# How to Calculate and Use Predictive Values in Veterinary Practice ## Quick Answer - Positive predictive value (PPV) tells you the chance that a positive test result means the animal truly has the disease, and it rises when disease prevalence in the tested population is higher. - Negative predictive value (NPV) tells you the chance that a negative test result means the animal is truly free of the disease, and it falls when prevalence rises. - Both values depend heavily on the population being tested, so a test with excellent performance in a referral hospital may perform poorly in a general practice screening setting. ## At a Glance Predictive values answer a different question than sensitivity and specificity. Sensitivity and specificity describe how well a test performs in animals known to have or not have a disease. Predictive values describe how much a clinician should trust a test result in an individual patient. The same test can have very different predictive values depending on how common the disease is in the population being tested. | Test characteristic | What it measures | Why it matters in practice | | --- | --- | --- | | Sensitivity | Proportion of diseased animals that test positive | A highly sensitive test rarely misses disease, useful for ruling out | | Specificity | Proportion of healthy animals that test negative | A highly specific test rarely gives false positives, useful for confirming | | Positive predictive value | Proportion of positive test results that are true positives | Determines how much weight to place on a positive result in your patient population | | Negative predictive value | Proportion of negative test results that are true negatives | Determines how confident you can be that a negative result rules out disease | The relationship between these values is not fixed. When disease prevalence increases in a population, PPV increases and NPV decreases. When prevalence decreases, PPV decreases and NPV increases. This is why published sensitivity and specificity values from one study cannot be directly applied to a different population without considering prevalence. ## Understanding Predictive Values in Veterinary Diagnostics Veterinary clinicians order laboratory tests to answer a practical question about a specific animal. The test result must be interpreted in the context of that animal's signalment, history, physical examination findings, and the prevalence of the disease in the population the animal represents. Predictive values provide the mathematical framework for this interpretation. The American Veterinary Medical Association emphasizes that pet owners rely on veterinarians to guide them through preventive care and diagnostic decisions, and clear communication about what test results mean is part of that responsibility. When a clinician understands predictive values, they can explain to an owner why a positive test result does not always mean the pet has the disease, and why a negative result does not always mean the pet is healthy. The World Small Animal Veterinary Association publishes global guidelines for companion animal practice that stress the importance of evidence-based clinical decision making. Understanding the statistical properties of diagnostic tests is a core component of this approach. A test result is only useful when the clinician knows how to interpret it in the specific patient before them. ### The Difference Between Test Performance and Test Interpretation Sensitivity and specificity are properties of the test itself. They are determined by comparing test results against a gold standard in a group of animals with known disease status. These values do not change based on the population being tested, assuming the test is used in a similar manner and the disease manifests similarly. Predictive values are properties of the test applied to a specific population. They depend on the prevalence of the disease in that population. A test with 95% sensitivity and 95% specificity sounds excellent, but if it is used to screen a population where only 1% of animals have the disease, the PPV will be only about 16%. This means more than 80% of positive results will be false positives. This distinction matters for everyday clinical decisions. A clinician who understands predictive values will not order a test indiscriminately. They will consider the pretest probability of disease based on the animal's presentation and use the predictive values to determine how much the test result should change their clinical suspicion. ### Why Prevalence Drives Predictive Values Prevalence is the proportion of animals in a given population that have the disease at a particular time. In veterinary medicine, prevalence varies widely by geographic region, species, breed, age, management system, and referral status. Consider a herd of pigs with a known history of Actinobacillus pleuropneumoniae infection. The prevalence of lung colonization in that herd will be much higher than in a herd with no known exposure. A tonsillar culture test with a PPV of 94.6% in experimentally infected nursery pigs would have a much lower PPV in a herd where the true prevalence is low, because the proportion of false positives relative to true positives would increase. The same logic applies to companion animal medicine. A urine normetanephrine test for pheochromocytoma in dogs has a PPV of 83.3% in a referral population where 19 of 32 dogs had the disease. In a general practice population where adrenal tumors are less common, the PPV would