Accurately estimating disease prevalence in finite populations is essential in epidemiology and ecological studies. Standard methods based on the binomial model often rely on assumptions of infinite populations and perfect diagnostic tests, which are frequently violated in practice. When sampling is without replacement from a finite population, the hypergeometric distribution provides the appropriate framework, but diagnostic misclassification must also be accounted for. In this paper, we present a framework for prevalence estimation in finite populations under imperfect diagnostic testing. The framework incorporates misclassification through diagnostic sensitivity and specificity, which may be treated either as fixed or as parameters estimated from independent studies. We evaluate multiple inferential approaches, including exact hypergeometric confidence intervals, a computationally efficient simulation-based approximation, a Bayesian formulation yielding posterior distributions, and a profile likelihood method that jointly accounts for uncertainty in sensitivity and specificity. Using extensive simulations we assess coverage, interval precision, and point estimate accuracy. Results show that hypergeometric-based methods consistently yield narrower and better-calibrated intervals than binomial-based alternatives. The profile likelihood approach achieves near-nominal coverage while appropriately propagating diagnostic uncertainty. An application to epidemiological surveillance data illustrates the practical relevance of the proposed methods.

In recent years, an increasing number of publications on the analysis of binary data have applied methods that take misclassification into account. However, potential misclassification is often ignored in study design due to the lack of sample size formulas or software. This may lead to a considerable loss of power in studies that only account for misclassification at the analysis stage. We argue that analyses correcting for misclassification should be used in combination with appropriate sample size adjustment in the design phase of the studies. We illustrate the importance of this by comparing the required sample sizes with and without misclassification, and provide an appropriate sample size procedure implemented as an R function for the one-sample and two-sample tests for binary endpoints. The sample size is calculated from the presumed binomial parameters (p0 and pa for one-sample and p1 and p2 for two-sample tests), the required power, and the probabilities of correct classification, sensitivity (Se), and specificity (Sp). Our results show that misclassification may drastically affect the necessary sample size in both testing scenarios.

IntroductionHeat stress in hutch-reared dairy calves (Bos taurus) is highly relevant due to its adverse effects on animal welfare, health, growth, and economic outcomes. This study aimed to provide arguments for protecting calves against heat stress. It was hypothesized that the thermal stress caused by high ambient temperature in summer months negatively affects the survival rate in preweaning calves.MethodsIn a retrospective study, we investigated how calf mortality varied by calendar month and between thermoneutral and heat stress periods on a large-scale Hungarian dairy farm (data of 46,899 calves between 1991 and 2015).ResultsThe daily mortality rate was higher in the summer (8.7–11.9 deaths per 10,000 calf days) and winter months (10.7–12.5 deaths per 10,000 calf-days) than in the spring (6.8–9.2 deaths per 10,000 calf-days) and autumn months (7.1–9.5 deaths per 10,000 calf-days). The distribution of calf deaths per calendar month differed between the 0–14-day and 15–60-day age groups. The mortality risk ratio was highest in July (6.92). The mortality risk in the 0–14-day age group was twice as high in periods with a daily mean temperature above 22°C than in periods with a daily mean of 5–18°C.ConclusionsHeat stress abatement is advised in outdoor calf rearing when the mean daily temperature reaches 22°C, which, due to global warming, will be a common characteristic of summer weather in a continental region.

It is postulated that there is negative correlation between milk yield and reproductive performance. However, some studies definitely doubt this causality. The aim of our study was to investigate the relationship between milk production and fertility on three dairy farms. The production parameter was the milk yield (in kg), and fertility was expressed by the number of inseminations per conception (AI index), as well as by the length of the service period (in days). A total of 13 012 lactations from cows with their first three lactations completed were analysed. The number of inseminations was significantly correlated with the milk yield and with the studied farm (p < 0.0001), but its correlation with the lactation number was not significant (p = 0.9477). A similar relationship was found after evaluating the length of the service period. A multiplicative model showed that a 2000 kg milk increase extended the service period by 9% and increased the AI index by 13%. Thereafter, using quartiles of the cows, the service period of the highest-producing group rose by 41.5 days, and the AI index by almost 1, compared to the lowest quartile. Our results indicate a definitive decline in reproductive indicators parallel to an increase in milk production but did not prove an inevitable correlation.

