Bayesian sample size determination for prevalence and diagnostic test studies in the absence of a gold standard test

Planning studies involving diagnostic tests is complicated by the fact that virtually no test provides perfectly accurate results. The misclassification induced by imperfect sensitivities and specificities of diagnostic tests must be taken into account, whether the primary goal of the study is to estimate the prevalence of a disease in a population or to investigate the properties of a new diagnostic test. Previous work on sample size requirements for estimating the prevalence of disease in the case of a single imperfect test showed very large discrepancies in size when compared to methods that assume a perfect test. In this article we extend these methods to include two conditionally independent imperfect tests, and apply several different criteria for Bayesian sample size determination to the design of such studies. We consider both disease prevalence studies and studies designed to estimate the sensitivity and specificity of diagnostic tests. As the problem is typically nonidentifiable, we investigate the limits on the accuracy of parameter estimation as the sample size approaches infinity. Through two examples from infectious diseases, we illustrate the changes in sample sizes that arise when two tests are applied to individuals in a study rather than a single test. Although smaller sample sizes are often found in the two-test situation, they can still be prohibitively large unless accurate information is available about the sensitivities and specificities of the tests being used.     

A Bayesian-based approach for spatio-temporal modeling of county level prevalence of Schistosoma japonicum infection in Jiangsu province, China

Spatio-temporal variations of Schistosoma japonicum infection risk in Jiangsu province, China, were examined and the relationships between key climatic factors and infection prevalence at the county level were determined. The parasitological data were collected annually by means of cross-sectional surveys carried out in 47 counties from 1990 to 1998. Climatic factors, namely land surface temperature (LST) and normalized difference vegetation index (NDVI), were obtained from remote sensing satellite sensors. Bayesian spatio-temporal models were employed to analyze the data. The best fitting model showed that spatial autocorrelation in Jiangsu province decreased dramatically from 1990 to 1992 and increased gradually thereafter. A likely explanation of this finding arises from the large-scale administration of praziquantel for morbidity control of schistosomiasis. Our analysis suggested a negative association between NDVI and risk of S. japonicum infection. On the other hand, an increase in LST contributed to a significant increase in S. japonicum infection prevalence. We conclude that combining geographic information system, remote sensing and Bayesian-based statistical approaches facilitate integrated risk modeling of S. japonicum, which in turn is of relevance for allocation of scarce resources for control of schistosomiasis japonica in Jiangsu province and elsewhere in China, where the disease remains of public health and economic significance.     

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

BackgroundLike 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.Methods/Principal findingsIn 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.Conclusions/SignificanceUtilization 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.     

Bayesian estimation of community prevalences of Schistosoma japonicum infection in China

A Bayesian approach to overcome the imperfections of an immunological test (an antibody-based ELISA) and a parasitological test (Kato-Katz) in the detection of Schistosoma japonicum infection, was used to estimate community prevalences of S. japonicum infection in China. At the same time, the similarity between the prevalence estimates based on data from ELISA alone and those using data from both ELISA and Kato-Katz tests was explored. The database from the third nationwide sampling survey of schistosomiasis in China, 2004, was used for analysis, in which a total of 239 endemic villages were sampled from seven endemic provinces through a stratified cluster sampling technique and 250,987 residents aged from 6 to 65 years, were examined by ELISA followed by a Kato-Katz test applied to the seropositives. Bayesian hierarchical models incorporating random effects to reflect the nested data structure and uncertainty about test properties were employed to analyse the data. Our analysis suggested that using data from ELISA alone or both ELISA and Kato-Katz tests resulted in similar prevalence estimates, probably owing to the lack of sensitivity of Kato-Katz and the fact that Kato-Katz was only applied to the seropositives. We conclude that it is feasible to employ only ELISA, instead of combined ELISA and Kato-Katz tests, to estimate prevalence of S. japonicum infection in large-scale epidemiological settings. This study confirmed heterogeneity in the prevalence of S. japonicum infection in space by the fact that the estimated prevalences of S. japonicum infection in the sampled villages ranged from 0.02% to about 56% (posterior median). It is indicated that the disease remains a threat in some areas along the Yangtze River, although great achievements have been made in the control programme of schistosomiasis in China.     

