In Chapter 6, you learned that clinical research organizes research questions into a hierarchy. In addition, you learned that clinical trials may be described as occurring in phases, and these phases are applied to exercise science and rehabilitation research. You also were introduced to randomization, which is critical for strengthening the internal validity of a research study. Finally, you learned how to calculate numbers needed to treat and numbers needed to harm, which are two important clinical measures for determining the effectiveness of an intervention.
In this chapter you will learn to apply diagnostic statistics for accuracy of clinical tests. You will learn that sensitivity and specificity indicate the percentage of persons correctly identified as having and not having, respectively, a condition based on a clinical test result. In addition, positive and negative predictive values will be defined and you will learn how these outcomes tell you the percentage of persons with positive outcomes who actually have a condition and the percentage of persons with negative test results, who do not have a condition, respectively. These two predictive statistics are important for "ruling in" or "ruling out" the condition. Related to these accuracy statistics are the positive and negative likelihood ratios, which tell you how many more or fewer times likely a person is to have or not have a condition, respectively. These two statistics are powerful and critical for ruling in (positive likelihood ratio) and ruling out (negative likelihood ratio) a condition. An additional way to examine the accuracy of a clinical test is to plot a receiver operating characteristic curve. This curve plots the sensitivity of a test versus the false-positive rate (test identifies a person as having a condition, but the person really does not have it) to determine the accuracy of a test for identifying those persons with a given condition. Finally, a measure known as area under the curve quantifies the degree of accuracy for a test based on the receiver operating characteristic plot. A high area under the curve value is indicative of an accurate test.
After completing this chapter, you will be able to answer these questions:
What are the differences among sensitivity, specificity, positive predictive score, and negative predictive score?
What are the differences among scores and rates for true-positive, false-positive, true-negative, and false-negative outcomes?
What do positive and negative likelihood ratios tell you about the accuracy of a clinical test?
What is a receiver operating characteristic curve, and how is it used to quantify accuracy?
How do you determine a clinical meaningful threshold (i.e., cutoff score) on a receiver operating characteristic curve that identifies persons with a given condition?
How does an area under the curve value relate to the accuracy of a clinical test?