Biomedical subjects

Wesley O Johnson

Publications and source records attributed to Wesley O Johnson.

17 recordsLinked to original sources

PubMed2006

Sample size calculations for disease freedom and prevalence estimation surveys.

We developed a Bayesian approach to sample size calculations for studies designed to estimate disease prevalence that uses a hierarchical model for estimating the proportion of infected clusters (cluster-level prevalence) within a country or region. The clusters may, for instance, be villages within a region, cities within a state, or herds within a country. Our model allows for clusters with zero prevalence and for variability in prevalences among infected clusters. Moreover, uncertainty about diagnostic test accuracy and within-cluster prevalences is accounted for in the model. A predictive approach is used to address the issue of sample size selection in human and animal health surveys. We present sample size calculations for surveys designed to substantiate freedom of a region from an infectious agent (disease freedom surveys) and for surveys designed to estimate cluster-level prevalence of an endemic disease (prevalence estimation surveys). In disease freedom surveys, for instance, assuming the cluster-level prevalence for a particular infectious agent in the region is greater than a maximum acceptable threshold, a sample size combination consisting of the number of clusters sampled and number of subjects sampled per cluster can be determined for which authorities conducting the survey detect this excessive cluster-level prevalence with high predictive probability. The method is straightforward to implement using the Splus/R library emBedBUGS together with WinBUGS.

Animals↗

PubMed2006

Diagnosis using predictive probabilities without cut-offs.

Standard diagnostic test procedures involve dichotomization of serologic test results. The critical value or cut-off is determined to optimize a trade off between sensitivity and specificity of the resulting test. When sampled units from a population are tested, they are allocated as either infected or not according to the test outcome. Units with values high above the cut-off are treated the same as units with values just barely above the cut-off, and similarly for values below the cut-off. There is an inherent information loss in dichotomization. We thus develop a diagnostic screening method based on data that are not dichotomized within the Bayesian paradigm. Our method determines the predictive probability of infection for each individual in a sample based on having observed a specific serologic test result and provides inferences about the prevalence of infection in the population sampled. Our fully Bayesian method is briefly compared with a previously developed frequentist method. We illustrate the methodology with serologic data that have been previously analysed in the veterinary literature, and also discuss applications to screening for disease in humans. The method applies more generally to a variation of the classic parametric 2-population discriminant analysis problem. Here, in addition to training data, additional units are sampled and the goal is to determine their population status, and the prevalence(s) of the subpopulation(s) from which they were sampled.

Animals↗

PubMed2005

Sample size determination for estimation of the accuracy of two conditionally independent tests in the absence of a gold standard.

We developed an Excel spreadsheet template (available at http://www.epi.ucdavis.edu/diagnostictests/) to calculate sample sizes to estimate sensitivity and specificity with desired precision in the absence of a gold standard. Calculations are predicated on the use of two conditionally independent tests for screening animals from two populations and are based on the methods of Hui and Walter(1980). Sample size calculations rely on asymptotic normality of maximum likelihood (ML) estimates of parameters. Spreadsheets for calculating standard errors for the parameter estimates and for providing ML estimates using cross-tabulated data also are included. An example of application of the methods to bovine paratuberculosis is presented.

Animals↗

PubMed2005

Bayesian inference for prevalence and diagnostic test accuracy based on dual-pooled screening.

We propose a useful protocol for the problem of screening populations for low-prevalence characteristics such as HIV or drugs. Current HIV screening of blood that has been donated for transfusion involves the testing of individual blood units with an inexpensive enzyme-linked immunosorbent assay test and follow-up with a more accurate and more expensive western blot test for only those units that tested positive. Our cost-effective pooling strategy would enhance current methods by making it possible to accurately estimate the sensitivity and specificity of the initial screening test, and the proportion of defective units that have passed through the system. We also provide a method of estimating the distribution of prevalences for the characteristic throughout the population or subpopulations of interest.

AIDS Serodiagnosis↗

PubMed2004

Diagnostic test accuracy and prevalence inferences based on joint and sequential testing with finite population sampling.

