A study compares the detection of colon polyps in 50 subjects by using both CT colonography and conventional colonoscopy (gold standard). As a part of the study, 2 independent radiologists assess the CT images to diagnose polyps. Each radiologist is blinded to the other radiologist's reading. The kappa statistic for these raters is 0.73. Which of the following does the kappa statistic best describe?
A)Power
B)Reliability
C)Specificity
D)Statistical significance
E)Validity

Question

A study compares the detection of colon polyps in 50 subjects by using both CT colonography and conventional colonoscopy (gold standard).  As a part of the study, 2 independent radiologists assess the CT images to diagnose polyps.  Each radiologist is blinded to the other radiologist's reading.  The kappa statistic for these raters is 0.73.  Which of the following does the kappa statistic best describe?
A)Power
B)Reliability
C)Specificity
D)Statistical significance
E)Validity

Quizplus · Accepted Answer

Explanation: When the 2 radiologists review the CT images, their interpretations of the findings may agree or disagree due to chance alone. The kappa statistic is a quantitative measure of inter-rater reliability (sometimes referred to as inter-rater concordance). It reflects the extent to which inter-rater agreement represents an improvement on chance agreement alone. Kappa values range from -1 (perfect disagreement) to +1 (perfect agreement), with kappa = 0 suggesting agreement due to chance, kappa <0 suggesting less than chance agreement (possibly intentional disagreement), and kappa >0 suggesting greater than chance agreement. The following (arbitrary) cutoffs are often used for kappa: 0-0.20 represents negligible improvement over chance agreement, 0.21-0.40 is minimal, 0.41-0.60 is fair, 0.61-0.80 is good, and >0.80 is excellent. The other answer options relate to diagnostic test characteristics and statistical significance. (Choice A) The power of a diagnostic test is its ability to detect a difference when one actually exists. Power is given by (1−β), where β is the probability of making a type II error (failure to reject a null hypothesis when it is false). (Choice C) The specificity of a test is the proportion of results that are negative when no disease is present. In tests with high specificity, a positive result suggests that disease is actually present (few false positives). (Choice D) The level of statistical significance is denoted by α, which is usually set at 0.05. An α = 0.05 means that investigators are willing to accept a 5% chance that the null hypothesis is being rejected incorrectly (type I error). If the p-value of a measurement is <α, then the measurement is considered statistically significant and the null hypothesis is rejected. (Choice E) Tests with high validity (accuracy) provide correct results on average. Tests with high reliability provide consistent results (even though they may not be correct). Tests with high validity and reliability consistently provide correct results. validity and reliability Educational objective: The kappa statistic represents the extent to which inter-rater agreement is an improvement on chance agreement alone. It is a quantitative measure of inter-rater reliability. References: Understanding interobserver agreement: the kappa statistic. (http://www.ncbi.nlm.nih.gov/pubmed/15883903)

A Study Compares the Detection of Colon Polyps in 50