be lower. The test has not changed, but the population has, and the interpretation must change accordingly. ## Building a 2x2 Table for Predictive Value Calculation The 2x2 table is the standard tool for organizing diagnostic test data. It allows the clinician to calculate sensitivity, specificity, PPV, and NPV from the same set of numbers. The table has four cells representing the combinations of test result and true disease status. | | Disease present | Disease absent | | --- | --- | --- | | Test positive | True positive (TP) | False positive (FP) | | Test negative | False negative (FN) | True negative (TN) | From these four numbers, the clinician can calculate all four test performance measures. Sensitivity is TP divided by all diseased animals (TP plus FN). Specificity is TN divided by all healthy animals (TN plus FP). PPV is TP divided by all positive test results (TP plus FP). NPV is TN divided by all negative test results (TN plus FN). ### Step-by-Step Calculation Method To calculate predictive values from a 2x2 table, follow these steps. First, identify the true disease status of each animal using the gold standard test or postmortem examination. This is the reference against which the test being evaluated is compared. Second, record the test result for each animal. Each animal falls into one of the four cells based on the combination of test result and true disease status. Third, calculate the prevalence of disease in the study population by dividing the total number of diseased animals (TP plus FN) by the total number of animals in the study. Fourth, calculate PPV by dividing the number of true positives by the total number of positive test results. Multiply by 100 to express as a percentage. Fifth, calculate NPV by dividing the number of true negatives by the total number of negative test results. Multiply by 100 to express as a percentage. ### Worked Example from Canine Portosystemic Shunt Diagnosis A study of ultrasonographic diagnosis of portosystemic shunting in dogs and cats provides a useful worked example. The study included 85 dogs and 17 cats with clinically suspected portosystemic shunt. A shunt was confirmed in 50 dogs and 9 cats. Six dogs and one cat had hepatic microvascular dysplasia, and 29 dogs and 7 cats had a normal portal system. Ultrasonography was 92% sensitive and 98% specific for identifying portosystemic shunt. The PPV was 98% and the NPV was 89%. The overall accuracy was 95%. To understand these numbers, consider what they mean in practice. In this referral population where portosystemic shunt was strongly suspected based on clinical signs, a positive ultrasound result was highly reliable. The 98% PPV means that 98 out of 100 positive ultrasound results were true positives. The 89% NPV means that 89 out of 100 negative ultrasound results were true negatives. The study also reported that the combination of a small liver, large kidneys, and uroliths had a PPV of 100% and an NPV of 51% for congenital portosystemic shunt in dogs. This means that when all three findings were present, the diagnosis was confirmed in every case. But the absence of this combination did not rule out the disease, since the NPV was only 51%. ### Worked Example from Porcine Respiratory Disease A study of tonsillar culture for detecting Actinobacillus pleuropneumoniae in nursery pigs provides a second worked example with different characteristics. In this study, 163 German Landrace nursery pigs were experimentally exposed to A. pleuropneumoniae serotype 7 by aerosol. The study found that 74.8% of pigs tested positive in both tonsillar and lung samples, 7.4% remained completely negative, 4.3% had positive tonsils but negative lung cultures, and 13.5% had positive lung cultures but negative tonsillar samples. The diagnostic sensitivity of tonsillar culture for detecting positive lung colonization was 84.7%, and the diagnostic specificity was 66.7%. The PPV was 94.6% and the NPV was 35.3%. The low NPV is particularly instructive. In this experimentally infected population where the true prevalence of lung colonization was high, a negative tonsillar culture did not reliably rule out lung colonization. More than one in three pigs with a negative tonsillar culture actually had A. pleuropneumoniae in their lung tissue. This is because the sensitivity of the test was not perfect, and the high prevalence meant that false negatives were relatively common. ## Prevalence and Its Effect on Predictive Values The relationship between prevalence and predictive values is the single most important concept for clinicians to understand. A test that performs well in one population can perform poorly in another solely because of differences in disease prevalence. ### How Prevalence Changes PPV and NPV When prevalence is high, most positive test results will be true positives because the pool of actually diseased animals is large. The PPV will be high. At the same time, the NPV will be lower because the pool of actually healthy animals is small, and even a small proportion of false negatives will represent a meaningful fraction of all negative results. When prevalence is low, the opposite occurs. Most positive test results will be false positives because the pool of actually healthy animals is large. The PPV will be low. The NPV will be high because the pool of actually healthy animals is large, and false negatives will be a small fraction of all negative results. This is why screening tests are problematic in low-prevalence populations. A test with 99% specificity