37.7%) and Sp (92.9%) for coliforms, as well as the kit's Se (60.4%) for Streptococcus spp. However, qPCR may require more technical skills and turnaround time for final results. Use of the on-farm mastitis test kit evaluated in the present study could enhance sustainable antimicrobial drug use by rapidly identifying Streptococcus mastitis for targeted treatment. Furthermore, the kit may be used in a Staph. aureus outbreak where cows can be rapidly screened to identify cases for segregation or culling during an outbreak and kit-negative cows further confirmed by milk culture or qPCR. However, the cost-effectiveness of such an approach has not been investigated.

Evaluation of the Accuracy of Cognitive Screening Tests in Detecting Dementia Associated with Alzheimer's Disease: A Hierarchical Bayesian Latent Class Meta-Analysis.

Montreal Cognitive Assessment (MoCA) and Mini-Mental State Examination (MMSE) are neuropsychological tests commonly used by physicians for screening cognitive dysfunction of Alzheimer's disease (AD). Due to different imperfect reference standards, the performance of MoCA and MMSE do not reach consensus. It is necessary to evaluate the consistence and differentiation of MoCA and MMSE in the absence of a gold standard for AD. We aimed to assess the accuracy of MoCA and MMSE in screening AD without a gold standard reference test. Studies were identified from PubMed, Web of Science, CNKI, Chinese Wanfang Database, China Science and Technology Journal Database, and Cochrane Library. Our search was limited to studies published in English and Chinese before August 2021. A hierarchical Bayesian latent class model was performed in meta-analysis when the gold standard was absent. A total of 67 studies comprising 5,554 individuals evaluated for MoCA and 76,862 for MMSE were included in this meta-analysis. The pooled sensitivity was 0.934 (95% CI 0.905 to 0.954) for MoCA and 0.883 (95% CI 0.859 to 0.903) for MMSE, while the pooled specificity was 0.899 (95% CI 0.859 to 0.928) for MoCA and 0.903 (95% CI 0.879 to 0.923) for MMSE. MoCA was useful to rule out dementia associated with AD with lower negative likelihood ratio (LR-) (0.074, 95% CI 0.051 to 0.108). MoCA showed better performance with higher diagnostic odds ratio (DOR) (124.903, 95% CI 67.459 to 231.260). MoCA had better performance than MMSE in screening dementia associated with AD from patients with mild cognitive impairment or healthy controls.

Automated Bot Detection Using Bayesian Latent Class Models in Online Surveys.

Behavioral scientists have become increasingly reliant on online survey platforms such as Amazon's Mechanical Turk (Mturk). These platforms have many advantages, for example it provides ease of access to difficult to sample populations, a large pool of participants, and an easy to use implementation. A major drawback is the existence of bots that are used to complete online surveys for financial gain. These bots contaminate data and need to be identified in order to draw valid conclusions from data obtained with these platforms. In this article, we will provide a Bayesian latent class joint modeling approach that can be routinely applied to identify bots and simultaneously estimate a model of interest. This method can be used to separate the bots' response patterns from real human responses that were provided in line with the item content. The model has the advantage that it is very flexible and is based on plausible assumptions that are met in most empirical settings. We will provide a simulation study that investigates the performance of the model under several relevant scenarios including sample size, proportion of bots, and model complexity. We will show that ignoring bots will lead to severe parameter bias whereas the Bayesian latent class model results in unbiased estimates and thus controls this source of bias. We will illustrate the model and its capabilities with data from an empirical political ideation survey with known bots. We will discuss the implications of the findings with regard to future data collection via online platforms.

Population Size Estimation From Capture-Recapture Studies Using shinyrecap: Design and Implementation of a Web-Based Graphical User Interface.