Climate change co-occurs with an advancement of avian breeding season (indexed as laying dates or fledging dates) in the temperate zone, suggesting a causality between them. Here, we investigate whether the long-term shifts in nestling (chick) ringing dates also mirror this phenomenon. This index is biased by inherent shortcomings, such as the non-independence of dates (in nestmates, colony members), poor accuracy (long period suitable for ringing), and strange shape of distributions. These shortcomings can be reduced by applying the median of annual ringing dates as an index of breeding phenology. The advantage of this index is that data are available for long periods and large sample sizes. By accepting certain compromise between statistical discipline and fieldwork realities, we examined changes in the breeding phenology of 9 bird species from 1951 to 2020 in Hungary. We found that the annual median of ringing dates advanced significantly (by 9–14 days) in the Black-headed Gull, Common Kestrel, Barn Swallow, Great Tit, and Eurasian Blue Tit. Contrarily, no significant (all P > 0.16) changes occurred in the case of the Common Tern, Black-crowned Night-heron, Common Buzzard, and Long-eared Owl. We also found that the proportion of Great Tits’ second brood has been reduced in recent decades.

In this study, the incidence, timing and risk factors associated with abortion and perinatal mortality (PM) were described in dromedary camels under intensive management. In addition, overall pregnancy losses were also summarized and weekly risk of pregnancy wastage was determined throughout gestation. Data were collected over 11 breeding seasons from September 2006 through June 2017 at the world's largest camel dairy farm. A total of 229 abortions were observed (5.05%) out of 4533 pregnancies after 60 days (d) of gestation. Most abortions were singleton (n = 199, 86.9%), but twin abortions were also recorded in 30 cases (13.1%). Abortions showed a pronounced seasonal distribution, with a peak in August. The age category (P < 0.01), breed or ecotype of the female (P < 0.05) and bull influenced the occurrence of singleton abortions. Dromedaries with twins tended to abort earlier than those with a singleton fetus (median = 232.5 d vs. 257 d, P = 0.053). Perinatal mortality was observed in 174 cases (3.84%) out of 4533 pregnancies after 60 d of gestation. The condition included the premature birth of non-viable calves after shorter than normal gestation (330-350 d, n = 26, 14.9%), the birth of well-developed but dead calves after normal gestation length (n = 120, 69.0%) and neonates that died within 48 h after delivery (n = 28, 16.1%). The frequency distribution of PM was parallel with that of parturitions. The most important predisposing factor for PM was difficult calving. Thirty-nine percent (68 out of 174) of these losses were associated with dystocia. In addition, age category (P < 0.05) and parity of the female (P < 0.01), month of delivery (P < 0.05) and breeding season (P < 0.05) also affected the incidence of PM. The cause of 60 cases of PM (1.4% of all deliveries) could not be determined and was considered idiopathic. In conclusion, one-third of total pregnancy losses occurred during mid to late gestation. Approximately 10% of pregnancies after Day 60 failed, and 90% resulted in the birth of a live calf that survived beyond 48 h. More than half of these pregnancy losses were abortions before 330 d of gestation, and approximately 40% were classified as PM. The weekly mean risk of pregnancy loss after 100 d of gestation remained only a fraction of that observed during the first 2-3 months.(c) 2022 Elsevier Inc. All rights reserved.

A tanulmányban ismertetünk egy új konfidenciaintervallumot a betegségek prevalenciabecslésének azon eseteire, amikor a felhasznált diagnosztikai teszt szenzitivitását és specificitását a vizsgálati mintától független mintákból becsüljük. Az új eljárás alapja a profil-likelihood módszer volt, az intervallum lefedési valószínűségének javítása érdekében pedig további korrekciót is alkalmaztunk. Az intervallum lefedési valószínűségét és várható hosszát szimulációval értékeltük, és összehasonlítottuk két másik, ugyanerre a problémára javasolt módszerrel, nevezetesen Lang és Reiczigel (2014), illetve Flor és munkatársai (2020) módszereivel. Az új intervallum várható hossza rövidebb, mint a Lang−Reiczigel-féle intervallumé, miközben lefedési valószínűségük nagyjából megegyezik. A Flor-féle intervallummal összehasonlítva a várható hossz hasonló, de az új intervallum lefedési valószínűsége nagyobb. Összességében tehát az új intervallum mindkét versenytársánál jobbnak bizonyult.

