Indeterminate Test Results: The articles were examined for two standards regarding the reporting and analysis of indeterminate results. The first standard required a statement regarding the existence and frequency of indeterminate results generated in the study. Buy proventil Click Here The second standard required that the study indicated whether indeterminate results were included or excluded in calculations of test accuracy.
Test Reproducibility: This standard depended on the nature of the diagnostic test under evaluation. For tests that depended on observer interpretation (eg, imaging studies), at least some of the test subjects needed to have been evaluated for a summary measure of observer variability. For tests performed without observer interpretation, a summary measure of instrument variability needed to be provided. Test Acceptability: New diagnostic tests may have high diagnostic accuracies but low clinical utility. This standard required that the article made a statement regarding the acceptability of the new test in terms of factors such as minimum detection levels and cross-reactivity for biochemical tests, degree of measurement errors, required personnel and equipment, cost, acceptability to patients, dose, and pharmacokinetics of any drugs or agents. Head-to-Head Comparisons: Some articles compared the relative diagnostic accuracies of two or more diagnostic tests or a new test compared with an established test relative to a reference standard. For such studies, articles were examined for fulfillment of two standards. The first standard required that all study patients or a random sample of study patients were subjected to all of the compared diagnostic tests. The second standard determined whether appropriate statistical methods were used to compare the diagnostic accuracies of the evaluated tests. Acceptable methods included the calculation of the area under the curve (AUC) generated by receiver operating characteristic (ROC) analysis for each test with appropriate statistical comparisons or calculation of confidence intervals for the summary measures of diagnostic accuracy. Hypothesis testing statistical measures that calculated a p value to determine if the compared diagnostic tests identified different patient subgroups were not considered suitable measures of the relative diagnostic accuracies of the evaluated tests.