Cow Tales: What’s in a diagnostic test | TSLN.com

Cow Tales: What’s in a diagnostic test

Kenny Barrett Jr., DVM, MS

Veterinarians always seem eager to sample some sort of biological stuff. From stomach foam to fetal membranes, it seems we can’t wait to test any number of different things. We need tests. We have bubble sheets and multiple choice forms permanently etched in our grey matter from eight years in the ivory tower. We enjoy tests. They help us further hone our skills in the “practice” of medicine. After all, we are only 51 percent sure most of the time. No wonder it takes eight years.

All joking aside, diagnostic tests are a real asset to veterinary medicine. There are tests to identify the nasty causes of pneumonia, measure trace mineral status, or even identify a potential STD. However, every test is subject to error, just ask a feedlot manager about the pen of “guaranteed open” heifers.

Diagnostic test falsities can occur at any point from sample collection to interpretation. Samples need to be collected with careful attention to detail. Veterinarians need to sample the right location, obtain an adequate amount of sample, and avoid contamination. Samples are then packaged, stored, and shipped to an appropriate lab in a timely fashion. Frozen packages can arrive on a Saturday in July only to thaw before being received. Conversely, samples can also be inadvertently frozen during shipment. Many people handle samples at the laboratory, with each step providing an additional opportunity for error to occur. No test is perfect, but all tests can provide valuable information.

Sensitivity and specificity are calculated measures for each particular test. They help us determine how valuable the test is. “Sensitivity” describes how likely a test will accurately detect a true positive. For example, how likely is a producer to accurately identify all the sick calves in a given weaning pen? In this example the producer’s observation is the test we are evaluating. Some ranchers have better honed observation skills and their sensitivity is greater.

“Specificity” refers to a test’s ability to accurately identify a true negative. I like to think of specificity as the size of the loop. Poor ropers tend to throw a bigger loop. A big loop helps ensure we catch the one we are after but we are also more likely to catch another animal. A big loop is not nearly as specific. Similarly, step back and take aim at a target with a 12-gauge and then a 22. We are more likely to hit the target with a shotgun. However, the dispersed shot pattern of a shotgun also means we will hit a lot of other things as well. The rifle is much more specific.

Every diagnostic test is subject to sensitivity and specificity. Of course we would like all of our tests to be perfect with 100 percent sensitivity and 100 percent specificity. Unfortunately these ideal tests do not exist. All tests are subject to some error although many tests near 100 percent for both measures. Usually, sensitivity and specificity work against each other. The more sensitive we make a test the more likely we are to identify a false positive; decrease specificity.

Recommended Stories For You

Back to the weaning pen example, we often need to determine whether we should just treat the sick calves, treat the whole pen to prevent healthy calves from becoming sick, and/or treat with booster vaccinations during a respiratory break in a group of calves. Veterinarians have their own breakpoints for moving through the decision tree. However, if we treat the whole pen we have created a scenario with 100 percent sensitivity and 0 percent specificity. Unfortunately, there are too many times when this is the right thing to do. Once we have evaluated a particular test we can apply the results to a test population – a herd of cows, a bull battery, or feedlot pen.

We can use the sensitivity and specificity of a test to calculate its positive and negative predictive value. In short, what is the likelihood a positive test is truly positive and a negative is truly negative, respectively. These values are calculated using sensitivity and specificity but depend largely on the prevalence of the condition being tested.

“Prevalence” is the frequency a positive test result is likely to occur in a given population. That population could be a small group of cattle or the national cowherd. Consider BVD-persistent infection (PI). We would not typically expect to find and BVD-PI calf in most well-vaccinated herds. On the other hand, PIs are more likely in a pen of feedlot calves purchased from three salebarns with 20 owners. These calves are high risk for a reason. The higher the prevalence of a disease or condition, the greater the positive predictive value and lower the negative predictive value. In other words, a positive PI ear notch is more likely to truly be a positive if it came from the feedlot pen with a greater expected prevalence.

Another example would be trichomoniasis, usually referred to as “trich,” a venereal disease of cattle. There is a low prevalence of trich in the U.S. with small pockets of increased disease rate. A positive trich test has a low positive predictive value while a negative test has a much greater negative predictive value. Tests with a low positive predictive value are not without value. Oftentimes these types of tests are used as screening tests because they are less expensive. A more costly, but more specific test can then be used to identify the true positives from those testing positive with the first test.

All in all, diagnostic tests are extremely important in veterinary medicine. Always work with your veterinarian when testing aspects of your ranching operation. Don’t be afraid to ask what the results mean and what the implications are. Discuss what you can test for, when tests are indicated, and what outcomes you should expect.

