SEARCH · Search PubMed

Results for “Bayes Theorem”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 181 records · Page 10Linked to original sources

PubMed1986

Evaluating batteries of short-term genotoxicity tests.

Selecting a battery of short-term genotoxicity tests suitable for screening unknown chemicals for carcinogenicity can be a large combinatorial task because of the great number of short-term tests currently available. Biological criteria, such as requirements for different targets and endpoints, can reduce the number of possible combinations. An independent yet potentially complementary approach which we have developed uses Bayes' theorem to predict carcinogenicity from results (positive or negative) in short-term tests. Batteries can be evaluated by their predictivity, calculated with Bayes' theorem from the sensitivities and specificities of the component tests. Our analyses indicate that tests which contribute most to a battery's predictivity are those which are both sensitive and specific, which we call Class I tests. Because few of the currently available tests are Class I, we have extended our analyses to consider when other types of tests must be substituted for Class I tests, the purpose of a particular test program will influence the choice. In order to obtain the best predictions for all chemicals, a battery should include an equal number of Class II tests (i.e. those that are sensitive but are not specific) and Class III tests (i.e. those that are not sensitive but are specific). However, for the purpose of reducing the number of carcinogens erroneously classified as non-carcinogenic, Class II tests contribute about twice as much to the predictivity of a battery as do Class III tests, and for the purpose of reducing the number of non-carcinogens erroneously classified as carcinogenic, Class II tests contribute about half as much as Class III tests.(ABSTRACT TRUNCATED AT 250 WORDS)

Mutagenicity Tests↗

PubMed1983

Clinical trials and statistical verdicts: probable grounds for appeal.

Conventional interpretation of clinical trials relies heavily on the classic p value. The p value, however, represents only a false-positive rate, and does not tell the probability that the investigator's hypothesis is correct, given his observations. This more relevant posterior probability can be quantified by an extension of Bayes' theorem to the analysis of statistical tests, in a manner similar to that already widely used for diagnostic tests. Reanalysis of several published clinical trials according to Bayes' theorem shows several important limitations of classic statistical analysis. Classic analysis is most misleading when the hypothesis in question is already unlikely to be true, when the baseline event rate is low, or when the observed differences are small. In such cases, false-positive and false-negative conclusions occur frequently, even when the study is large, when interpretation is based solely on the p value. These errors can be minimized if revised policies for analysis and reporting of clinical trials are adopted that overcome the known limitations of classic statistical theory with applicable bayesian conventions.

Aspirin↗

PubMed2001

Serum alanine aminotransferase in hepatitis c screening of patients on hemodialysis.

It is recommended that patients on hemodialysis (HD) therapy undergo regular screening for hepatitis C virus (HCV) infection by using alanine aminotransferase (ALT) values. However, the utility of using ALT values in this setting is unknown. The aim of this prospective study at the University of California Los Angeles Hepatitis Screening Program is to determine the sensitivity, specificity, and predictive values of an elevated ALT level for the diagnosis of HCV infection in HD patients. We screened 2,440 HD patients from 39 dialysis centers for viral infection by using hepatitis antibody serological testing and ALT values. We found the sensitivity and specificity of a newly elevated ALT level for acute HCV infection to be 83% and 90%, respectively. According to Bayes' theorem, the positive predictive value was 4% and the positive likelihood ratio was 8.74. For chronic HCV infection, the sensitivity of a newly elevated aminotransferase level was 21%, and specificity was 91%. The positive predictive value was 16% (according to Bayes' theorem), and the positive likelihood ratio was 2.47. The negative predictive value of a newly elevated aminotransferase value was 99% for acute HCV infection and 94% for chronic HCV infection. Our results indicate that although a newly elevated aminotransferase level is sensitive and specific for acute HCV infection, its positive predictive value is inadequate. A newly elevated aminotransferase level was neither sensitive nor positively predictive of chronic infection. Therefore, an elevated ALT level is an ineffective method for screening for HCV infection in HD patients.

