As an illustration of how effective a test is depending upon its Sensitivity and Specificity , the following hypothetical test must find the 10 abnormal dogs among 100 normal dogs. Please note, in real life the ratio of normal to abnormal is typically much higher.

The purpose of a biomarker for disease is to properly identify the abnormal disease state, or in essence "find the needle in the haystack". If a test were perfect, it would separate out all the abnormal dogs from the normal dogs (ie, free of the disease the test is looking for) for its intended use. 

If a test has high specificity, when positive you can have confidence the disease state is present. With moderate sensitivity, some of the abnormal dogs will be incorrectly identified as normal. 

If a test has high sensitivity but low specificity, the test will include most of the abnormal dogs when positive, but it will also include many normal dogs or have a high False Positive Rate . In this situation, with a positive test, the veterinarian will need to further work up and treat dogs that are otherwise free of the disease the test is looking for. As such, valuable time and money is lost pursing one disease state when the dog may have another disease for which treatment may not be given.

An effectively worthless test is one that has a low sensitivity and specificity, which is the same as if you were just flipping a coin.