will produce one false positive for every 100 healthy animals tested. If the disease prevalence is 1%, then for every 100 animals tested, there will be one true positive and approximately one false positive. The PPV will be about 50%, meaning a positive result is only as likely to be true as it is to be false. ### Applying Prevalence to Herd and Population Decisions In production animal medicine, prevalence considerations affect herd-level testing decisions. The World Organisation for Animal Health emphasizes the importance of surveillance and reporting for animal health and welfare. When designing a surveillance program, the expected prevalence of the target disease in the population determines whether a given test will be useful. For a disease with very low prevalence in a region, testing with a highly specific test will generate many false positives that require confirmatory testing. This increases the cost of the surveillance program and can create unnecessary concern among producers. Understanding the expected PPV allows the program designer to plan for confirmatory testing and to interpret positive results appropriately. In companion animal practice, prevalence considerations affect individual patient decisions. A test for a rare disease should only be ordered when there is a specific clinical reason to suspect the disease. Ordering a panel of tests for rare diseases in a healthy animal will generate false positives that cause owner anxiety and lead to unnecessary additional testing. ### The Referral Population Effect Referral populations have higher disease prevalence than general populations because animals are referred specifically because they are sick or because a disease is suspected. This means that published predictive values from referral hospital studies will overestimate the PPV that a general practitioner will see. The pheochromocytoma study illustrates this effect. The study included 32 client-owned dogs with adrenal tumors, of which 19 had pheochromocytoma and 13 had other adrenal tumors. This is a prevalence of 59.4% in the study population. The PPV of the urine normetanephrine test was 83.3% in this population. In general practice, the prevalence of pheochromocytoma among dogs with adrenal masses is lower, and the PPV of the same test would be correspondingly lower. A general practitioner who applies the published PPV to their own patient population will overestimate the reliability of a positive test result. ## Sensitivity and Specificity as Inputs to Predictive Values Sensitivity and specificity are the raw materials from which predictive values are derived. Understanding how they interact with prevalence helps the clinician predict how a test will perform before ordering it. ### High Sensitivity Tests and Ruling Out Disease A highly sensitive test rarely misses disease. When a highly sensitive test is negative, the clinician can be reasonably confident that the animal does not have the disease. This is expressed in the mnemonic SnNOut, which stands for sensitive test, negative result, rules out. The NPV of a highly sensitive test will be high, especially when prevalence is low. This is because the test rarely produces false negatives, and in a low-prevalence population, most negative results will be true negatives. The tonsillar culture study for A. pleuropneumoniae illustrates the limitation of this approach. The sensitivity was 84.7%, which is not high enough to reliably rule out lung colonization. The NPV was only 35.3% in the high-prevalence experimental population. A clinician who used a negative tonsillar culture to rule out lung colonization would be wrong more than 60% of the time. ### High Specificity Tests and Ruling In Disease A highly specific test rarely produces false positives. When a highly specific test is positive, the clinician can be reasonably confident that the animal has the disease. This is expressed in the mnemonic SpPIn, which stands for specific test, positive result, rules in. The PPV of a highly specific test will be high, especially when prevalence is high. This is because the test rarely produces false positives, and in a high-prevalence population, most positive results will be true positives. The portosystemic shunt study illustrates this principle. The specificity of ultrasonography was 98%, and the PPV was 98% in the referral population. A positive ultrasound result was highly reliable for confirming the diagnosis. ### The Tradeoff Between Sensitivity and Specificity Most tests involve a tradeoff between sensitivity and specificity. A test that is made more sensitive by lowering the threshold for a positive result will also become less specific, because more healthy animals will exceed the lower threshold. The pheochromocytoma study provides a concrete example. The study found that using a cutoff of 4 times the upper reference limit for urine normetanephrine would miss a significant number of diagnoses. Only 5 of 19 pheochromocytoma patients had urine normetanephrine concentrations above this cutoff. The study authors concluded that previously published guidelines using this cutoff would lead to missed diagnoses. The spot urine normetanephrine-to-creatinine ratio had a sensitivity of 78.9% and a specificity of 76.9% for discriminating between pheochromocytoma and nonpheochromocytoma. The PPV was 83.3% and the NPV was 71.4%. These moderate values reflect the inherent overlap between the two groups of dogs. ## Practical Workflow for Using Predictive Values in Clinical Decisions The following