Antibody tests for Covid-19 are becoming widely available across the United States. Some companies, like Quest Diagnostics, even offer at-home kits. Whether you take one in a doctor’s office or your bedroom, an antibody test can, ideally, determine whether you’ve previously been infected with Covid-19. It does so by searching your blood for antibodies, which are proteins the immune system makes when it encounters the coronavirus.

While it’s likely people develop some immunity after being infected with the coronavirus, experts still don’t know whether you need a certain amount of antibodies to fight off another infection. And even if you do have the right amount, it also isn’t clear yet how long its protection lasts. (The pandemic was only officially declared on March 11, so there hasn’t been time to study long-term immunity yet.)

If you’ve felt even remotely sick in the past few months, just knowing whether you ever had Covid-19 could be a welcome relief. But there’s one big problem: These tests can be so inaccurate that they might not tell you anything meaningful.

Guidance updated this week by the U.S. Centers for Disease Control and Prevention (CDC) states that antibody tests are useful for determining the proportion of people in a population that have been sick before, as well as other patterns of disease spread. It also says that the tests should not be used to determine whether a person is immune.

As the first tests began appearing as early as late February, scientists raised concerns about their accuracy, which varied widely — too widely for the tests to be useful in making decisions about personal health. The same holds true today, though the accuracy of some new tests is more promising than others. In May, the FDA began pulling antibody tests from the market after updating its guidelines to weed out products making fraudulent claims.

Here’s the latest on what we know about antibody tests and what we’re still waiting to find out.

Just because you have antibodies doesn’t mean you’re protected from Covid-19, so following social distancing and proper hygiene is just as important as it was before you got the test.

Results

If you decide to take an antibody test, you’ll get one of three results: positive, negative, and equivocal. Interpreting these results can be tricky, as I wrote on Medium’s Coronavirus blog in early May and am republishing here, as the science on coronavirus testing hasn’t budged since.

Positive

If you test positive for Covid-19 antibodies, it means you were previously infected with Covid-19. What it doesn’t mean is that you are now definitely immune to Covid-19. The CDC’s updated guidance says that a positive antibody test result “likely indicates at least some degree of immunity,” but it also points out that “until the durability and duration of immunity is established, it cannot be assumed that individuals with truly positive antibody test results are protected from future infection.” This means that while people likely develop some immunity to Covid-19, scientists still don’t know whether the mere presence of antibodies means people can’t get the disease again or how long any immunity lasts. This may seem confusing since infection to diseases like chickenpox or measles confers lifelong immunity, but the body, unfortunately, doesn’t remember all viruses equally well.

Scientists also don’t know whether you need a certain amount of antibodies in your blood to fight off an infection if you encounter the virus again. Most antibody tests can only give you a yes or no answer, and even if you did get a test that told you how many antibodies you had, it’s unclear how many you’d actually need to mount an immune response.

A positive result, however, could make you eligible to donate your blood plasma in order to help people who are very sick with Covid-19. Small trials involving “convalescent plasma” use antibodies from recovered patients to treat people with severe illness.

The short of it is: Just because you have antibodies doesn’t mean you’re protected from Covid-19, so following social distancing and proper hygiene is just as important as it was before you got the test.

Negative

If you test negative for Covid-19 antibodies, it could mean that you have never been infected with Covid-19. But it could also mean that you were infected previously and produced antibodies but lost them over time, that you were infected and your immune system didn’t make antibodies or didn’t make enough to be detected, or that you are currently infected but your body hasn’t started to make antibodies yet.

Unfortunately, there are still a lot of unknowns when it comes to immunity against Covid-19, so interpreting test results is tricky. For one thing, it isn’t clear how long antibodies last, once you make them; it’s possible that antibodies fade away over time, much like our protection from viruses that cause the common cold disappears, leaving us susceptible to infection year after year.

Scientists also don’t know for certain if everyone makes antibodies against the virus, or if everyone makes the same amount. As previously reported in Elemental, some papers show that everyone makes antibodies, and experts say it’s likely most people produce some immunity. But there is at least one preprint showing that 10 out of 175 people who recovered from Covid-19 didn’t. Experts also say that people produce different amounts of antibodies depending on their gender, age, and underlying diseases. So, it is possible that you were infected but didn’t make antibodies, or that you didn’t make enough of them for the test to detect, producing a negative result.

As mentioned earlier, it takes one or two weeks for the immune system to start making antibodies, once it detects an infection. That means that a test can come back negative for antibodies if a person is currently infected — which is why you should wait a few weeks to get an antibody test if you suspect you were infected (but didn’t get a diagnostic test to confirm it).

The bottom line, though, is that a negative result means you likely don’t have antibodies against the virus, which means it’s as important as ever to follow social distancing and proper hygiene.

Equivocal

An equivocal or “borderline” result is one where test results can’t be interpreted as either positive or negative. The reason for this rather unsatisfying outcome could be that the antibody test detected an antibody for a different illness, that the immune system hasn’t made an adequate amount of antibodies, or that the test is detecting a waning immune response. If you get this result, you’ll likely be encouraged to get retested in a few weeks.

In the meantime, and just as with positive and negative results, you should continue to follow social distancing rules and proper hygiene, just as before.

