China
Africa
Explore chapters and articles related to this topic
Meta-Analysis of Diagnostic Tests
Published in Christopher H. Schmid, Theo Stijnen, Ian R. White, Handbook of Meta-Analysis, 2020
Yulun Liu, Xiaoye Ma, Yong Chen, Theo Stijnen, Haitao Chu
with Se on the y-axis and (1 − Sp) on the x-axis. As shown in equation (19.1), the corresponding sensitivity can be derived to any given value of specificity, thereby giving the entire SROC curve. When comparing the performance of multiple diagnostic tests, the SROC curve may be useful to describe the trade-off of sensitivity against specificity (Bossuyt et al., 2013). For example, given the same value of specificity, the corresponding sensitivities of each diagnostic test can be calculated, and their differences can be reported at specified values of the specificity. In addition, summary values for the diagnostic odds ratio can be computed and their corresponding confidence intervals as well.
D
Published in Filomena Pereira-Maxwell, Medical Statistics, 2018
A clinical, laboratory, radiological or other type of test, which is carried out for the purpose of establishing an actual diagnosis as to the presence or absence of disease. Unlike screening, diagnostic testing is usually prompted by the presence of signs and/or symptoms of disease. Test results are often dichotomized according to the categories of a binary variable as, for example, positive/negative, above threshold/below threshold. For this type of test, diagnostic accuracy is measured by its sensitivity and specificity, but its usefulness in practice is given by its predictive ability, given the actual prevalence of disease. A measure of overall accuracy is the diagnostic odds ratio. For test results on an ordinal or quantitative scale, the choice of diagnostic threshold may be made by plotting sensitivity (detection rate) vs. false-positive rate for different cut-off points - what is known as a ROC curve. However, a different utilization of these test results is also possible, given by the likelihood of different test results conditional on the presence (or absence) of disease, that results in the calculation of post-test odds and post-test probabilities of disease for different test results. Errors or misclassification with respect to disease status may lead to underestimation of exposure and treatment effects. See also classification table, false-positive rate, false-negative rate, predictive values, likelihood ratios. MACHIN & CAMPBELL (2005) give details of the design of studies to establish diagnostic accuracy. See Guyatt, Sackett & Haynes, in HAYNES et al. (eds., 2006), for further details and discussion, with a focus on the evaluation of diagnostic tests. Deeks, in EGGER, DAVEY SMITH & ALTMAN (eds., 2001), discusses the undertaking of systematic reviews of studies evaluating diagnostic and screening tests, to which special quality and bias assessment criteria apply, and which require specific methodology. A comprehensive overview of statistical methods for the evaluation of diagnostic and screening tests is also given. See STARD statement for reporting guidelines for studies of diagnostic accuracy.
Commentary: statistical analysis of 2 × 2 tables in biomarker studies 2) study design and statistical tests
Published in Biomarkers, 2022
The diagnostic odds ratio (DOR) is a single measure of diagnostic accuracy similar to the odds ratio widely used in epidemiology. It is the ratio of the odds of + ve results in cases with the condition relative to the odds of + ve results in cases without the condition (Glas et al. 2003). It is calculated as DOR = (TP/FN)/(FP/TN) or ad/bc. Alternative formulations use the sensitivity and the specificity or the PPV and NPV while it is also equal to ratio of the positive and negative likelihood ratios (LR+/LR–); illustrating, again, the inter-relationship of the statistics associated with the 2 × 2 table. A ratio of 1 provides no diagnostic evidence. The lower value is 0 while the upper limit is infinity. The value rises steeply when sensitivity or specificity are close to 1. Confidence intervals can be derived, although the upper limits may be unbounded. It does not depend on prevalence but does on the criteria used for defining the binary (±) ‘end points’ (Šimundić 2009). It can be used for meta-analyses across studies. Although Glas et al. (2003) recommend it, discussing its use with logistic regression, Sackett et al. (1996) and Pepe et al. (2004) cautioned against it, partly because the same DOR value can be obtained with different combinations of sensitivity and specificity. It is important to appreciate that, although large odds ratio such as 3.0 indicate strong associations in epidemiological studies, they do not necessarily provide evidence of good classification in diagnostic studies.
The application of advanced imaging techniques in glaucoma
Published in Expert Review of Ophthalmology, 2022
Su Ling Young, Nikhil Jain, Andrew J Tatham
The ALIENOR study was another single-gate design study, which used a population-based approach [22]. Schweitzer and colleagues examined the diagnostic ability of SDOCT peripapillary retinal nerve fiber layer (cpRNFL) parameters in 532 individuals over 65 years old. The diagnostic odds ratio (DOR), which is the ratio of the odds of positivity in disease relative to the odds of positivity in non-disease, was used as a key parameter for test comparison. Higher values of DOR indicate better discriminating performance. Superotemporal cpRNFL thickness was the best parameter for detecting glaucoma, with a DOR of 25.31 (positive likelihood ratio (LR+), 4.65; negative likelihood ratio (LR-), 0.18), followed by average cpRNFL thickness (DOR = 24.80, LR+, 6.36; LR−, 0.26). Applying the normative database provided by the machine to all cpRNFL parameters outside normal limits, with at least one abnormal RNFL parameter (at P < 0.05) the DOR increased further to 31.03 (LR+, 1.75; LR−, 0.06), performing better than superotemporal cpRNFL thickness alone [22].
Comparison of clinical usefulness of serum Ca125 and CA19-9 in pancreatic adenocarcinoma diagnosis: meta-analysis and systematic review of literature
Published in Biomarkers, 2021
Aleksander Skulimowski, Adam Durczyński, Janusz Strzelczyk, Piotr Hogendorf
The point estimate for CA19-9 has the following parameters:Sensitivity: 0.748 (95%CI: 0.676–0.809).Specificity: 0.782 (95%CI: 0.716–0.836).Area under the curve (AUC) was estimated for 0.832.Using the calculated hsROC, we applied it to further calculate the mean DOR, PLR and NLR.Diagnostic odds ratio: 10.9 (7.56–15.1).Positive likelihood ratio: 3.46 (2.72–4.4).Negative likelihood ratio: 0.324 (0.252–0.403).