workflow helps the clinician apply predictive values to individual patient decisions. It can be adapted for companion animal practice, production animal medicine, and laboratory quality assurance. ### Step 1: Establish the Pretest Probability Before ordering a test, estimate the probability that the animal has the disease based on signalment, history, physical examination findings, and local disease prevalence. This is the pretest probability. For example, a young dog with a small liver, large kidneys, and uroliths has a high pretest probability of congenital portosystemic shunt. The study found that this combination of findings had a PPV of 100% for congenital portosystemic shunt in dogs. A dog with none of these findings has a lower pretest probability. For a pig herd with a known history of A. pleuropneumoniae, the pretest probability of lung colonization in a nursery pig is higher than in a herd with no known exposure. The clinician should adjust the interpretation of tonsillar culture results accordingly. ### Step 2: Select the Appropriate Test Choose a test with known sensitivity and specificity for the disease in question. Consider whether the test has been validated in the species and population being tested. A test validated in one species may not perform identically in another species. The World Small Animal Veterinary Association global guidelines provide information on diagnostic approaches for common companion animal conditions. The Merck Veterinary Manual provides authoritative background on disease diagnosis and prevention across species. These sources can help the clinician select appropriate tests. ### Step 3: Calculate or Look Up the Predictive Values If the prevalence of the disease in the relevant population is known, the clinician can calculate the expected PPV and NPV using the sensitivity and specificity of the test. If published predictive values are available from a study in a similar population, these can be used as estimates. The clinician should be cautious about applying published predictive values from a referral population to a general practice population. The prevalence will almost certainly be different, and the predictive values will change accordingly. ### Step 4: Interpret the Test Result in Context When the test result returns, interpret it using the predictive values that apply to the patient's population. A positive result in a high-prevalence population is more likely to be a true positive than the same result in a low-prevalence population. The clinician should also consider the consequences of a wrong answer. If a false negative would lead to a missed diagnosis with serious consequences, the clinician may want to confirm a negative result with a second test or repeat testing. If a false positive would lead to unnecessary treatment or surgery, the clinician may want to confirm a positive result before proceeding. ### Step 5: Communicate the Result to the Owner or Producer Explain the test result in terms the owner or producer can understand. The American Veterinary Medical Association provides resources for pet owners that emphasize the importance of clear communication between veterinarians and owners. Cornell University College of Veterinary Medicine also provides educational resources for animal owners. For a positive result with a PPV of 80%, explain that 8 out of 10 animals with this test result actually have the disease, and that confirmatory testing may be recommended. For a negative result with an NPV of 90%, explain that 9 out of 10 animals with this test result are truly free of the disease, but that clinical signs should still be monitored. ## Common Failure Patterns in Predictive Value Interpretation Clinicians make predictable errors when interpreting predictive values. Recognizing these patterns helps avoid them. ### Applying Referral Population Values to General Practice The most common error is applying published predictive values from a referral hospital study to a general practice population. The prevalence of disease in a referral population is almost always higher than in a general population, so the PPV will be lower in general practice than the published value suggests. The pheochromocytoma study illustrates this problem. The study population had a 59.4% prevalence of pheochromocytoma among dogs with adrenal tumors. A general practitioner seeing dogs with adrenal masses will encounter a lower prevalence, and the PPV of the urine normetanephrine test will be lower than 83.3%. ### Ignoring Prevalence When Interpreting Results A second common error is ignoring prevalence entirely and interpreting test results as if they were definitive. A positive result on a test with 95% specificity still has a substantial chance of being false if the disease is rare. Consider a hypothetical test for a disease with 1% prevalence in the population. If the test has 95% sensitivity and 95% specificity, the PPV is only 16.1%. This means that 84 of every 100 positive results are false positives. A clinician who treats every positive result as a diagnosis will cause substantial harm through unnecessary treatment. ### Using Sensitivity and Specificity Interchangeably with Predictive Values A third common error is confusing sensitivity and specificity with predictive values. Sensitivity and specificity describe the test, while predictive values describe the test result in a population. A test with 99% sensitivity does not have a 99% PPV. The PPV depends on prevalence. ### Overinterpreting Negative Results in