The two-test two-population model, originally formulated by Hui and Walter, for estimation of test accuracy and prevalence estimation assumes conditionally independent tests, constant accuracy across populations and binomial sampling. The binomial assumption is incorrect if all individuals in a population e.g. child-care centre, village in Africa, or a cattle herd are sampled or if the sample size is large relative to population size. In this paper, we develop statistical methods for evaluating diagnostic test accuracy and prevalence estimation based on finite sample data in the absence of a gold standard. Moreover, two tests are often applied simultaneously for the purpose of obtaining a 'joint' testing strategy that has either higher overall sensitivity or specificity than either of the two tests considered singly. Sequential versions of such strategies are often applied in order to reduce the cost of testing. We thus discuss joint (simultaneous and sequential) testing strategies and inference for them. Using the developed methods, we analyse two real and one simulated data sets, and we compare 'hypergeometric' and 'binomial-based' inferences. Our findings indicate that the posterior standard deviations for prevalence (but not sensitivity and specificity) based on finite population sampling tend to be smaller than their counterparts for infinite population sampling. Finally, we make recommendations about how small the sample size should be relative to the population size to warrant use of the binomial model for prevalence estimation.

Agglutination Tests↗

PubMed2004

Factors affecting menstrual cycle characteristics.

This 1989-1991 study in California and Utah used daily urinary metabolites of estrogen and progesterone and computer algorithms to assess ovulatory status and day of ovulation. The authors examined the associations of risk factors with menstrual cycle characteristics for 309 working women aged 20-44 years who collected a median of five cycles each of daily urine samples. Linear mixed models were used to assess continuous menstrual outcomes. Compared with women less than age 35 years, women aged 35 years or older had a significantly decreased (-0.94 days, 95% confidence interval: -1.83, -0.05) adjusted mean cycle length. Age modified the effects of smoking, physical activity, ethnicity, and alcohol consumption on mean follicular phase length. Asian women had a significantly longer (1.65 days, 95% confidence interval: 0.54, 2.76) adjusted mean cycle length compared with Caucasian women. Compared with women who did not consume alcoholic drinks, women who did had a significantly shorter (-1.26 days, 95% confidence interval: -2.21, -0.31) adjusted mean cycle length. Mean cycle and phase lengths were significantly associated with length of the prior luteal phase. These results indicate that potentially modifiable risk factors, as well as immutable host factors, are associated with menstrual cycle characteristics that may in turn be related to subsequent disease risk.

Adult↗

PubMed2004

Bayesian analysis of risk factors for anovulation.

Two algorithms for assessing ovulatory status using daily urinary levels of oestrogen and progesterone metabolites have been applied to non-clinic-based, free-living populations of women. These relatively new methods for assessing ovarian function have been used to assess the potential adverse effects of occupational and environmental exposures, such as smoking, on the reproductive health of women. One algorithm has been validated against serum hormone measurements and gives good sensitivity and specificity for anovulation. However, a gold standard is generally not available in epidemiologic field studies in which these daily urine samples are collected. In this paper, we used Bayesian methods to estimate: (i) the probability of occurrence of anovulation, (ii) the sensitivity and specificity of the two algorithms, and (iii) the association between anovulation and smoking and other risk factors in the absence of a perfect test. We evaluated the two published algorithms for assessing ovulatory status, based on their cross-classified results applied to one randomly selected cycle from each woman in a sample of 338 employed women. We first assumed that the algorithms were independent, conditional on ovulatory status. Then, we used a dependence model to allow for correlation between the results of the two algorithms. We implemented a Bayesian logistic regression analysis that allowed the outcome measurement to be partially imperfect. We incorporated the posterior distributions for algorithm accuracy obtained from the dependence model as prior distributions for this logistic regression model. Then, we compared the results with those obtained from a standard multiple logistic approach using the algorithm determination of ovulatory status as if it were perfect. Our results indicated that increasing physical activity was associated with a significantly increased risk of anovulation; and smokers had a potentially, but not statistically significant, increased occurrence of anovulation.

Adult↗

PubMed2004

Modelling risk when binary outcomes are subject to error.