Veterinarians always seem eager to sample some sort of biological stuff. From stomach foam to fetal membranes, it seems we can’t wait to test any number of different things. We need tests. We have bubble sheets and multiple choice forms permanently etched in our grey matter from eight years in the ivory tower. We enjoy tests. They help us further hone our skills in the “practice” of medicine. After all, we are only 51 percent sure most of the time. No wonder it takes eight years.

All joking aside, diagnostic tests are a real asset to veterinary medicine. There are tests to identify the nasty causes of pneumonia, measure trace mineral status, or even identify a potential STD. However, every test is subject to error, just ask a feedlot manager about the pen of “guaranteed open” heifers.

Diagnostic test falsities can occur at any point from sample collection to interpretation. Samples need to be collected with careful attention to detail. Veterinarians need to sample the right location, obtain an adequate amount of sample, and avoid contamination. Samples are then packaged, stored, and shipped to an appropriate lab in a timely fashion. Frozen packages can arrive on a Saturday in July only to thaw before being received. Conversely, samples can also be inadvertently frozen during shipment. Many people handle samples at the laboratory, with each step providing an additional opportunity for error to occur. No test is perfect, but all tests can provide valuable information.

Sensitivity and specificity are calculated measures for each particular test. They help us determine how valuable the test is. “Sensitivity” describes how likely a test will accurately detect a true positive. For example, how likely is a producer to accurately identify all the sick calves in a given weaning pen? In this example the producer’s observation is the test we are evaluating. Some ranchers have better honed observation skills and their sensitivity is greater.

“Specificity” refers to a test’s ability to accurately identify a true negative. I like to think of specificity as the size of the loop. Poor ropers tend to throw a bigger loop. A big loop helps ensure we catch the one we are after but we are also more likely to catch another animal. A big loop is not nearly as specific. Similarly, step back and take aim at a target with a 12-gauge and then a 22. We are more likely to hit the target with a shotgun. However, the dispersed shot pattern of a shotgun also means we will hit a lot of other things as well. The rifle is much more specific.

Every diagnostic test is subject to sensitivity and specificity. Of course we would like all of our tests to be perfect with 100 percent sensitivity and 100 percent specificity. Unfortunately these ideal tests do not exist. All tests are subject to some error although many tests near 100 percent for both measures. Usually, sensitivity and specificity work against each other. The more sensitive we make a test the more likely we are to identify a false positive; decrease specificity.

Back to the weaning pen example, we often need to determine whether we should just treat the sick calves, treat the whole pen to prevent healthy calves from becoming sick, and/or treat with booster vaccinations during a respiratory break in a group of calves. Veterinarians have their own breakpoints for moving through the decision tree. However, if we treat the whole pen we have created a scenario with 100 percent sensitivity and 0 percent specificity. Unfortunately, there are too many times when this is the right thing to do. Once we have evaluated a particular test we can apply the results to a test population – a herd of cows, a bull battery, or feedlot pen.

We can use the sensitivity and specificity of a test to calculate its positive and negative predictive value. In short, what is the likelihood a positive test is truly positive and a negative is truly negative, respectively. These values are calculated using sensitivity and specificity but depend largely on the prevalence of the condition being tested.

“Prevalence” is the frequency a positive test result is likely to occur in a given population. That population could be a small group of cattle or the national cowherd. Consider BVD-persistent infection (PI). We would not typically expect to find and BVD-PI calf in most well-vaccinated herds. On the other hand, PIs are more likely in a pen of feedlot calves purchased from three salebarns with 20 owners. These calves are high risk for a reason. The higher the prevalence of a disease or condition, the greater the positive predictive value and lower the negative predictive value. In other words, a positive PI ear notch is more likely to truly be a positive if it came from the feedlot pen with a greater expected prevalence.

Another example would be trichomoniasis, usually referred to as “trich,” a venereal disease of cattle. There is a low prevalence of trich in the U.S. with small pockets of increased disease rate. A positive trich test has a low positive predictive value while a negative test has a much greater negative predictive value. Tests with a low positive predictive value are not without value. Oftentimes these types of tests are used as screening tests because they are less expensive. A more costly, but more specific test can then be used to identify the true positives from those testing positive with the first test.

All in all, diagnostic tests are extremely important in veterinary medicine. Always work with your veterinarian when testing aspects of your ranching operation. Don’t be afraid to ask what the results mean and what the implications are. Discuss what you can test for, when tests are indicated, and what outcomes you should expect.

kenny barrett jr. is a veterinarian at the belle fourche veterinary clinic in belle fourche, sd and pens “cow tales” monthly. learn more about the clinic on the web at http://www.bfvetclinic.com, or drop them an e-mail at: office@bfvetclinic.com to suggest a topic for the next installment of “cow tales.”

Go back to article