High-Prevalence Populations A fourth common error is overinterpreting negative results in high-prevalence populations. The tonsillar culture study for A. pleuropneumoniae provides a clear example. The NPV was only 35.3% in the experimentally infected population. A negative tonsillar culture did not rule out lung colonization in most cases. This error is particularly dangerous in production animal medicine, where a negative test result might lead a producer to believe a herd is free of a disease when it is not. The World Organisation for Animal Health emphasizes the importance of accurate surveillance and reporting, and understanding the limitations of negative test results is part of this responsibility. ## Records and Measurements for Predictive Value Assessment Maintaining records of test results and outcomes allows the clinician to assess how well predictive values apply to their own population. This is a form of local validation that improves clinical decision making over time. ### What to Record For each diagnostic test ordered, record the following information. Record the signalment of the animal, including species, breed, age, and sex. Record the reason for testing, including the clinical signs that prompted the test. Record the test result, including the specific value if the test is quantitative. Record the final diagnosis, including the results of any confirmatory testing or postmortem examination. For production animal medicine, record the herd or group identification, the production stage, and the management system. Record the reason for testing, such as routine surveillance, suspected disease outbreak, or pre-movement testing. Record the test results and any subsequent disease events in the group. ### How to Use the Records Periodically review the records to calculate the actual PPV and NPV of tests in your population. Compare these values to published values from the literature. If the actual PPV is lower than the published value, the prevalence of disease in your population is likely lower than in the study population. This review also helps identify tests that are not performing as expected. If a test has a much lower PPV than expected, the test may be producing more false positives than anticipated, or the disease prevalence may be lower than assumed. Either way, the clinician should adjust their interpretation of future test results. ### Limitations of Local Records Local records have limitations. The number of animals tested may be small, so the calculated predictive values will have wide confidence intervals. The gold standard used to confirm disease may be imperfect, so some animals will be misclassified. The population may change over time, so historical predictive values may not apply to current patients. Despite these limitations, local records provide valuable information that complements published studies. The Cornell University College of Veterinary Medicine and other academic institutions emphasize the importance of evidence-based practice, and local data are part of the evidence base. ## Welfare and Safety Context for Predictive Value Use The interpretation of diagnostic tests has direct welfare and safety implications for animals and for the people who care for them. ### Avoiding Unnecessary Treatment A false positive test result can lead to unnecessary treatment. In companion animal practice, this might mean unnecessary surgery, such as an adrenalectomy for a dog that does not have a pheochromocytoma. The pheochromocytoma study found that 21% of cases were biochemically silent, meaning the tumor did not produce elevated hormone levels. This reinforces the importance of confirmatory testing before surgical intervention. In production animal medicine, a false positive result might lead to culling of animals that are not actually infected, or unnecessary treatment of a group with antimicrobials. The World Organisation for Animal Health emphasizes the importance of responsible antimicrobial use, and avoiding unnecessary treatment is part of this responsibility. ### Avoiding Missed Diagnoses A false negative test result can lead to a missed diagnosis. In companion animal practice, this might mean failing to diagnose a portosystemic shunt in a dog with clinical signs. The portosystemic shunt study found that the NPV of ultrasonography was 89%, meaning that 11% of negative results were false negatives. In production animal medicine, a false negative result might mean failing to detect A. pleuropneumoniae in a group of pigs, allowing the disease to spread. The tonsillar culture study found that the NPV was only 35.3% in the high-prevalence experimental population, meaning that most negative results were false negatives. ### Communicating Uncertainty to Owners and Producers Clinicians have an ethical obligation to communicate uncertainty honestly. The American Veterinary Medical Association emphasizes the importance of the veterinarian-client-patient relationship, and honest communication about test limitations is part of this relationship. When a test result is uncertain, explain the uncertainty to the owner or producer. Explain that the test is not perfect, that the result must be interpreted in the context of the animal's clinical signs, and that additional testing may be recommended. This honest approach builds trust and helps the owner or producer make informed decisions. ## Professional Escalation Criteria There are situations where the clinician should seek additional expertise or refer the case to a specialist. ### When