BackgroundPopulation size estimates (PSE) provide critical information in determining resource allocation for HIV services geared toward those at high risk of HIV, including female sex workers, men who have sex with men, and people who inject drugs. Capture-recapture (CRC) is often used to estimate the size of these often-hidden populations. Compared with the commonly used 2-source CRC, CRC relying on 3 (or more) samples (3S-CRC) can provide more robust PSE but involve far more complex statistical analysis.ObjectiveThis study aims to design and describe the Shiny application (shinyrecap), a user-friendly interface that can be used by field epidemiologists to produce PSE.Methodsshinyrecap is built on the Shiny web application framework for R. This allows it to seamlessly integrate with the sophisticated CRC statistical packages (eg, Rcapture, dga, LCMCR). Additionally, the application may be accessed online or run locally on the user’s machine.ResultsThe application enables users to engage in sample size calculation based on a simulation framework. It assists in the proper formatting of collected data by providing a tool to convert commonly used formats to that used by the analysis software. A wide variety of methodologies are supported by the analysis tool, including log-linear, Bayesian model averaging, and Bayesian latent class models. For each methodology, diagnostics and model checking interfaces are provided.ConclusionsThrough a use case, we demonstrated the broad utility of this powerful tool with 3S-CRC data to produce PSE for female sex workers in a subnational unit of a country in sub-Saharan Africa.

Field evaluation of diagnostic sensitivity (DSe) and specificity (DSp) of common tests for amoebic gill disease (AGD) and complex gill disease (CGD) in cultured Atlantic salmon (Salmo salar) in Scotland using Bayesian latent class models

Amoebic gill disease (AGD) and complex gill disease (CGD) are the most significant marine gill diseases in salmon aquaculture in Scotland. Little is published about diagnostic performance of tests to detect these diseases, making it difficult to interpret test results. We estimated diagnostic sensitivity (DSe) and specificity (DSp) of common tests for AGD (gross AGD score, qPCR for Neoparamoeba perurans, histopathology) and CGD (gross proliferative gill disease (PGD) score, gross total gill score, histopathology). Because specifications in our sampling protocol implemented to encourage consistency across the farms might affect diagnostic performance of histopathology (historically the reference standard for gill diseases), we used Bayesian latent class models without reference standard. Cases and non-cases were based on less, medium, and severe stringent case definitions, representing different cut-off levels for the different tests. Gross gill scores for both diseases were excellent in designating non-diseased fish, DSps were generally around 1. To detect CGD, DSe of gross total gill score and gross PGD score were between respectively 0.81 (0.73 – 0.91 lower to upper 95% credible interval) and 0.53 (0.46 – 0.64) for medium stringent case definitions, and to detect AGD the DSe for the gross AGD score was between 0.53 (0.48–0.57) and 0.14 (0.07 – 0.22) for respectively the less and severe stringent case definition. Thus, gross gill scores were medium to good in designating truly diseased fish, implying some false negatives are expected. For CGD the DSe for gross total gill scores were the highest, for AGD it was the qPCR test at a DSe of 0.92 (0.86 – 0.99). For both diseases, DSe was lowest for histopathology, e.g. 0.23 (0.16 – 0.30) for AGD and 0.1 (0.07 – 0.14) for CGD under medium stringent case definitions, perhaps due to collecting the second gill arch on the right rather than the worst affected arch, whilst PCR sampling and gross gill scoring included multiple (PCR) or all (gross scoring) gill arches. The diagnostic goals of these tests differ; gross gill scoring provides a low-cost presumptive diagnosis, PCR a non-lethal confirmation of the presence of a specific pathogen and histopathology provides information on the underlying aetiology of gill damage as well as the extent, severity, and chronology of gill disease. An effective gill health surveillance strategy is likely to incorporate multiple diagnostic tools used in a complementary manner.