We present methods for binomial regression when the outcome is determined using the results of a single diagnostic test with imperfect sensitivity and specificity. We present our model, illustrate it with the analysis of real data, and provide an example of WinBUGS program code for performing such an analysis. Conditional means priors are used in order to allow for inclusion of prior data and expert opinion in the estimation of odds ratios, probabilities, risk ratios, risk differences, and diagnostic test sensitivity and specificity. A simple method of obtaining Bayes factors for link selection is presented. Methods are illustrated and compared with Bayesian ordinary binary regression using data from a study of the effectiveness of a smoking cessation program among pregnant women. Regression coefficient estimates are shown to change noticeably when expert prior knowledge and imperfect sensitivity and specificity are incorporated into the model.

Adult↗

PubMed2004

Sample size calculations for surveys to substantiate freedom of populations from infectious agents.

We develop a Bayesian approach to sample size computations for surveys designed to provide evidence of freedom from a disease or from an infectious agent. A population is considered "disease-free" when the prevalence or probability of disease is less than some threshold value. Prior distributions are specified for diagnostic test sensitivity and specificity and we test the null hypothesis that the prevalence is below the threshold. Sample size computations are developed using hypergeometric sampling for finite populations and binomial sampling for infinite populations. A normal approximation is also developed. Our procedures are compared with the frequentist methods of Cameron and Baldock (1998a, Preventive Veterinary Medicine34, 1-17.) using an example of foot-and-mouth disease. User-friendly programs for sample size calculation and analysis of survey data are available at http://www.epi.ucdavis.edu/diagnostictests/.

Animals↗

PubMed2004

Effect of multiple sampling on diagnostic sensitivity.

Overall diagnostic sensitivity is the probability that a diagnostic procedure will detect an agent if the tested animal is indeed infected. The overall or effective sensitivity is a function of both the probability that the assay will detect the agent if it is present in the sample tested and the probability that the agent will be present in the sample tested if the animal is infected with the agent. Thus, even with a highly sensitive assay, the probability of detecting an infected animal may be low or nil if the sampling procedure failed to capture the agent in samples tested by the assay. In this article, it is demonstrated how increased frequency of testing, such as testing multiple subsamples, can have a profound effect on increasing the overall sensitivity of a diagnostic procedure.

Animals↗

PubMed2003

A mixture model for bovine abortion and foetal survival.

The effect of spontaneous abortion on the dairy industry is substantial, costing the industry on the order of US dollars 200 million per year in California alone. We analyse data from a cohort study of nine dairy herds in Central California. A key feature of the analysis is the observation that only a relatively small proportion of cows will abort (around 10;15 per cent), so that it is inappropriate to analyse the time-to-abortion (TTA) data as if it were standard censored survival data, with cows that fail to abort by the end of the study treated as censored observations. We thus broaden the scope to consider the analysis of foetal lifetime distribution (FLD) data for the cows, with the dual goals of characterizing the effects of various risk factors on (i). the likelihood of abortion and, conditional on abortion status, on (ii). the risk of early versus late abortion. A single model is developed to accomplish both goals with two sets of specific herd effects modelled as random effects. Because multimodal foetal hazard functions are expected for the TTA data, both a parametric mixture model and a non-parametric model are developed. Furthermore, the two sets of analyses are linked because of anticipated dependence between the random herd effects. All modelling and inferences are accomplished using modern Bayesian methods.

Abortion, Veterinary↗

PubMed2003

Quantification, risk factors, and health impact of natural congenital infection with bovine viral diarrhea virus in dairy calves.

OBJECTIVES: To estimate risk and identify risk factors for congenital infection with bovine viral diarrhea virus (BVDV) not resulting in persistent infection and examine effect of congenital infection on health of dairy calves. ANIMALS: 466 calves. PROCEDURES: Calves from 2 intensively managed drylot dairies with different vaccination programs and endemic BVDV infection were sampled before ingesting colostrum and tested with their dams for BVDV and BVDV serum-neutralizing antibodies. Records of treatments and death up to 10 months of age were obtained from calf ranch or dairy personnel. Risk factors for congenital infection, including dam parity and BVDV titer, were examined by use of logistic regression analysis. Effect of congenital infection on morbidity and mortality rates was examined by use of survival analysis methods. RESULTS: Fetal infection was identified in 10.1% of calves, of which 0.5% had persistent infection and 9.6% had congenital infection. Although dependent on herd, congenital infection was associated with high BVDV type 2 titers in dams at calving and with multiparous dams. Calves with congenital infection had 2-fold higher risk of a severe illness, compared with calves without congenital infection. CONCLUSIONS AND CLINICAL RELEVANCE: The unexpectedly high proportion of apparently healthy calves found to be congenitally infected provided an estimate of the amount of fetal infection via exposure of dams and thus virus transmission in the herds. Findings indicate that congenital infection with BVDV may have a negative impact on calf health, with subsequent impact on herd health.