to Refer for Specialist Evaluation Refer to a specialist when the test result is discordant with the clinical picture. If the test is positive but the animal has no clinical signs of the disease, or if the test is negative but the animal has strong clinical signs of the disease, the clinician should consider referral for further evaluation. Refer to a specialist when the disease is uncommon and the clinician has limited experience with its diagnosis and management. The Merck Veterinary Manual provides background information on many diseases, but specialist consultation may be appropriate for complex cases. Refer to a specialist when the consequences of a wrong diagnosis are severe. For example, a dog with a suspected pheochromocytoma should be referred to a surgeon with experience in adrenalectomy, because the surgery carries significant risk and the diagnosis must be secure before proceeding. ### When to Seek Laboratory Consultation Consult the diagnostic laboratory when there is uncertainty about test interpretation. Laboratory professionals can provide information about test performance, expected values in different populations, and factors that might affect test results. Consult the laboratory when a test result seems inconsistent with the clinical picture. The laboratory may be able to repeat the test, perform additional testing, or provide guidance on interpretation. ### When to Escalate in Production Animal Medicine In production animal medicine, escalate to a veterinary specialist or diagnostic laboratory when a disease outbreak is suspected or when test results have major economic consequences. The World Organisation for Animal Health provides guidance on disease surveillance and reporting, and some diseases are reportable to animal health authorities. If a test result suggests the presence of a reportable disease, the clinician should contact the appropriate animal health authority immediately. The World Organisation for Animal Health maintains standards for disease reporting and notification. ## Limitations of Predictive Values in Veterinary Practice Predictive values are powerful tools, but they have important limitations that the clinician must understand. ### Dependence on Accurate Prevalence Estimates Predictive values depend on accurate prevalence estimates, and prevalence estimates are often uncertain. The true prevalence of a disease in a given population may not be known, and the clinician must rely on estimates from the literature, local experience, or clinical judgment. When the prevalence estimate is uncertain, the predictive values are also uncertain. The clinician should consider a range of possible prevalence values and how the predictive values change across that range. ### Dependence on the Gold Standard Predictive values are calculated using a gold standard to determine true disease status. If the gold standard is imperfect, the calculated predictive values will be inaccurate. Some animals will be misclassified as diseased or healthy, and this misclassification will affect the calculated values. The portosystemic shunt study used a combination of diagnostic imaging, surgical findings, and histopathology as the gold standard. The pheochromocytoma study used histopathology as the gold standard. These are generally reliable gold standards, but they are not perfect. ### Variation Across Subpopulations Predictive values can vary across subpopulations even when the overall prevalence is the same. For example, the prevalence of hypothyroidism in Eurasian dogs was found to be 3.9% overall, but the prevalence was higher in dogs with positive thyroglobulin autoantibody status. The predictive value of a thyroid test will differ between these two groups. The study of hypothyroidism in Eurasian dogs found that 22.0% of dogs with positive TgAA status were already hypothyroid on initial examination, and 42.5% of TgAA-positive dogs developed hypothyroidism on follow-up. This information allows the clinician to interpret thyroid test results differently in TgAA-positive dogs. ### Changes Over Time Predictive values can change over time as the prevalence of disease changes. A disease that is being successfully controlled through vaccination or biosecurity will have a declining prevalence, and the PPV of a positive test result will decline as well. The clinician should periodically reassess the prevalence of disease in their population and adjust their interpretation of test results accordingly. ## Frequently Asked Questions ### What is the difference between sensitivity and positive predictive value? Sensitivity measures the proportion of diseased animals that test positive, while positive predictive value measures the proportion of positive test results that are true positives. Sensitivity is a property of the test, while PPV depends on both the test and the prevalence of disease in the population being tested. A test can have high sensitivity and low PPV if the disease is rare. ### Why does disease prevalence affect predictive values? Prevalence affects the relative number of true positives and false positives among positive test results. When prevalence is high, most positive results are true positives because the pool of diseased animals is large. When prevalence is low, most positive results are false positives because the pool of healthy animals is large. The same logic applies in reverse for negative predictive value. ### How do I calculate positive predictive value from a 2x2 table? Divide the number of true positives by the total