Richard Lerner is a veterinarian and epidemiologist. He left small-animal practice early in 2009 and now works with the Commonwealth of Massachusetts on food safety issues, especially those regarding animal agriculture.
Here’s a question I routinely ask myself: Why am I doing exactly what it is I’m doing? Recently, the activity I’ve been questioning involves running the 4DX test made available by IDEXX. (The “4” in the name refers to the number of diseases for which it tests: heartworm, Lyme, Ehrlichia canis and anaplasmosis. The first of these is spread via mosquitoes, the other three by ticks.)
I’ve never owned a veterinary clinic, so I’ve never been The Decider; when debating the “what needs to be done on a routine visit” question, someone above my pay grade has set the rules. I’d like to say that the decisions all have sound scientific and medical reasoning; alas, I cannot. (Even at the most august of institutions, I’ve been asked to give vaccinations for economic rather than medical reasons.)
Preventing disease is good, and tests help us detect diseases. Therefore, tests are a good thing, right? Were it that simple! Mass screening, like mass vaccination, is rarely an unqualified good. Remember, screening means checking clinically healthy animals for markers of a disease, and positive results lead to further confirmatory tests. When we test a sick animal for diseases that may cause a particular constellation of symptoms, we’re not screening—we’re diagnosing.
So when is it appropriate to screen? First, the disease in question needs to be significant—unlike, say, strep throat, a common childhood infection caused by the bacterium Streptococcus pyogenes. Children go to the doctor every year for checkups, but they are not routinely screened for the presence of strep bacteria because we don’t see strep throat as a serious disease. The possibility of complications doesn’t warrant the cost of testing every asymptomatic child.
Second, there needs to be a significant Detectable Preclinical Phase (DPCP), a period of time between the initial infection and the point at which symptoms appear, which allows us to intervene before the clinical phase starts. The classic in canine medicine is heartworm disease; we can detect the disease from a drop of blood and then treat positive cases before the dog becomes overtly ill.
Third, the screening test should be easy, noninvasive and inexpensive; for humans, the classic is blood-pressure screening.
Fourth, the test has to be accurate; false positives are not benign false alarms—they are followed by confirmatory diagnostics, worry and, often, unnecessary treatment.
The real bugaboo lies in what is called the Positive Predictive Value (PPV) of a test, or how likely it is that a positive result really is positive. Roughly speaking, PPV is a real-world measure that varies depending on how common a disease is. Sensitivity and specificity are laboratory values, a comparison of the test in question with a “gold standard” test. Tests like the 4DX are marketed by IDEXX based on sensitivity and specificity, not on the PPV.
Back to our 4DX: What is its PPV? To know that, we would need to know how prevalent ehrlichiosis and anaplasmosis are. Their occurrence varies by region, and in general, we don’t know. Nor do we know how often infection actually results in disease. We do know that these diseases are uncommon, if not rare, in most of the country, which lowers the test’s PPV. Moreover, in cases that do occur, most are acute and therefore don’t have a significant detectable preclinical phase—if the dog is infected, he will exhibit signs long before the next annual visit. In short, we have a test that, at best, shows whether a dog may have been exposed to these diseases, not whether or not he actually has them.
It should be clear that regular screening for any acute onset disease, including acute onset Lyme disease, is not necessary, and may even be counterproductive. As to whether chronic Lyme disease is actually a recognizable clinical entity in the dog: Well, the profession may need to reassess this. There is growing evidence in human medicine that “late Lyme,” or a chronic form of the disease, is far less prevalent than previously thought. While dogs are not people, there is no reason at this point to think it’s different for them.
Unfortunately, many vets treat or order additional diagnostics based on a positive test. This is often inappropriate. In the words of noted veterinary immunologist Ronald Schultz, “Treat the dog, not the test.” In the mind of many practitioners, a test that creates more doubts and questions than it resolves is worse than neutral.
If you want your dog tested for heartworm, and not for Lyme, ehrlichiosis, or anaplasmosis, that should be your right. Should we be on the lookout for these diseases? Absolutely. But whether or not veterinary clients should have to pay for this surveillance with tests of unproven value is a debatable matter. And as long as your veterinarian is selling the test, he is hardly an unbiased party.
Author disclosures: First, I have owned a small (i.e., not enough for anyone to care if I use my proxy votes) number of IDEXX shares for years. Second, IDEXX has been generous, assisting me with research that I hope will benefit both people and dogs in developing nations. Finally, my ownership of IDEXX shares reflects the fact that I believe their science is sound. This article concerns what marketing departments and veterinarians choose to do with the tools scientists provide.