Potential misclassification of a binary outcome measure is often ignored in study design, causing considerable loss of power, and threatening the quality of research. Although there exist studies taking misclassification into account in data analysis, we argue that it should be accounted for already in sample size calculation. We illustrate this by comparing sample sizes needed with and without misclassification in case of the binomial test. Our sample size procedure, implemented as an R function, calculates exact power, and accounts for non-monotonicity of power as a function of sample size, and for potential drop-out or lack of data in the study. The necessary sample size is computed from the null proportion p , the assumed true proportion p , and the probabilities of correct classification, sensitivity ( Se) and specificity ( Sp) . Our results show that misclassification may drastically affect the necessary sample size. For p <0.5, the effect of specificity is stronger than that of sensitivity, whereas for p >.5 it is the other way round. Effects are strongest when p is near 0 or 1, especially for one-sided tests with p located farther from 0.5 than the null value p . For example, even with Se = Sp = 99%, p = 0.01, and left-sided alternative, sample size is more than fourfold of that without misclassification (3-fold if p =0.02; 1.4-fold if p =0.05).

We present a new confidence interval for the prevalence of a disease for a situation when sensitivity and specificity of the diagnostic test are estimated from validation samples independent of the study sample. The new interval is based on profile likelihood and incorporates an adjustment improving the coverage probability. Its coverage probability and expected length were assessed by simulation and compared to two other methods for this problem, namely those by Lang and Reiczigel (2014) and Flor et al. (2020). Expected length of the new interval is less than that of the Lang and Reiczigel interval while its coverage is about the same. Comparison to the Flor interval resulted in similar expected length but higher coverage probabilities for the new interval. All in all, the new interval proved to be better than both its competitors.

BACKGROUND:Albuminuria is an important marker of renal damage and can precede proteinuria; thus, it can be a useful analyte in the early diagnosis of kidney diseases. Albuminuria has also been found in dogs with hypertension, inflammatory, infectious, and neoplastic diseases. OBJECTIVES:The aim of this study was to establish a reference interval (RI) for albuminuria in dogs. METHODS:One hundred sixty-four clinically healthy dogs were enrolled in the study. Urinary albumin was determined by the immunoturbidimetric method, and albumin excretion was expressed as the urinary albumin-to-creatinine (UAC) ratio. The RI for UAC was established. RESULTS:After exclusions, 124 dogs from 32 breeds remained. The median UAC of the study population was 3.0 mg/g (range: 0-48). The RI was defined as 0-19 mg/g (with a 90% CI for the upper limit of 13-28 mg/g). No significant difference was found between male and female dogs or between different age and body weight groups. The results of Sighthounds (n = 30) and Beagle dogs (n = 23) did not differ from the other breeds. CONCLUSION:The canine RI of UAC is similar but somewhat narrower than the human RI.

This retrospective study was performed on 71 dogs which had been admitted for heartworm screening or with clinical suspicion of heartworm disease. The examination methods included polymerase chain re-action (PCR) to identify Dirofilaria immitis and/or Dirofilaria repens infections and a heartworm antigen (Ag) test (VetScan). By using PCR, 26 dogs were found positive only for Dirofilaria immitis (Group 1), while 21 dogs for both D. immitis and D. repens (Group 2). Group 3 included 24 dogs with D. repens infection only according to the PCR results. The sensitivity of the VetScan Ag test for the Group 1 and 2 animals proved to be 97.7% (95% Blaker confidence interval; CI 89.0%-99.9%). The specificity of the VetScan Ag test, calculated from the results of Group 3, was found to be 66.7% (95% CI 45.6%-83.1%), which was lower than that reported from the USA, where D. repens does not occur. In cases when PCR results were positive for D. repens but negative for D. immitis, the occult dirofilariosis was the likely explanation for the positive D. immitis Ag tests. These observations highlight the importance of per-forming more Ag tests simultaneously in those areas where both Dirofilaria species are present.