number of positive test results, which is the sum of true positives and false positives. Multiply by 100 to express as a percentage. For example, if there are 50 true positives and 1 false positive, the PPV is 50 divided by 51, which equals 98%. ### Why was the negative predictive value so low in the tonsillar culture study for Actinobacillus pleuropneumoniae? The NPV was 35.3% because the sensitivity of tonsillar culture was only 84.7% and the prevalence of lung colonization was high in the experimentally infected population. A negative tonsillar culture did not reliably rule out lung colonization because the test missed the organism in 15.3% of colonized pigs, and the high prevalence meant that false negatives were a substantial fraction of all negative results. ### Can I use published predictive values from a study in my own practice? Published predictive values can be used as estimates, but they must be adjusted for the prevalence of disease in your population. Published values from referral hospital studies will overestimate the PPV in a general practice population because referral populations have higher disease prevalence. Calculate the expected predictive values using the sensitivity and specificity from the study and the estimated prevalence in your population. ### What is the clinical significance of a biochemically silent pheochromocytoma? A biochemically silent pheochromocytoma does not produce elevated hormone levels, so it will not be detected by hormone testing. The pheochromocytoma study found that 21% of cases were biochemically silent. This means that a normal urine normetanephrine result does not rule out pheochromocytoma, and the clinician should consider other diagnostic modalities such as imaging when clinical suspicion is high. ### How should I interpret a positive test result when the disease is rare? When the disease is rare, a positive test result is more likely to be a false positive than a true positive, even when the test has high specificity. Confirm the result with a second test that has different test characteristics, or refer the animal for specialist evaluation. Discuss the uncertainty with the owner or producer and explain the need for confirmatory testing. ### What records should I keep to assess predictive values in my own population? Record the signalment of each animal, the reason for testing, the test result, and the final diagnosis confirmed by gold standard testing. For production animals, record the herd or group identification and the production stage. Periodically review these records to calculate the actual PPV and NPV of tests in your population and compare them to published values. ## Related Veterinary Guides - [Anesthetic Machine Leak Testing and Pressure Checks: A Step-by-Step Protocol](/knowledge/veterinary-medicine/anesthesia-analgesia/anesthetic-machine-leak-testing-pressure-checks) - [Urinalysis in Veterinary Practice: From Collection to Interpretation](/knowledge/veterinary-medicine/clinical-pathology/urinalysis-veterinary-practice-collection-interpretation) - [Capnography Waveform Interpretation in Veterinary Patients](/knowledge/veterinary-medicine/anesthesia-analgesia/capnography-waveform-interpretation-veterinary-patients) - [Zoonotic Disease Risk Assessment in Veterinary Practice](/knowledge/veterinary-medicine/veterinary-public-health/zoonotic-disease-risk-assessment-in-veterinary-practice) - [Zoonotic Disease Risk Communication in Veterinary Practice](/knowledge/veterinary-medicine/veterinary-public-health/zoonotic-disease-risk-communication-in-veterinary-practice) ## References and Further Reading - [Pet Care](https://www.avma.org/resources-tools/pet-owners). American Veterinary Medical Association. - [AAHA Guidelines](https://www.aaha.org/resources). American Animal Hospital Association. - [Global Guidelines](https://wsava.org/global-guidelines). World Small Animal Veterinary Association. - [Merck Veterinary Manual](https://www.merckvetmanual.com/). Merck Veterinary Manual. - [Cornell University College of Veterinary Medicine](https://www.vet.cornell.edu/). Cornell University. - [Animal Health and Welfare](https://www.woah.org/en/what-we-do/animal-health-and-welfare). World Organisation for Animal Health. - [Ultrasonographic diagnosis of portosystemic shunting in dogs and cats.](https://pubmed.ncbi.nlm.nih.gov/15487568). Veterinary radiology & ultrasound : the official journal of the American College of Veterinary Radiology and the International Veterinary Radiology Association, 2004. - [Evaluation of the predictive value of tonsil examination by bacteriological culture for detecting positive lung colonization status of nursery pigs exposed to Actinobacillus pleuropneumoniae by experimental aerosol infection.](https://pubmed.ncbi.nlm.nih.gov/29954395). BMC veterinary research, 2018. - [High specificity and sensitivity of spot urine normetanephrine-to-creatinine ratios in the diagnosis of canine pheochromocytoma.](https://pubmed.ncbi.nlm.nih.gov/39536456). Journal of the American Veterinary Medical Association, 2025. - [Laboratory indicators of hypothyroidism and TgAA-positivity in the Eurasian dog breed.](https://pubmed.ncbi.nlm.nih.gov/36693083). PloS one, 2023. - [Performance of QuantiFERON tests for detecting latent tuberculosis infections: A Meta-analysis.](https://pubmed.ncbi.nlm.nih.gov/41144091). Immunologic research, 2025. > This article is educational and is not a substitute for veterinary diagnosis or treatment. Contact a veterinarian for advice about an individual animal.