Search DrRajHealth

Saturday, May 16, 2020

FDA Cautions About Accuracy Of Widely Used Abbott Coronavirus Test

COVID-19 Testing, testing, testing. Is it worth it? We've been asking this question on the healthcare frontlines since day one of this pandemic. More specifically, is a negative diagnostic test for COVID-19 a true negative or a false negative? A false negative is analogous to an enemy combatant successfully getting past your security checkpoint because you didn't identify them as such and now they are on your base and will destroy your troops from within. So, what frontline health care is really interested is a test with a near 100% true negative result. This true negative result is called the "specificity" of the test. This is in contradistinction to the sensitivity of the test (previously discussed in an earlier post). The reason why a highly sensitive test is not the most essential but a highly specific test is is because when a disease has a high prevalence in the community, the expectation is that many, if not most, are infected whether we have a test that says so or not. So, the high population prevalence of the disease is a highly sensitive "pre-test" already. An analogy would be weather forcasting. If you wish to know if it's raining in your neighborhood and you look outside and rain is falling from the sky, why do you need a further test to tell you it's raining? The prevalence of rain is 100% so, any test that tells you it's not raining, including the meteorologist on TV saying it's not raining, would not change your assessment and it would be considered a "false negative." Now, if you looked outside and it was not raining but there were dark, heavy storm clouds and the smell of water in the air, you might have a "pre-test" probability of 80% that it's going to rain. If the meteorologist concurs that it will rain today, that might move your pre-test expectation of rain closer to 100% (this is called the positive predictive value (PPV) of a test because the positive results of that test further further strengthened your pre-test expectations). But, if the meteorologist says, it's overcast with a low chance of rain, your pre-test expectation of rain may drop lower than 80%--perhaps the rain clouds will blow on over and deliver rain elsewhere, but not to you (this is called the negative predictive value (NPV) of a test because the test did not support your pre-test expectations). If the meteorologist offers no insight and shrugs her shoulder and says, "maybe it will rain, maybe not." Then that does not change your own pre-test expectations. You're still no better off than looking out the window and sticking with your assumption. So, when we're in a pandemic, why are we so concerned with a diagnostic test to tell us the disease is here? Afterall, isn't that what a pandemic is--the disease is everywhere? If you've left your backyard, you're at risk for getting it. Plus, what is the utility of a single, one-time test anyway? Once negative, is not always negative. Let's say you go and get tested today and the test is negative for the virus, but while walking out the testing center, an incoming person sneezes in your face, what's the value of that test you just got? It may have been instantly invalidated moments after you took it. And, what about the next day and the day after that and the week after that as you go about your life in the community--shopping, driving, neighborhood walks, interacting with other family and friends who are doing likewise? Are you going to get tested on a daily or weekly basis? Is that practical for each individual in society to get serial testing? Do we have sufficient testing supplies for that and sufficient laboratory capacity to run all those tests? It's more pragmatic to test those who are symptomatic and those undergoing high risk medical procedures where the risk of infecting others in the procedure room is high. Otherwise, for the rest of the community, continue droplet precautions--mask, hand hygiene, maintain social distance. There are reasons for diagnostic testing community-wide such as to define the case fatility rate as well as the infectivity vs virulence of the virus. More to come.
from Rajesh Harrykissoon, MD

No comments:

Post a Comment