Animals↗

PubMed2003

Estimation of receiver-operating characteristic curves to determine accuracy of a competitive enzyme-linked immunosorbent assay for the serodiagnosis of Brucella infection in domestic water buffalo (Bubalus bubalis) and cattle.

OBJECTIVE: To estimate receiver-operating characteristic (ROC) curves for a competitive ELISA (c-ELISA) that is used in serodiagnosis of brucellosis in water buffalo and cattle, to determine the most appropriate positive cutoff value for the c-ELISA in confirmation of infection, and to evaluate species differences in c-ELISA function. SAMPLE POPULATION: Sera from 4 herds of cattle (n = 391) and 4 herds of water buffalo (381). PROCEDURE: Serum samples were evaluated for Brucella-specific antibodies by use of a c-ELISA. On the basis of previous serologic test results, iterative simulation modeling was used to classify animals as positive or negative for Brucella infection without the use of a gold standard. Accuracy of c-ELISA for diagnosis of infection was compared between cattle and water buffalo by comparison of areas under ROC curves. RESULTS: A positive cutoff value of 30% inhibition for c-ELISA yielded sensitivity and specificity estimates, respectively, of 83.9 and 92.6% for cattle and 91.4 and 95.4% for water buffalo. A positive cutoff value of 35% inhibition yielded sensitivity and specificity estimates, respectively, of 83.9 and 96.2% for cattle and 88.0 and 974% for water buffalo. Areas under ROC curves were 0.94 and 0.98 for cattle and water buffalo, respectively. CONCLUSIONS AND CLINICAL RELEVANCE: ROC curves can be estimated by use of iterative simulation methods to determine optimal cutoff values for diagnostic tests with quantitative outcomes. A cutoff value of 35% inhibition for the c-ELISA was found to be most appropriate for confirmation of Brucella infection in cattle and water buffalo.

Animals↗

PubMed2002

Predicted ages of dairy calves when colostrum-derived bovine viral diarrhea virus antibodies would no longer offer protection against disease or interfere with vaccination.

OBJECTIVE: To develop models that could be used to predict, for dairy calves, the age at which colostrum-derived bovine viral diarrhea virus (BVDV) antibodies would no longer offer protection against infection or interfere with vaccination. DESIGN: Prospective observational field study. ANIMALS: 466 calves in 2 California dairy herds. PROCEDURE: Serum BVDV neutralizing antibody titers were measured from birth through 300 days of age. The age by which colostrum-derived BVDV antibodies had decayed sufficiently that calves were considered susceptible to BVDV infection (ie, titer < or = 1:16) or calves became seronegative was modeled with survival analysis methods. Mixed-effects regression analysis was used to model colostrum-derived BVDV antibody titer for any given age. RESULTS: Half the calves in both herds became seronegative for BVDV type I by 141 days of age and for BVDV type II by 114 days of age. Rate of antibody decay was significantly associated with antibody titer at 1 to 3 days of age and with whether calves were congenitally infected with BVDV. Three-month-old calves were predicted to have a mean BVDV type-I antibody titer of 1:32 and a mean BVDV type-II antibody titer of 1:16. CONCLUSIONS AND CLINICAL RELEVANCE: Results provide an improved understanding of the decay of BVDV-specific colostrum-derived antibodies in dairy calves raised under typical field conditions. Knowledge of the age when the calf herd becomes susceptible can be useful when designing vaccination programs aimed at minimizing negative effects of colostrum-derived antibodies on vaccine efficacy while maximizing overall calf herd immunity.

Age Factors↗

PubMed2002

Predictive influence in the accelerated failure time model.

