Understanding the Maps: Key Factors that Influence the Resultsby Heather Walden MS, PhD
The CAPC Parasite Prevalence Maps are designed to show the proportion of pets tested which test positive for a given infection using available assays. This data is very useful to veterinarians and others trying to better understand the prevalence of parasites in a given practice area, but as with any survey, the results can be influenced by a number of factors, including the number of pets tested, the history of the pets prior to testing, the reason the pets were tested, and the assays used. Understanding each of these factors is critical to accurately interpreting the geographic distribution of these important disease agents.
1. Sample Size
The number of tests conducted is the denominator used and, as such, directly influences the percentage of dogs which will be reported as testing positive. Inadequate sample size is the most common limitation in being able to accurately interpret results from any survey tool. When no pets are tested, we cannot draw any conclusions about the prevalence of disease. However, when too few pets are tested, the small size of the denominator (sample size) will artificially inflate a few positive samples or obscure the importance of positives not yet detected. In other words, small sample size precludes accurate interpretation of any survey. In general, as sample size increases, so does the statistical power, or validity, of the data.
Because of these complications, caution should be employed in interpreting any percentages generated by testing a small number of pets, particularly when the pathogen presence is entirely unexpected. While these unexpected results could be due to recent spread or recognition of a disease agent in a given area, they can also result from a number of other factors, including the background or source of the pets tested or the nature of the tests used (see below) rather than true presence of infection.
2. History of Pet Prior to Testing
Certain pets are more likely to test positive than others. For example, a pet that was recently rescued or adopted may not have been protected from infection previously and thus is more likely to harbor parasites. In addition, pets that have recently moved into an area may arrive with infections acquired previously. Dogs imported from a high prevalence area for heartworm and tested after arrival may result in a number of “new” heartworm cases appearing in a location where infection historically was relatively rare. Similarly, dogs that have antibodies to the agent of Lyme disease may move with their owners to another region, such as the southern United States, where infection with the disease is much less common.
Complete historical information is not available for each of the dogs tested and included on the maps. However, unexpected positive results, particularly when clustered around areas with frequent population turnover such as metropolitan areas, military bases, retirement communities, and college towns, should be interpreted with caution as they may represent infections in translocated pets rather than locally acquired infections.
3. Motivation for Testing
The diagnostic assays from which the results are used to create the maps may be performed by veterinarians for diagnosis of suspected infection or for routine screening. Veterinarians are judicious in recommending diagnostic assays for a given patient and respectful of limited client resources. Accordingly, pets that are more likely to harbor infections, including those exhibiting clinical signs of disease or those with previous infections facilitated by lifestyle or geographic location, are more likely to receive a test from a veterinarian. When pets that are more likely to be infected are disproportionately tested, the number of positive results increases as does the calculated percentage of positive test results. This increase in positive results is more likely to be seen in areas where the tests are used more commonly for diagnostic verification of a suspected infection than for routine screening. At this time, the data do not include the reason for testing and variations in motivation throughout the United States may bias local prevalence up or down.
4. Nature of assays used and impact of local or population-based prevalence of infection
Although assays used in testing pets often appear to be similar, individual tests vary in sensitivity and specificity, which can influence overall results. Tests with poor sensitivity may underestimate the prevalence of infection in a given population, while tests with poor specificity can overestimate the presence of infection as false positive results are recorded. In addition, the predictive value of a given assay, or the actual precision in use, is dependent not only on individual test performance but also the prevalence of infection in the population considered.
For a given test, if sensitivity and specificity are held constant, predictive value tracks with prevalence; higher prevalence results in a higher positive predictive value, meaning any given positive result is, on average, more likely to be a true positive so long as infection is fairly common. When infections are rare, the positive predictive value decreases and the negative predictive value increases. In effect, the tests perform differently in different geographic areas and in different populations of animals, and the results should be interpreted accordingly.
For example, in the southern United States, heartworm infection is relatively common in adult dogs not on preventive. In this group of dogs, a positive test result is likely to be a true positive (high positive predictive value). However, for a dog maintained on heartworm preventive year-round, infection is relatively rare, and thus a negative test result is likely a true negative (high negative predictive value) while a positive test result is more likely to be a false positive (lower positive predictive value). Because false positive results can occur on antigen tests, positive antigen test results should always be confirmed by performing a second, independent, laboratory-based assay prior to instituting treatment. This confirmation is particularly important in areas where heartworm infection is less common and in individual dogs for which the positive result in unexpected, i.e. dogs maintained on heartworm preventive.
5. Antigen and Antibody
The assays detect two different types of biomarkers in the pet's blood, antigen or antibody. Antigen is a protein released by a pathogen and detected by the test. In general, when a test is positive, and detects antigen, the animal is considered to have an active infection. Antibody is a component of the pet's immune system that is produced in response to infection by the pathogen and detected by the test. Antibodies often persist for some time after infection, and may be detected during an active infection and after the infection has been resolved. Therefore, a positive result for a test that detects antibodies only indicates that the animal has been exposed to the pathogen at some point in the past. Other clinical findings help determine whether the animal has an active infection.
Understanding the difference between the two types of tests is very important when interpreting the maps. The canine heartworm test is antigen-based, so the prevalence map is indicative of actively infected animals. The feline heartworm tests may be antigen or antibody-based and must be interpreted accordingly. The tick-borne pathogen tests are antibody-based, so the prevalence maps are indicative of either an active or a resolved infection with the pathogen.
In epidemiologic studies, there is a study population and a target population. The study population includes people or animals that are directly observed during the course of the study. It is also a representation of the target population, or general population. The study population is, therefore, a sample of the target population. The test results used to produce the prevalence maps are from animals that were presented to a veterinary clinic for a variety of reasons (see 3 above: Motivation for Testing). The vast majority of these animals are owned, while a smaller subset of the animals may be in custody of a rescue organization or a shelter. As such, the prevalence maps only represent the owned dog population presented to a veterinary clinic. Owned dogs not presented to a veterinary clinic or not tested, stray and feral dogs, and those in the custody of rescues or shelters that do not present to the veterinary clinic are not represented. It is probable that preventive care is provided less to these animals, though no large-scale studies exist on this subject. The estimated prevalence provided by the maps is conservative and it is believed that the prevalence of each pathogen is much higher in the whole population.