A mechanistic spatio-temporal modeling of COVID-19 data

Understanding the evolution of an epidemic is essential to implement timely and efficient preventive measures. The availability of epidemiological data at a fine spatio-temporal scale is both novel and highly useful in this regard. Indeed, having geocoded data at the case level opens the door to analyze the spread of the disease on an individual basis, allowing the detection of specific outbreaks or, in general, of some interactions between cases that are not observable if aggregated data are used. Point processes are the natural tool to perform such analyses. We analyze a spatio-temporal point pattern of Coronavirus disease 2019 (COVID-19) cases detected in Valencia (Spain) during the first 11 months (February 2020 to January 2021) of the pandemic. In particular, we propose a mechanistic spatio-temporal model for the first-order intensity function of the point process. This model includes separate estimates of the overall temporal and spatial intensities of the model and a spatio-temporal interaction term. For the latter, while similar studies have considered different forms of this term solely based on the physical distances between the events, we have also incorporated mobility data to better capture the characteristics of human populations. The results suggest that there has only been a mild level of spatio-temporal interaction between cases in the study area, which to a large extent corresponds to people living in the same residential location. Extending our proposed model to larger areas could help us gain knowledge on the propagation of COVID-19 across cities with high mobility levels.     

A Bayesian method for comparing and combining binary classifiers in the absence of a gold standard

ABSTRACT: BACKGROUND: Many problems in bioinformatics involve classification based on features such as sequence, structure or morphology. Given multiple classifiers, two crucial questions arise: how does their performance compare, and how can they best be combined to produce a better classifier? A classifier can be evaluated in terms of sensitivity and specificity using benchmark, or gold standard, data, that is, data for which the true classification is known. However, a gold standard is not always available. Here we demonstrate that a Bayesian model for comparing medical diagnostics without a gold standard can be successfully applied in the bioinformatics domain, to genomic scale data sets. We present a new implementation, which unlike previous implementations is applicable to any number of classifiers. We apply this model, for the first time, to the problem of finding the globally optimal logical combination of classifiers. RESULTS: We compared three classifiers of protein subcellular localisation, and evaluated our estimates of sensitivity and specificity against estimates obtained using a gold standard. The method overestimated sensitivity and specificity with only a small discrepancy, and correctly ranked the classifiers. Diagnostic tests for swine flu were then compared on a small data set. Lastly, classifiers for a genome-wide association study of macular degeneration with 541094 SNPs were analysed. In all cases, run times were feasible, and results precise. The optimal logical combination of classifiers was also determined for all three data sets. Code and data are available from http://bioinformatics.monash.edu.au/downloads/. CONCLUSIONS: The examples demonstrate the methods are suitable for both small and large data sets, applicable to the wide range of bioinformatics classification problems, and robust to dependence between classifiers. In all three test cases, the globally optimal logical combination of the classifiers was found to be their union, according to three out of four ranking criteria. We propose as a general rule of thumb that the union of classifiers will be close to optimal.     

Congenital transmission of Schistosoma japonicum in the rabbit

Fourteen pregnant rabbits were each infected with 300 cercariae of Schistosoma japonicum and divided into two groups. Group M (n = 8) was infected during mid-gestation (the organogenetic stage) and group L (n = 6) was infected during late-gestation (the post-organogenetic stage). Mother rabbits and rabbit kittens were killed 45-60 days after infection and perfused in order to obtain worm counts. Furthermore, faecal egg counts and tissue egg counts from livers were obtained from the mother rabbits as well as the rabbit kittens. All mother rabbits became infected harbouring 207.6 +/- 20.2 and 220.0 +/- 27.5 adult worms in group M and L, respectively. In groups M and L, 13.5% and 46.7% of the kittens were infected, respectively. In 12 of 14 litters at least one kitten was infected. The infected kittens harboured between one and three adult S. japonicum. The livers of the kittens infected with a worm pair displaced lesions as a result of egg deposition. The results, therefore, show that congenital transmission of S. japonicumcan occur in rabbits. The close anatomical resemblance between the rabbit and human placenta may be indicative of the presence of congenital transmission of S. japonicum infection in humans.     