The large-scale farming and high population density on farms require both fast and cost-effective screening and monitoring methods to detect infectious diseases, while also ensuring that the number of animals tested within a herd is as large as possible. This paper reviews the methods and possibilities of testing grouped or aggregated i.e. pooled samples. A pool can be artificial or natural e.g. aggregating equal volume of sera samples is artificial pooling, while bulk milk or oral fluid collected by rope can be considered as a natural pool. Following a brief historical review, the currently used methods of collection and examination of pooled samples and the statistical evaluation of their laboratory results are discussed. Testing of pooled samples retrospect until 1943 in human medicine, however, the world-wide spreading of HIV from the 1980-s gave this method a real boost. The laboratory tests became sensitive and specific enough to detect even one positive sample in a large group of negatives. On the other hand, some of these tests were quite expensive, therefore the intention of reducing the cost per test was justified.Based on the aim of the pooling one can distinguish(a) screening or classification, when the status of each individual subject should be established (e.g. BVD PI screening) or(b) prevalence estimation within a population (e.g. what proportion of a vector population is affected by an infectious agent) or(c) confirmation of the infected or free status of a population (a herd or a flock e.g. a poultry flock is infected with Salmonella enteritidis). Neither the prevalence nor the individual animal status is investigated.Papers about the testing of oral fluid and processing fluid samples from pigs, as well as the Trichinella examination at slaughter are reviewed. Pooled sample testing for Bovine Viral Diarrhoea (BVD) and paratuberculosis as well as poultry flock check for salmonellosis are described.Statistical methods of calculation of individual prevalence estimation from pool prevalence including confidence intervals and optimum pool size estimation are discussed.

In this paper, we described the incidence of early pregnancy loss (EPL) both after natural mating and embryo transfer, evaluated risk factors, and summarized the outcome of twin pregnancies throughout gestation in dromedaries under reproductive care. Data were collected over seven breeding seasons at the world's largest camel dairy farm (study 1). In addition, we determined the timing of EPL and monitored serum progesterone (P4) concentration between Days 13-70 of gestation during one breeding season (study 2). In the first study, out of 2970 pregnancies, 507 cases (17.1%) of EPL were diagnosed with transrectal ultrasonography. The rate of EPL after natural mating and embryo transfer was 16.1% (n = 422 out of 2616) and 24.0% (n = 85 out of 354), respectively. Twin pregnancies were detected in 215 cases (7.2% of all gestations), and 57 of those (26.5%) underwent complete EPL. Almost half of the early losses (n = 243; 47.9%) occurred before 30 d of gestation. Another 43.2% (n = 219) of EPL was diagnosed during the next month, and 8.9% (n = 45) occurred after 60 d of gestation. Multivariable mixed effects logistic regression models revealed that the breeding season (year) and twin pregnancy were the most important exposure variables affecting the rate of EPL (P < 0.001). The effect of some male camels was also demonstrated while other factors, such as type of breeding, age category, month of mating, breed/ecotype and reproductive history did not prove to have a significant influence. In the second study, the overall rate of EPL was 24.5% (n = 34 of 139). There was no difference in the incidence of EPL between ET recipient (24.2%, n = 23 of 95) and mated (25%, n = 11 of 44) camels. Weekly rate of EPL ranged from 0.9% to 4.8% with a decreasing tendency, and approx. 41% of the animals (n = 14 of 34) had some ultrasonographic signs of impending EPL 1 week before the final diagnosis. Mean serum P4 concentration in camels with subsequent EPL was 5.3 +/- 0.1 ng/ml compared to 5.6 +/- 0.04 ng/ml in normal pregnant dromedaries. Day of gestation and future EPL influenced serum P4 levels (P < 0.001) with an interaction between the two fixed factors (P < 0.05). At the time of the final diagnosis of EPL, mean serum P4 concentration was 2.8 +/- 0.44 ng/ml. Although twinning had an unfavorable prognosis with a total pregnancy loss of 36.7%, it was not entirely detrimental for the final outcome of gestation as two-thirds of twin pregnancies (n = 136 out of 212) resulted in the birth of a live calf. (c) 2021 Elsevier Inc. All rights reserved.