Reproductive tract disease in Irish grazing dairy cows: Retrospective observational study examining its association with reproductive performance and accuracy of 2 diagnostic tests

The detection of reproductive tract disease (RTD) 3 wk postpartum is important because of its effect on subsequent reproductive outcomes. Numerous methods for the diagnosis of RTD are described, some of which are more practical and instantaneous in terms of diagnosis. Two of these methods involve identification of purulent vaginal discharge (PVD) and evidence of ultrasonographic uterine changes indicative of endometritis (UE). The objectives of our retrospective observational study were (1) to assess the association of PVD or UE score at the prebreeding examination (PBE) with the hazard of pregnancy within the subsequent breeding season; (2) to determine the test sensitivity (Se) and specificity (Sp) at the point of sampling of both tests using a Bayesian latent class model; and (3) to determine the effect of varying positivity thresholds on test accuracy. To achieve these objectives, we analyzed an initial data set of 5,049 PBE from 2,460 spring-calved cows in 8 herds between 2014 and 2018. Each PBE was conducted once between 25 and 86 d in milk. At each PBE, vaginal discharge was obtained with a Metricheck device (Simcro) whereas uterine contents were assessed using transrectal ultrasonography. Purulent vaginal discharge was scored on a scale of 0 to 3 depending on discharge character, and UE was scored on a scale of 0 to 4 depending on the presence and consistency of intraluminal fluid. Cows with scores of ≥2 in either test had received treatment. Fertility data were available from 4,756 PBE after data exclusion. The association between PVD or UE score at the PBE and subsequent hazard of pregnancy was analyzed using a Cox proportional hazards model. Cows with a PVD score of 2 or 3 were less likely to conceive than cows with a PVD score 0 [score 2 hazard ratio (HR) = 0.74; 95% confidence interval (CI): 0.59-0.94; score 3 HR = 0.65; 95% CI: 0.51-0.84]. Cows with a UE score of 1, 2, 3, or 4 were less likely to conceive than cows with a UE score of 0 (score 1 HR = 0.82; 95% CI: 0.73-0.93; score 2 HR = 0.79; 95% CI: 0.62-1.00; score 3 HR = 0.43; 95% CI: 0.43-0.90; score 4 HR = 0.39; 95% CI: 0.26-0.58). To determine the Se and Sp of PVD or UE score for diagnosis of RTD at the time of PBE, a Bayesian latent class model was fitted on 2,460 individual cow PBE. Flat priors were used for the Se and Sp of UE, whereas informative priors were used for PVD Se (mode = 65%, 5th percentile = 45%) and Sp (mode = 90%, 5th percentile = 80%) and RTD prevalence (mode = 20%, 5th percentile = 10%). Posterior estimates (median and 95% Bayesian probability intervals; BPI) were obtained using 'rjags' (R Studio). The optimal test thresholds (PVD and UE score ≥1) were selected by assessing the effect of different thresholds on test estimates and using a misclassification cost analysis. Based on these, median (95% BPI) Se for PVD and UE score ≥1 were 44% (29-60%) and 67% (33-100%), respectively. Median Sp for PVD and UE score ≥1 were 90% (86-93%) and 91% (86-93%), respectively. Higher scores in both tests were associated with impaired fertility, and UE scoring with a threshold of ≥1 had the highest test Se and Sp estimates although test Se was conditional on days in milk when the PBE occurred.

Bayesian latent class models for identifying canine visceral leishmaniosis using diagnostic tests in the absence of a gold standard.

Like many infectious diseases, there is no practical gold standard for diagnosing clinical visceral leishmaniasis (VL). Latent class modeling has been proposed to estimate a latent gold standard for identifying disease. These proposed models for VL have leveraged information from diagnostic tests with dichotomous serological and PCR assays, but have not employed continuous diagnostic test information. In this paper, we employ Bayesian latent class models to improve the identification of canine visceral leishmaniasis using the dichotomous PCR assay and the Dual Path Platform (DPP) serology test. The DPP test has historically been used as a dichotomous assay, but can also yield numerical information via the DPP reader. Using data collected from a cohort of hunting dogs across the United States, which were identified as having either negative or symptomatic disease, we evaluate the impact of including numerical DPP reader information as a proxy for immune response. We find that inclusion of DPP reader information allows us to illustrate changes in immune response as a function of age. Utilization of continuous DPP reader information can improve the correct discrimination between individuals that are negative for disease and those with clinical VL. These models provide a promising avenue for diagnostic testing in contexts with multiple, imperfect diagnostic tests. Specifically, they can easily be applied to human visceral leishmaniasis when diagnostic test results are available. Also, appropriate diagnosis of canine visceral leishmaniasis has important consequences for curtailing spread of disease to humans.