Chromosomal differentiation of the Schistosoma japonicum complex

The C-banding pattern, location of telomere sequence and chiasma frequency of four species of the Schistosoma japonicum complex were compared with those of two African species, Schistosoma mansoni and Schistosoma haematobium. In the six species, C-banding patterns of seven autosomes and the two sex chromosomes (Z and W) showed relatively species-specific and geographical (Asian and African) differences. Particularly, a plausible pathway of alteration of chromosome 2 revealed a direction from the A-chromosome to the M- chromosome in terms of rearrangements of pericentric inversion and elimination of constitutive heterochromatin (AM inversion). This chromosome change suggested hypothetically that the S. japonicum complex is the original type, and the African species represents the derived type. Moreover, the mosaic construct of the Asian and African types in Schistosoma sinensium chromosomes prompted us to propose that the species might have been formed by hybrid speciation of the genomes of Asian and African species. Localisation of telomeric repeats enabled Asian and African schistosomes to be distinguished clearly by simple terminal location and by terminal and interstitial locations, respectively. Change of chiasma frequency in the S. japonicum complex might be caused by the reduction of interstitial chiasmate (Xi) in the larger chromosomes, 1 and Z (or W), and the change seems to have progressed to Japan from South East Asia. These data enabled us to predict a tentative evolutionary pathway of schistosomes at the cytogenetic level.     

High prevalence of Schistosoma japonicum by perfusion in naturally exposed water buffalo in a region of the Philippines endemic for human schistosomiasis

In the past decade, ecological surveys emphasized rats and dogs as the most significant animal reservoirs for Schistosoma japonicum (S.j) in the Philippines. However, recent studies demonstrated 51–91% prevalence of schistosomiasis among water buffalo using qPCR in the Sj endemic regions in the Philippines. In order to resolve the inconsistency of reported surveys regarding Sj endemicity among carabao, a domestic water buffalo that is the most important draught animal, we introduced 42 schistosome negative water buffalo to Macanip, Jaro municipality, Leyte, the Philippines, a subsistence rice-farming village that has been the focus of schistosomiasis japonica studies of our group for the past 20 years. We conducted perfusion to the remaining 34 buffalo that survived 10 months of nature exposure and Typhoon Haiyan. Thirty-three water buffalo were found to be positive with at least 1 pair of worms from the mesenteric vein. The infection rate is 97%, with the worm burden of 94 (95% confidence interval, 49–138 worms) worms. To our knowledge, this is the first report about S. japonicum worm burden in naturally infected water buffalo in the Philippines. The fact that with less than one-year of exposure, in this human schistosomiasis endemic area, only 1 out of 34 water buffalo was uninfected is striking. Urgent attention is needed for a cost-effective technique for monitoring Sj infection in animals and humans. Meanwhile, intervention implementation, including water buffalo treatment and vaccination, should be taken into consideration.     

