Home   About Us   Divisions   Ask a Med-U-Nurse   Health Library   Find a Doctor   For Physicians   Job Opportunities   Make a Difference   News & Events    News Room    Kids Connection Newsletter    Support Groups    Special Events    Parenting Newsletter    Parenting Classes    Community Classes    Grand Rounds Schedule   Research   Residents & Fellows   Your Hospital Visit 171 Ashley Ave. Charleston, SC 29425 843-792-1414 800-424-MUSC		Print Version February 2007 Evidence-Based Tip Diagnosis: How good is this test? Usually, when we use the word "diagnosis," we are referring to the process of making a diagnosis or performing a differential diagnosis. In Evidence-Based Medicine, however, when we pose a diagnosis question, we want to know how well a particular test does in telling us if our patient does or does not have a certain condition. Diagnostic studies give us information that tells us how well the test does, compared to "the truth," as represented by the "gold standard." Bandolier's EBM Glossary defines gold standard as "A method, procedure or measurement that is widely accepted as being the best available." For example, a recent study at Vanderbilt University Medical Center compared the rapid flu test (influenza rapid antigen detection) to culture and reverse-transcription polymerase chain reaction. Two important percentages result from this comparison: sensitivity and specificity. In this study, the sensitivity tells us how many of the patients who had a positive culture (the gold standard) tested positive on the rapid flu test. Out of 41 children detected by the gold standard, 26 tested positive on the rapid test. Dividing 26 by 41, we get a sensitivity of 63 percent. The specificity tells us how many of the patients who had a negative culture (the gold standard) had a negative result on the rapid flu test. Out of the 229 negatives from the gold standard, 223 children had a negative rapid test. Dividing 223 by 229, we get a specificity of 97 percent. Sensitivity tells us how many people with the disease will have a positive diagnostic test, and specificity tells us how many people without the disease will have a negative diagnostic test. So these numbers can give us a feel for how good a test is. However, they do NOT give us the likelihood that a patient does or does not have the disease. For the rapid flu test, even though this test has a high specificity, its low sensitivity can decrease its usefulness in ruling in the disease. To make these numbers more useful, we need to do a little more math. We need to calculate the likelihood ratio. Next month: The Likelihood Ratio: what are the odds that my patient has the disease I am testing for?
	Terms and Conditions | Privacy Policy | Site Map | Contact Us