Clinical Test for Celiac Disease
The sensitivity of a clinical test designed to discriminate between individuals with or without celiac disease would be based on its ability to correctly identify patients with the disease (Lalkhen & McClusky, 2008). In the example provided, 200 patients known to have celiac disease and 200 controls known not to have the disease were used to test the efficacy of the new clinical test. The results are as follows:
True positives = 190/200 = 0.95 or 95%
False positives = 50/200 = 0.25 or 25%
True negatives = 150/200 = 0.75 or 75%
False negatives = 10/200 = 0.05 or 5%
Sensitivity is determined by dividing true positives by the sum of true positives and false negatives (Lalkhen & McClusky, 2008), or in this example 190/(190+10) = 0.95 or 95%.
The specificity of a clinical test describes its ability to exclude individuals from false positive results (Lalkhen & McClusky, 2008). Specificity is calculated by dividing true negatives by the sum of true negatives and false positives, or in this example 150/(150+50) = 0.75 or 75%. The sensitivity and specificity for the new celiac disease test is therefore 95% and 75%, respectively, which would be considered highly sensitive and moderately specific.
As Lalkhen and McClusky (2008) discuss, high sensitivity is important for serious diseases for which treatments are available, but if false positives puts patients at risk for harm from treatments, then a second test, if available, with low sensitivity and high specificity, would be used to cull false positives from the patient population. Celiac disease would not fall within this category, because most patients can live a full life after eliminating gluten from their diets (Celiac Disease Foundation, n.d.). A diagnostic test with moderate sensitivity and specificity would be sufficient, while false positives could be eliminated by lack of responsiveness to a gluten-free diet.
The positive predictive value (PPV) of a clinical test is useful because it provides a numerical value for the likelihood that someone with the disease will test positive (Lalkhen & McClusky, 2008). PPV is calculated by dividing true positives by the sum of true positives and false positives, or in the example above: 190/(190+50) = 0.792 or 79.2%. This is a good PPV because nearly 80% of individuals with the disease will test positive.
Applying the above sensitivity and specificity to a population of 10,000 individuals and disease prevalence…