Routine Kato-Katz technique underestimates the prevalence of Schistosoma japonicum: a case study in an endemic area of the People's Republic of China

There is an evidence that the Kato-Katz technique lacks sensitivity and may hence be an unsuitable method for the assessment of the 'real infection status' in community with low-intensity infections. In this study, six Kato-Katz thick smears (examination of two stool samples with three thick smears each) were used as the diagnostic 'gold' standard for estimating the prevalence of Schistosoma japonicum infection and the results were compared with results based on fewer Kato-Katz thick smear readings. A total of 1055 individuals in 2005 and 725 in 2006 from an endemic village were recruited for the study. The observed prevalence increased gradually with the number of Kato-Katz thick smears examined, and hence the rate of underestimation decreased accordingly. The prevalence based on single Kato-Katz thick smear readings was significantly lower than that obtained using five or six thick smears. The rate of underestimation based on using two and three Kato-Katz thick smears, a typical diagnostic effort in the national schistosomiasis control programme, was about 36.0% (28.4-48.9%) and 25.0% (15.9-40.7%). The number of Kato-Katz thick smears required to secure detection of a S. japonicum infection varies with the infection intensity level. Indeed, examination of a single thick smear was sufficient when the geometric mean of the fecal content of eggs per gram (EPG) was 250 or higher in infected individuals, while six Kato-Katz thick smears were required when the EPG score was lower than 10. In conclusion, our results confirm that the prevalence of S. japonicum infection in a community is generally considerably "underestimated". Moreover, our findings provide a benchmark for the proper application of the Kato-Katz technique and the rational evaluation of the epidemic situation, as well as a scientific basis for constructing a mathematic diagnostic model.     

Bayesian latent trait modeling of migraine symptom data

Definition of disease phenotype is a necessary preliminary to research into genetic causes of a complex disease. Clinical diagnosis of migraine is currently based on diagnostic criteria developed by the International Headache Society. Previously, we examined the natural clustering of these diagnostic symptoms using latent class analysis (LCA) and found that a four-class model was preferred. However, the classes can be ordered such that all symptoms progressively intensify, suggesting that a single continuous variable representing disease severity may provide a better model. Here, we compare two models: item response theory and LCA, each constructed within a Bayesian context. A deviance information criterion is used to assess model fit. We phenotyped our population sample using these models, estimated heritability and conducted genome-wide linkage analysis using Merlin-qtl. LCA with four classes was again preferred. After transformation, phenotypic trait values derived from both models are highly correlated (correlation = 0.99) and consequently results from subsequent genetic analyses were similar. Heritability was estimated at 0.37, while multipoint linkage analysis produced genome-wide significant linkage to chromosome 7q31-q33 and suggestive linkage to chromosomes 1 and 2. We argue that such continuous measures are a powerful tool for identifying genes contributing to migraine susceptibility.     

Bayesian approaches to modeling the conditional dependence between multiple diagnostic tests

Many analyses of results from multiple diagnostic tests assume the tests are statistically independent conditional on the true disease status of the subject. This assumption may be violated in practice, especially in situations where none of the tests is a perfectly accurate gold standard. Classical inference for models accounting for the conditional dependence between tests requires that results from at least four different tests be used in order to obtain an identifiable solution, but it is not always feasible to have results from this many tests. We use a Bayesian approach to draw inferences about the disease prevalence and test properties while adjusting for the possibility of conditional dependence between tests, particularly when we have only two tests. We propose both fixed and random effects models. Since with fewer than four tests the problem is nonidentifiable, the posterior distributions are strongly dependent on the prior information about the test properties and the disease prevalence, even with large sample sizes. If the degree of correlation between the tests is known a priori with high precision, then our methods adjust for the dependence between the tests. Otherwise, our methods provide adjusted inferences that incorporate all of the uncertainty inherent in the problem, typically resulting in wider interval estimates. We illustrate our methods using data from a study on the prevalence of Strongyloides infection among Cambodian refugees to Canada.     