Alanine Transaminase↗

PubMed1987

Human versus machine: a comparison of a computer 'expert system' with human experts in the diagnosis of vaginal discharge.

A model system has been designed which generates 'cases' of vaginal discharge. Each such case is presented to a human for diagnosis, and this is then compared with a computer diagnosis using two forms of Bayes' theorem. Six subjects (2 medical; 4 non-medical) participated in the trial and each examined 100 successive 'cases'. When the humans had forewarning of the trial and full access to the knowledge-base their performance was superior to that of Bayes' theorem using positive features only and equivalent to that using both positive and negative features. When the trial was repeated without forewarning the human performance was markedly inferior to that of the machine. It is concluded: that human and computer-aided diagnosis can be of approximately equal efficiency for complex and non-definitive data; that the imperfections of human memory give an obvious potential advantage to the machine in this type of situation.

Computer Simulation↗

PubMed1990

A comparison of different methods for calculating overall risk scores from risk factors ascertained in a computerized obstetric information system.

Current obstetric risk-scoring systems depend upon simple addition of weighted or unweighted risk factors. In this study a population of 994 pregnant women (470 primiparae; 524 multiparae) has been examined in order to compare the use of weighted and unweighted risk scores, together with Bayes theorem, in estimating an overall risk score from the presence of individual risk factors. The findings have been related to fetal outcome in these pregnancies and expressed as receiver-operating characteristic (ROC) curves for different cutoff levels between normal and abnormal. In primiparae, the use of weighted risk factors, with or without Bayes theorem, was clearly superior to the use of unweighted factors. In multiparae there was only a marginal difference between the three approaches. The increasing use of computerised information system in antenatal care should make calculation of risk scores for every pregnant woman a practical option.

Adolescent↗

PubMed2003

Diagnostic value of symptoms in patients with suspected upper digestive tract disease.

OBJECTIVE: to develop a predictive model based on clinical data for patients with suspected upper digestive tract disease. PATIENTS AND METHODS: forty-seven clinical data were obtained before endoscopy from 283 patients with suspected upper digestive disease (153 men; mean age 55 years, limits 17-92 years). A clinical prediction of diagnosis was made before endoscopy. On the basis of the endoscopic findings, the patients were divided into three diagnostic groups: absence of significant disease (group I), significant benign disease (group II) and malignant disease (group III). The probability rate of belonging to each one of the three groups was obtained for each patient by using Bayes' theorem. RESULTS: the endoscopic findings were classified according to their clinical importance: 121 patients (43%) belonged to group I, 137 (48%) were included in group II; and 25 (9%) in group III. The clinical prediction correctly classified 61% of the patients (group I: 56%, group II: 62% and group III: 76%) The coincidence between prediction obtained using Bayes' theorem and the actual diagnosis was 61% (group I: 65%, group II: 51%, group III: 92%). The predictive model was useful to confirm or not the clinical prediction. CONCLUSION: objective analysis of clinical data can be useful to support clinical judgment, mainly in patients with neoplasia. However, the model is not adequate to improve indication of upper endoscopy since many patients are misclassified.

Adolescent↗

PubMed1996

Microsatellite typing in a paternity case against a deceased man whose two brothers were available for testing.

To resolve a deficiency case of disputed paternity, we examined the mother, child, and two sibs of the deceased alleged father. In all of the 11 conventional marker systems and three microsatellite systems (ACTBP2, FGA, and D11S488) applied, the child and the paternal sibs had alleles is common. We deduced first the possible ACTBP2 genotypes of the deceased parents of the two sibs from their test results. Next, from all the combinations of the deduced genotypes, we estimated the frequencies of the possible genotypes of the alleged father and, with Bayes' theorem, calculated the probabilities of the alleged father having the genotypes compatible with paternity. Using these probabilities, we worked out the probability of the child's genotype resulting from the mating of the mother with the alleged father. Meanwhile, we calculated the probability of the child's genotype being expected of the mating of the mother with an unrelated random man. Finally, applying Bayes' theorem again, we obtained a probability of paternity of 0.992 for the ACTBP2 locus alone. Inclusion of the FGA locus in the probability calculations brought the overall probability of paternity to 0.998. The present study demonstrates that testing for hypervariable microsatellite loci greatly facilitates the solution of a deficiency case that would be difficult or time-consuming to solve by conventional marker typing.