We develop case deletion diagnostics for prediction of future observations in the accelerated failure time model. We view prediction to be an important inferential goal in a survival analysis and thus it is important to identify whether particular observations may be influencing the quality of predictions. We use the Kullback-Leibler divergence as a measure of the discrepancy between the estimated probability distributions for the full and the case-deleted samples. In particular, we focus on the effect of case deletion on estimated survival curves but where we regard the survival curve estimate as a vehicle for prediction. We also develop a diagnostic for assessing the effect of case deletion on inferences for the median time to failure. The estimated median can be used with both predictive and estimative purposes in mind. We also discuss the relationship between our suggested measures and the corresponding Cook distance measure, which was designed with the goal of assessing estimative influence. Several applications of the proposed diagnostics are presented.

Journal Article↗

PubMed2002

Comparison of serologic tests for detection of Brucella infections in cattle and water buffalo (Bubalus bubalis).

OBJECTIVE: To estimate sensitivity and specificity of 4 commonly used brucellosis screening tests in cattle and domestic water buffalo of Trinidad, and to compare test parameter estimates between cattle and water buffalo. ANIMALS: 391 cattle and 381 water buffalo. PROCEDURE: 4 Brucella-infected herds (2 cattle and 2 water buffalo) and 4 herds (2 of each species) considered to be brucellosis-free were selected. A minimum of 100 animals, or all animals > 1 year of age, were tested from each herd. Serum samples were evaluated for Brucella-specific antibodies by use of standard plate agglutination test (SPAT), card test (CT), buffered plate agglutination test (BPAT), and standard tube agglutination test (STAT). A Bayesian approach was used to estimate sensitivity and specificity of diagnostic tests without the use of a gold standard, assuming conditional independence of tests. RESULTS: Sensitivity and specificity estimates in cattle, respectively, were SPAT, 66.7 and 98.9; CT, 72.7 and 99.6; BPAT, 88.1 and 98.1; and STAT, 80.2 and 99.3. Corresponding test estimates in water buffalo, respectively, were SPAT, 51.4 and 99.3; CT, 90.4 and 99.4; BPAT, 96.3 and 90.7; and STAT, 75.0 and 98.8. Sensitivity of the CT and specificity of the BPAT were different between cattle and water buffalo with at least 95% probability. CONCLUSIONS AND CLINICAL RELEVANCE: Brucellosis serologic test performance varied by species tested, but BPAT had the highest sensitivity for screening cattle and water buffalo. Sensitivity and specificity of more than 2 screening tests can be estimated simultaneously without a gold standard by use of Bayesian techniques.

Agglutination Tests↗

PubMed2002

A method of probability diagnostic assignment that applies bayes theorem for use in serologic diagnostics, using an example of Neospora caninum infection in cattle.

OBJECTIVE: To develop a method of probability diagnostic assignment (PDA) that uses continuous serologic measures and infection prevalence to estimate the probability of an animal being infected, using Neospora caninum as an example. ANIMALS: 196 N caninum-infected beef and dairy cattle and 553 cattle not infected with N caninum; 50 dairy cows that aborted and 50 herdmates that did not abort. PROCEDURE: Probability density functions corresponding to distributions of N caninum kinetic ELISA results from infected and uninfected cattle were estimated by maximum likelihood methods. Maximum likelihood methods also were used to estimate N caninum infection prevalence in a herd that had an excessive number of abortions. Density functions and the prevalence estimate were incorporated into Bayes formula to calculate the conditional probability that a cow with a particular ELISA value was infected with N caninum. RESULTS: Probability functions identified for infected and uninfected cattle were Weibull and inverse gamma functions, respectively. Herd prevalence was estimated, and probabilities of N caninum infection were determined for cows with various ELISA values. CONCLUSIONS AND CLINICAL RELEVANCE: Use of PDA offers an advantage to clinicians and diagnosticians over traditional seronegative or seropositive classifications used as a proxy for infection status by providing an assessment of the actual probability of infection. The PDA permits use of all diagnostic information inherent in an assay, thereby eliminating a need for estimates of sensitivity and specificity. The PDA also would have general utility in interpreting results of any diagnostic assay measured on a continuous or discrete scale.

Abortion, Veterinary↗