treatment required?
Is the disease life‐threatening?
Is there a zoonotic risk?
Diagnosis of existing disease states can also protect the health and welfare of individual animals and, in the case of transmissible or zoonotic diseases, the health of the entire shelter population, both animal and human. In addition to those indications described for screening tests, when it comes to disease diagnosis, practitioners should also consider how the results of a particular diagnostic test will alter the treatment or management plan for the individual animal or the shelter population at large.
4.4.2 Population Level Testing
There are two high‐level indications for diagnostic testing on a population level, including (i) disease surveillance and (ii) diagnosis and management of a disease outbreak. Surveillance has been defined as “the use of data to monitor health problems to facilitate their prevention or control” (United States Department of Health and Human Services 2012). It involves the active collection of data, such as clinical signs and diagnostic test results, within an at‐risk population and is used to set priorities, plan and conduct disease control programs, and assess the effectiveness of control efforts (United States Department of Health and Human Services 2012). In the animal shelter, disease surveillance can help identify and document common conditions in incoming animals, detect disease trends within the shelter, and identify emerging diseases (Hurley and Pesavento 2013).
Though disease surveillance often occurs informally, such as monitoring for general trends during daily rounds, scheduled and formalized surveillance can provide more objective, actionable data. One example of this might be a periodic evaluation of fecal samples from dogs admitted within one week to assess concordance with prophylactic deworming protocols. The mathematical goal of surveillance testing is to detect at least one animal with the infection or disease of interest and not to determine prevalence in the shelter population; therefore, the ideal number of animals to test should be calculated based on the lowest assumed rate of infection and the desired confidence limits. Sampling enough animals to ensure at least 90% probability of detection is recommended (Table 4.6).
In addition to testing clinically healthy animals, it is useful to perform disease surveillance in cohorts of animals with specific clinical syndromes. For example, collecting oropharyngeal swabs from cats with clinical signs of upper respiratory disease may help guide empirical treatments and direct husbandry practices in order to minimize disease transmission once particular pathogens are identified. This type of surveillance may help inform preventive care practices and be helpful in understanding baseline disease characteristics should frequency or severity of clinical syndromes increase or animals fail to respond to treatment as expected.
Table 4.6 Surveillance testing sample sizes for ≥90% probability of detection (National Research Council 1991).
| Assumed Infection Rate (%) | Number of Animals to Test |
|---|---|
| 2 | 120 |
| 3 | 80 |
| 4 | 60 |
| 5 | 45 |
| 10 | 25 |
| 15–20 | 15 |
| 25–30 | 10 |
| ≥40 | 5 |
The frequency of surveillance is dependent upon the importance of the specific disease to the management of the population, historical trends, and resource availability; typical intervals include monthly, quarterly, biannual, or annual testing (National Research Council 1991). Results should be interpreted in terms of time (e.g. seasonal disease trends), location (e.g. building, ward, housing unit), and the individuals affected (e.g. species, age, sex, clinical signs) (United States Department of Health and Human Services 2012).
Diagnostic testing plays a key role in individual animal risk assessment during a disease outbreak response. Definitive identification of the cause of an infectious disease outbreak is a critical step in its mitigation and control (O'Quin 2013). In such circumstances, consideration should be given to testing of animals that are presumed to be infected; exposed and at risk of infection; exposed but not at risk of infection; and those not yet exposed (Hurley and Pesavento 2013). Testing every animal is generally not necessary and may not yield more actionable results. Though the precise number of patients to test will depend on the stage of infection, diagnostic test accuracy, and disease prevalence, a minimum of three to six samples for both GI and respiratory diseases have been recommended in human outbreak investigations (British Columbia Provincial Infection Control Network 2011; Plantenga et al. 2011). See Chapter 6 for more information on this topic.
4.5 Accuracy and Testing Strategy
The accuracy of diagnostic test results is commonly assessed through calculation of sensitivity, specificity, and predictive value reported as percentages. Sensitivity refers to a test's ability to report a positive test result in an animal that is actually infected. It can be calculated by dividing the number of true positive test results by the total number of cases that are actually positive (true positives + false negatives). A test with high sensitivity will report few false negative results. Specificity refers to a test's ability to report a negative test result in an animal that is not infected. It can be calculated by dividing the number of true negative test results by the total number of cases that are actually negative (false positives + true negatives). A test with high specificity will report few false positive results. Sensitivity and specificity refer to the characteristics of a specific diagnostic test, clinical sign, or disease syndrome, and remain constant regardless of the population being evaluated (Rothman 2012).
Predictive value refers to the usefulness of a particular test, sign, or syndrome in classifying animals with the infection or disease of interest. As such, predictive value is highly dependent on the prevalence of disease within the population being evaluated (Rothman 2012). Whereas sensitivity and specificity report the percentage of accurate test results, predictive value reports the percentage of patients that are likely to have accurate results. Positive predictive value can be calculated by dividing the number of true positive test results by the total number of positive test results (true positives + false positives). A test will have a high positive predictive value when the disease is common in the population and the diagnostic test utilized has high specificity. Negative predictive value can be calculated by dividing the number of true negative test results by the total number of negative test results (true negatives + false negatives). A test will have a high negative predictive value when the disease is not common in the population and the diagnostic test utilized has high sensitivity. See Figure 4.2.
Figure 4.2 Sensitivity, specificity and predictive value. TP, true positives; TN, true negatives; FP, false positives; TN, true negatives.
Understanding these measures of accuracy is critical to designing rational diagnostic testing strategies for both individual animals and the shelter population as a whole. With each diagnostic test employed, there comes a risk of either false positive or false negative test results that may have