Schistosoma japonicum: the pathology of experimental infection

The pathology of experimental schistosomiasis japonica is reviewed and compared with the pathology of schistosomiasis japonica in man and to some aspects of schistosomiasis mansoni and schistosomiasis haematobia in experimental animals. The induction of granulomas around Schistosoma japonicum eggs depends upon cell mediated immunity, as do the reactions to Schistosoma mansoni and Schistosoma haematobium eggs. However, the modulation of the reaction to S. japonicum eggs can be greatly influenced by antibody, while antibody has no effect on the granulomas around S. mansoni eggs. Adult worm pairs of S. japonicum tend to cluster in the mesenteric venules, and most eggs are laid in a few sites. This leads to large, focal intestinal lesions similar to the discrete lesions produced by S. haematobium in the intestine and urinary tract but in contrast to the widespread, diffuse lesions produced by S. mansoni. Comparison with S. japonicum infection in humans is limited chiefly by our scant knowledge of the pathology produced by S. japonicum in infected persons. Most such comparisons are, in any case, limited by the marked differences in the reactions of various experimental host species to the infection and by differences in the reaction of a given host species to different strains of the parasite.     

Measuring obesity in the absence of a gold standard

Reliable measures of body composition are essential to develop effective policies to tackle obesity. The lack of an acceptable gold-standard for measuring fatness has made it difficult to evaluate alternative measures of obesity. We use latent class analysis to characterise existing diagnostics. Using data on US adults we show that measures based on body mass index and bioelectrical impedance analysis misclassify large numbers of individuals. For example, 45% of obese White women are misclassified as non-obese using body mass index, while over 50% of non-obese White women are misclassified as being obese using bioelectrical impedance analysis. In contrast the misclassification rates are low when waist circumference is used to measure obesity. These results have important implications for our understanding of differences in obesity rates across time and groups, as well as posing challenges for the econometric analysis of obesity.     

Sensitivity of diagnostic tests for human soil-transmitted helminth infections: a meta-analysis in the absence of a true gold standard

Reliable, sensitive and practical diagnostic tests are an essential tool in disease control programmes for mapping, impact evaluation and surveillance. To provide a robust global assessment of the relative performance of available diagnostic tools for the detection of soil-transmitted helminths, we conducted a meta-analysis comparing the sensitivities and the quantitative performance of the most commonly used copro-microscopic diagnostic methods for soil-transmitted helminths, namely Kato-Katz, direct microscopy, formol-ether concentration, McMaster, FLOTAC and Mini-FLOTAC. In the absence of a perfect reference standard, we employed a Bayesian latent class analysis to estimate the true, unobserved sensitivity of compared diagnostic tests for each of the soil-transmitted helminth species Ascaris lumbricoides, Trichuris trichiura and the hookworms. To investigate the influence of varying transmission settings we subsequently stratified the analysis by intensity of infection. Overall, sensitivity estimates varied between the different methods, ranging from 42.8% for direct microscopy to 92.7% for FLOTAC. The widely used double slide Kato-Katz method had a sensitivity of 74–95% for the three soil-transmitted helminth species at high infection intensity, however sensitivity dropped to 53–80% in low intensity settings, being lowest for hookworm and A. lumbricoides. The highest sensitivity, overall and in both intensity groups, was observed for the FLOTAC method, whereas the sensitivity of the Mini-FLOTAC method was comparable with the Kato-Katz method. FLOTAC average egg count estimates were significantly lower compared with Kato-Katz, while the compared McMaster counts varied. In conclusion, we demonstrate that the Kato-Katz and Mini-FLOTAC methods had comparable sensitivities. We further show that test sensitivity of the Kato-Katz method is reduced in low transmission settings.     

Structural characterization of glycosphingolipids from the eggs of Schistosoma mansoni and Schistosoma japonicum

The granulomatous pathology in human intestinal schistosomiasisis induced primarily by the egg antigens of schistosome, a parasitictrematode. Glycolipids and glycoproteins were extracted fromthe eggs of the two major species which infect human, Schistosomamansoni and Schistosoma japonicum, for structural characterizationbased on highly sensitive mass spectrometric analysis coupledwith chemical derivatization. Here, we demonstrate that a seriesof uniquely multifucosylated glycosphingolipids constitute themajor egg glycolipids of S.mansoni but not of S.japonicum. TheS.mansoni glycosphingolipids were found to be extended by varyingnumbers of an unusual repeating unit,