by admin on February 2, 2012
The next step in the evaluation of the reliability of diagnostic tests is to conduct systematic reviews on a group of related studies in order to understand the reliability of a test applied in different contexts, by different examiners, and on different patient samples. The preparation of systematic reviews requires the use of an accepted quality appraisal tool, which reviewers use to assess the methodological quality of included studies.
Quality appraisal tools have been developed for use in systematic reviews of diagnostic accuracy, with QUDAS being the recommended tool of choice. For systematic reviews of diagnostic reliability, QAREL is recommended.
If you enjoyed this post, make sure you subscribe to my RSS feed!
by admin on February 2, 2012
In order to determine if a test is reliable, the test must be subjected to a study of its reliability. Studies of diagnostic reliability are based on the comparison between the results of two or more independent examiners on a group of patients, or between two or more results obtained by the same examiner on the same group of patients. After the examination is concluded, the data from each examiner are used to estimate reliability.
Reliability is sometimes presumed to be a simple calculation of the number of times examination findings agree on a given set of test results. This approach, however, does not take into account the agreement that would occur between examination findings by chance alone. Since we are not interested in chance agreement, but in agreement beyond that of chance, it is necessary to use statistical tests that control for chance.
The type of statistical test used depends on the type of data generated by the diagnostic test. For continuous data, such as blood pressure or joint range of motion, the intraclass correlation coefficient (ICC) is used. For diagnostic tests that generate categorical data, such as ‘positive’ or ‘negative’, or “mild, “moderate”, or “severe”, the kappa statistic is used.
Many diagnostic tests that generate continuous data are converted into categorical data for ease of reporting. For example, a discrepancy in leg length is a continuous variable that is measured in millimetres, however leg length discrepancy is often rated as being either ‘absent’ or ‘present’. Similarly, a radial fissure in a lumbar intervertebral disc is a continuous variable, whereas it is the categorical grade of the fissure (grade 1, 2, 3 or 4) that has been shown to be associated with low back pain used for diagnostic purposes.
If you enjoyed this post, make sure you subscribe to my RSS feed!