Female↗

PubMed1996

Adjusting the risk for trisomy 21 by a simple ultrasound method using fetal long-bone biometry.

OBJECTIVE: To establish the efficacy of second-trimester fetal long-bone biometry (femur, humerus, tibia, and fibula length) in detecting trisomy 21 and to generate tables for adjusting the risk of trisomy 21 according to long-bone biometry. METHODS: Four long-bones--femur, humerus, tibia, and fibula--were measured ultrasonically in singleton fetuses before genetic amniocentesis. Fetuses with normal karyotypes were used to derive regression equations describing predicted lengths on the basis of the biparietal diameter measurement. The efficacy of each abnormally short bone, alone and in combination, was determined in 22 fetuses with trisomy 21 encountered during the study period. After the sensitivity and specificity of long-bone biometry were established, appropriate tables were generated by Bayes' theorem to adjust the risk of trisomy 21 in the second trimester depending on long-bone biometry. RESULTS: Of 515 patients between 14 and 23 weeks' gestation, 493 had normal fetal karyotypes and 22 had trisomy 21. The sensitivity of an abnormal ultrasound, as defined by the presence of one or more short bones, was 63.6% and the specificity was 78.5%. According to Bayes' theorem, genetic amniocentesis may not be recommended for women less than 40 years old in the presence of normal long-bone biometry (ie, all four bones normal). CONCLUSION: Second-trimester fetal long-bone biometry is useful in detecting trisomy 21 and may be used to adjust the a priori risk of both high- and low-risk women for trisomy 21 and, therefore, the need for genetic amniocentesis.

Anthropometry↗

PubMed1983

Problems and pitfalls of risk assessment in antenatal care.

Current obstetric risk-scoring systems do not make a precise prediction of the chances of an abnormal outcome and so cannot be used in formal decision analysis. We examine here the feasibility of using Bayes' theorem to provide an accurate assessment of fetal risk and conclude that two severe limitations effectively exclude this approach as a useful contribution to antenatal care: (1) inaccuracy of the data base: geographical variations and the 'treatment paradox' conspire to reduce the reliability on which any assessment may be made; (2) 'dependency' of the risk factors: most obstetric variables are interdependent and are not therefore amenable to analysis by means of Bayes' theorem. Although fetal risk will be assessed subjectively for the foreseeable future, obstetricians should be aware of the essentially mathematical nature of decision-making.

Computers↗

PubMed1986

Factors associated with nephrotoxicity and clinical outcome in patients receiving amikacin.

Data from 60 patients treated with amikacin were analyzed for factors associated with nephrotoxicity. In 42 of these patients, data were examined for factors associated with clinical outcome. Variables evaluated included patient weight, age, sex, serum creatinine level, creatinine clearance, duration of therapy, total dose, mean daily dose, organism minimum inhibitory concentration (MIC), mean peak levels, mean trough levels, mean area under the serum concentration-time curve (AUC), total AUC, mean AUC greater than MIC, total AUC greater than MIC, mean Schumacher's intensity factor (IF), total IF, In (mean maximum concentration [Cmax]/MIC). Model-dependent pharmacokinetic parameters were calculated by computer based on a one-compartment model. When the parameters were examined individually, duration of therapy and total AUC correlated significantly (P less than .05) with nephrotoxicity. In contrast, a stepwise discriminant function analysis identified only duration of therapy (P less than .001) as an important factor. Based on this model and on Bayes' theorem, the predictive accuracy of identifying "nephrotoxic" patients increased from 0.17 to 0.39. When examined individually, mean IF, MIC, total dose, mean daily dose, and ln (mean Cmax/MIC) correlated significantly (P less than .05) with cure. In contrast, a simultaneous multivariable analysis identified IF, MIC, and total dose according to one model and ln (mean Cmax/MIC) according to a second statistical model of parameters selected to have the greatest prospective value. Based on Bayes' theorem and the first model, the predictive accuracy of identifying patients not cured increased from 0.19 to 0.83. For the second model, the predictive accuracy increased from 0.19 to 0.50.(ABSTRACT TRUNCATED AT 250 WORDS)