Determination of an optimal ELISA cut-off for the diagnosis of Toxoplasma gondii infection in pigs using Bayesian latent class modelling of data from multiple diagnostic tests

Toxoplasma gondii infection in pigs is commonly diagnosed using serological tests that detect IgG antibodies targeted against the parasite. Such tests include enzyme-linked immunosorbent assay (ELISA), modified agglutination test (MAT), and western blot (WB), which are commercially available as rapid test kits. In this study, we evaluated the manufacturer recommended cut-off of ELISA-PrioCHECK test kit and determined a new optimal cut-off for identifying T. gondii infections in pigs. Assessment of the commercial ELISA kit was done by including data from two additional serological tests, MAT, and WB, applied to seven pig population categories with varying prevalences. A total of 233 plasma samples that were previously used in other studies for investigating T. gondii seroprevalence in pigs in Denmark were randomly selected for inclusion, including 95 samples that had previously been analysed with all three tests and an additional 138 samples that were analysed using the three serological tests for this study. In the absence of a gold standard test, a latent class model was fit to the data to obtain estimates of sensitivity and specificity for each of the tests along with prevalence in each of the populations. A cut-off that maximized the sensitivity and specificity of the ELISA test was then selected. The optimal cut-off value for percent of positive control (PP) in ELISA-PrioCHECK was estimated to be 27.7 PP, which is higher than the cut-off value of 20 PP that is recommended by the manufacturer. At this cut-off, the estimated sensitivities of ELISA, MAT and WB were 99.2% (96.3–100.0%), 96.3% (88.0–100.0%), and 89.8% (80.0–98.0%), respectively. The estimated specificities of ELISA, MAT and WB were 95.2% (92.5–97.6%), 99.6% (97.5–100.0%), and 98.2% (95.9–100.0%), respectively. Our findings have broad relevance to the use of the ELISA-PrioCHECK test kit for detecting Toxoplasma gondii infection in pigs.

Translating From Egg- to Antigen-Based Indicators for Schistosoma mansoni Elimination Targets: A Bayesian Latent Class Analysis Study.

Schistosomiasis is a parasitic disease affecting over 240-million people. World Health Organization (WHO) targets for Schistosoma mansoni elimination are based on Kato-Katz egg counts, without translation to the widely used, urine-based, point-of-care circulating cathodic antigen diagnostic (POC-CCA). We aimed to standardize POC-CCA score interpretation and translate them to Kato-Katz-based standards, broadening diagnostic utility in progress towards elimination. A Bayesian latent-class model was fit to data from 210 school-aged-children over four timepoints pre- to six-months-post-treatment. We used 1) Kato-Katz and established POC-CCA scoring (Negative, Trace, +, ++ and +++), and 2) Kato-Katz and G-Scores (a new, alternative POC-CCA scoring (G1 to G10)). We established the functional relationship between Kato-Katz counts and POC-CCA scores, and the score-associated probability of true infection. This was combined with measures of sensitivity, specificity, and the area under the curve to determine the optimal POC-CCA scoring system and positivity threshold. A simulation parametrized with model estimates established antigen-based elimination targets. True infection was associated with POC-CCA scores of ≥ + or ≥G3. POC-CCA scores cannot predict Kato-Katz counts because low infection intensities saturate the POC-CCA cassettes. Post-treatment POC-CCA sensitivity/specificity fluctuations indicate a changing relationship between egg excretion and antigen levels (living worms). Elimination targets can be identified by the POC-CCA score distribution in a population. A population with ≤2% ++/+++, or ≤0.5% G7 and above, indicates achieving current WHO Kato-Katz-based elimination targets. Population-level POC-CCA scores can be used to access WHO elimination targets prior to treatment. Caution should be exercised on an individual level and following treatment, as POC-CCAs lack resolution to discern between WHO Kato-Katz-based moderate- and high-intensity-infection categories, with limited use in certain settings and evaluations.