Exact two-tailed tests and two-sided confidence intervals (CIs) for a binomial proportion or Poisson parameter by Sterne (Biometrika 41:117–129, 1954) or Blaker (Can J Stat 28(4):783–798, 2000) are successful in reducing conservatism of the Clopper–Pearson method. However, the methods suffer from an inconsistency between the tests and the corresponding CIs: In some cases, a parameter value is rejected by the test, though it lies in the CI. The problem results from non-unimodality of the test p value functions. We propose a slight modification of the tests that avoids the inconsistency, while preserving nestedness and exactness. Fast and accurate algorithms for both the test modification and calculation of confidence bounds are presented together with their theoretical background.

The aggregated distributions of host-parasite systems require several different infection parameters to characterize them. We advise readers how to choose infection indices with clear and distinct biological interpretations, and recommend statistical tests to compare them across samples. A user-friendly and free software is available online to overcome technical difficulties.

Rudas, Clogg, and Lindsay (1994, J. R Stat Soc. Ser. B, 56 , 623) introduced the so-called mixture index of fit, also known as pi-star (π*), for quantifying the goodness of fit of a model. It is the lowest proportion of ‘contamination’ which, if removed from the population or from the sample, makes the fit of the model perfect. The mixture index of fit has been widely used in psychometric studies. We show that the asymptotic confidence limits proposed by Rudas et al . (1994, J. R Stat Soc. Ser. B, 56 , 623) as well as the jackknife confidence interval by Dayton ( 2003 , Br. J. Math. Stat. Psychol., 56 , 1) perform poorly, and propose a new bias-corrected point estimate, a bootstrap test and confidence limits for pi-star. The proposed confidence limits have coverage probability much closer to the nominal level than the other methods do. We illustrate the usefulness of the proposed method in practice by presenting some practical applications to log-linear models for contingency tables.

Rudas, Clogg, and Lindsay (1994, J. R Stat Soc. Ser. B, 56, 623) introduced the so-called mixture index of fit, also known as pi-star (π*), for quantifying the goodness of fit of a model. It is the lowest proportion of 'contamination' which, if removed from the population or from the sample, makes the fit of the model perfect. The mixture index of fit has been widely used in psychometric studies. We show that the asymptotic confidence limits proposed by Rudas et al. (1994, J. R Stat Soc. Ser. B, 56, 623) as well as the jackknife confidence interval by Dayton (, Br. J. Math. Stat. Psychol., 56, 1) perform poorly, and propose a new bias-corrected point estimate, a bootstrap test and confidence limits for pi-star. The proposed confidence limits have coverage probability much closer to the nominal level than the other methods do. We illustrate the usefulness of the proposed method in practice by presenting some practical applications to log-linear models for contingency tables.

We monitored the major chemical composition of bulk dromedary camel milk by FT-MIR spectroscopy over a 5-year period. The results highly correlated with those determined with reference methods (r > 0.985, p < 0.001). Production parameters showed significant (p < 0.001) seasonal and yearly changes. The overall mean fat, protein, lactose, solids-not-fat, and total solids concentrations of bulk dromedary camel milk were 2.87%, 2.94%, 4.15%, 8.00%, and 10.69%, respectively. Month of the year, year of the study, and level of production had a strong influence on bulk milk chemical composition and yield of milk components; however, the relative effect of season on composition was greater (proportion of variance app. 50%) compared to that of other factors of variation. The highest and lowest values were measured during winter and summer, respectively. Circannual variation in major milk components was associated with environmental conditions (photoperiod, temperature), whereas it was independent of nutritional factors.

In ecology, diversity is often measured as the mean rarity of species in a community. In behavioral sciences and parasitology, mean crowding is the size of the group to which a typical individual belongs. In this paper, focusing mostly on the mathematical aspect, we demonstrate that diversity and crowding are closely related notions. We show that mean crowding can be transformed into diversity and vice versa. Based on this general equivalence rule, notions, relationships, and methods developed in one field can be adapted to the other one. In relation to crowding, we introduce the notion "effective number of groups" that corresponds to the "effective number of species" used in diversity studies. We define new aggregation indices that mirror evenness indices known from diversity theory. We also construct aggregation profiles and orderings of populations based on aggregation indices. By uniting the mathematical interpretation of the ecological notion of diversity and the ethological notion of typical group size (or crowding, in parasitology), our insight opens a new avenue of both theoretical and methodological research. This is exemplified here using real-life abundance data of avian parasites.