Adult↗

PubMed1990

Diagnostic accuracy of radionuclide techniques in patients with equivocal electrocardiographic exercise testing.

The aim of the present study was to evaluate the yield of radionuclide studies for the diagnosis of coronary artery disease (coronary artery narrowing greater than 50%) in a prospective series of 73 patients with thoracic pain and equivocal electrocardiographic stress testing. In the study population, the prevalence of coronary artery disease was 51%. The sensitivity, specificity, diagnostic accuracy and the post-test probability difference curves according to the Bayes' theorem were calculated for 201-thallium exercise testing and radionuclide exercise equilibrium ventriculography. For the latter study the following criteria were considered: (1) increase in left ventricular ejection fraction less than 5%; (2) the criterion proposed by Rozanski; (3) decrease in regional ejection fraction; and (4) abnormalities in phase and amplitude analysis (Fourier). 201-thallium exercise testing was the most sensitive (97%) and accurate (86%) study. Radionuclide ventriculography sensitivity was always lower for any criterion, although its best result was for evaluation of regional ejection fraction (85%). The most specific study was Fourier analysis (97%), although its sensitivity was low (42%). The application of Bayes' theorem to these results shows that the highest post-test probability difference values were achieved with 201-thallium exercise testing for prevalences higher than 40% and with Fourier analysis for lower prevalences.

Adult↗

PubMed1983

Computer-assisted diagnosis in the noninvasive evaluation of patients with suspected coronary artery disease.

A microcomputer program called CADENZA, which employs Bayes' theorem to analyze and report the results of various clinical descriptors and noninvasive tests relative to the diagnosis of coronary artery disease, was evaluated in 1,097 consecutive patients without previous myocardial infarction. With this program, each patient was characterized by a probability for coronary artery disease, based on Framingham risk factor analysis, symptom characterization, electrocardiographic stress testing, cardiokymography, cardiac fluoroscopy, thallium perfusion scintigraphy and technetium equilibrium-gated blood pool scintigraphy. A total of 11,808 probability estimates derived from various combinations of the available observations were analyzed: 2,180 in 170 patients undergoing coronary angiography and 9,628 in 969 patients who completed a 1 year follow-up for coronary events. The predicted probability of disease correlated linearly with observed angiographic prevalence in the 170 patients who subsequently had coronary angiography (prevalence = [0.001 +/- 0.011] + [0.966 +/- 0.019] X probability). The difference between probability and prevalence averaged 3.1%, and the magnitude of this correlation was not affected by the type or amount of data analyzed. The prevalence of multivessel disease in these patients increased as a monotonic function of disease probability. Below a probability of 25%, single vessel disease was slightly more common than multivessel disease. Above a probability of 75%, multivessel disease predominated. In the 969 patients followed up for 1 year from the date of testing, the incidence of cardiac death and nonfatal infarction increased as a cubic function of disease probability (from approximately 0 to 8% per year for each). Above a probability of 90%, however, the standard deviation for predicting these events was wide. These data indicate that Bayes' theorem in general--and CADENZA in particular--is an accurate, clinically applicable means for quantifying the prevalence of angiographic coronary artery disease, the risk of multivessel disease and the incidence of morbid coronary events in the year after testing.

Coronary Angiography↗

PubMed1972