Sensitivity, Specificity, and Accuracy

Two terms often come up in conversations about an antibody test’s accuracy: sensitivity and specificity. One antibody test made by Abbott, for example, claims 99% accuracy, which the company backed up with data showing that it had “100% sensitivity” and “99.6% specificity.”

Sensitivity refers to how good a test actually is at detecting the antibody it’s looking for. A 100% sensitive test won’t accidentally miss antibodies and incorrectly come back negative — an error known as a false negative.

Specificity refers to a test’s ability to detect only the specific antibody it’s looking for — and not get fooled by other molecules. A 100% specific test wouldn’t accidentally give you a positive result when in fact you don’t have the antibodies in question, an error known as a false positive. These are especially dangerous errors because they could lead a person to mistakenly believe they have immunity to the virus when they actually don’t.

But unless a test has 100% sensitivity and 100% specificity — an impressive achievement — a test can’t unequivocally tell you whether or not you have antibodies. That’s because calculating the likelihood that a test gave you the correct result also depends on how many people in your area are actually sick with Covid-19, a measure known as its prevalence. You can find an explainer of the rather complex math behind this calculation here, but it’s helpful to consider that if the prevalence in your area is 1%, even a test with 99% sensitivity and specificity would only give you a correct result 50% of the time.

Perhaps the better question is: What are you going to do with your antibody test results?

Some tests are better than others

When antibody tests first showed up on the U.S. market in early March, the FDA’s rules were relatively loose, especially compared to its strict regulations that controversially held up diagnostic testing in the early days of the pandemic. The FDA’s early rules allowed vendors to sell their antibody tests as long as they self-vetted their product, which naturally opened the door to fraudulent tests.

Now, the FDA is now cracking down on antibody tests. New rules announced on May 4 stated that vendors “must apply for emergency FDA authorization for the test within 10 days of the rules change, or within 10 days of notifying FDA of the intent to sell new tests.” Emergency use authorization allows a product to be fast-tracked to market with less vetting in the case of emergencies, like a global pandemic. On May 21, the FDA named 27 antibody tests that could no longer be sold because the companies that made them didn’t comply with the new rules. The list of companies originally authorized to sell the tests numbered over 200.

There’s a handy table on the FDA website of antibody tests (also called “serology tests”) that have received Emergency Use Authorization. It currently lists 13 tests.

Details about the tests include the FDA’s estimates of each test’s sensitivity and specificity. They also include useful metrics called positive predictive value (PPV) and negative predictive value (NPV), which the FDA calculates by rolling together sensitivity, specificity, and an assumption about prevalence. They’re basically making an attempt to combine all these factors into a single rating that tells you whether a positive result really means you have antibodies (PPV) and whether a negative result really means you don’t (NPV). But keep in mind that the actual prevalence of your area matters a lot — and varies widely. There is no single resource that provides prevalence data for specific regions, so your local health department likely has the most reliable information on Covid-19 prevalence in your area.

If you’re looking for an antibody test to take, it’s definitely worth looking through the details on each test, but below are some highlights. Notably, there are no tests that have both 100% PPV and 100% NPV — which perhaps goes to show how hard it is to get both specificity and sensitivity right. And yes, the test names are a mouthful!

High PPV. Several tests got a score of 100% at 5% prevalence, which means that if you get a positive result from any of these tests and you live in a place where prevalence is 5%, you can be confident that you actually do have antibodies.

• EUROIMMUN’s SARS-COV-2 ELISA (IgG) test
• Mount Sinai Hospital’s Clinical Laboratory COVID-19 ELISA Antibody Test
• Ortho-Clinical Diagnostics’ VITROS Anti-SARS-CoV-2 IgG test
• Ortho-Clinical Diagnostics’ VITROS Immunodiagnostic Products Anti-SARS-CoV-2 Total Reagent Pack and Calibrator

High NPV. Some tests got a score of 100% at 5% prevalence, which means that if you get a negative result from any of these tests and you live somewhere with a prevalence of 5%, you can be confident that you really do not have antibodies.

• Abbott’s Alinity i SARS-CoV-2 IgG
• Abbott Architect SARS-CoV-2 IgG
• Ortho-Clinical Diagnostics VITROS Immunodiagnostic Products Anti-SARS-CoV-2 Total Reagent Pack and Calibrator
• Roche Elecsys Anti-SARS-CoV-2

What to do if you haven’t had symptoms

Given the varying accuracy of antibody tests and their dependence on widely varying prevalence, you’re probably wondering: Should I get an antibody test? It may be especially perplexing for people who haven’t had symptoms and are wondering whether they were sick without even knowing it.

But perhaps the better question is: What are you going to do with your antibody test results?

Even if we assume 100% accuracy, an antibody test result still can’t tell you if you’re in the clear and can’t get Covid-19 again. In April, the WHO said it did not recommend the use of antibody test results for patient care, and it hasn’t updated its guidance since.

What getting tested can do is help satisfy your curiosity about whether you were sick in the past — as long you can stomach the potential that the results might be wrong. Although cities are reopening across the country, an important thing to keep in mind if you’re thinking about getting tested is that going outside anywhere — especially to a doctor’s office where sick people are walking in and out — still carries some risk, since the virus hasn’t been eliminated in the United States. Especially if you venture outside, hand-washing, mask-wearing, and social distancing remain as important now as they did at the beginning of the pandemic.