When I was vaccinated against Covid-19, I felt a deep sense of relief: no more worries about personally catching the disease.

So when I noticed mild, Covid-19-like symptoms two months later — stuffy nose, chest congestion, and an upset stomach — I thought they were due to seasonal allergies. I was shocked a few days later when a test for Covid-19 done in preparation for an unrelated medical procedure came back positive.

Not believing the result, I got tested again. And again. Over a five-day period I had four PCR tests — two were positive, two were inconclusive — and a false negative rapid antigen test.

I quickly went about quarantining and notifying recent contacts. I soon began receiving multiple calls from my city’s Board of Health, whose representative told me she knew of several other people who had also tested positive even though they were fully vaccinated.

My experience shows that testing for Covid-19 isn’t perfect. Rapid antigen tests detect proteins that are part of SARS-CoV-2, the virus that causes Covid-19. They have a fast turnaround time, but aren’t as accurate as the gold-standard PCR test. In people with Covid-19 symptoms, rapid tests correctly detect the virus only 80% of the time. In those without symptoms, it drops to 40%.

PCR tests, in contrast, tend to have much higher sensitivities, above 95% across the board.

How samples are collected can also affect the results of Covid-19 testing. I know this happens because I have had many Covid-19 tests over the past year. At some test centers, the clinician inserted the swab into the very back of my nose, known as the nasopharyngeal space. Some say it feels like the swab is heading into the brain. In other tests centers, the clinician barely scraped the lower rim of my nose; that’s called a nasal swab. Those made me wonder if it collected any sample at all. I’ve since learned that nasal swabs can miss detecting the virus in those with low viral loads.

While vaccination confers essentially 100% protection from Covid-19-related hospitalization and death, it doesn’t entirely prevent people from catching the infection in the first place.

A recent study from the University of California, San Francisco, tracking more than 30,000 health care workers showed an infection rate after vaccination of approximately 1%, meaning that infection is rare but still a threat. The Minnesota Department of Health announced it was investigating several of these “breakthrough” cases, as has the Oregon Health Authority. In mid-April, the Centers for Disease Control and Prevention reported 5,800 breakthrough infections to date among the millions of Americans who have been fully vaccinated.

Reports of 95% efficacy rates in vaccine clinical trials do not necessarily translate to real-world effectiveness of that magnitude, and can create a false sense of reassurance when it comes to asymptomatic or mild infections. The Pfizer trials only tested volunteers for SARS-CoV-2 if they developed symptoms after getting the vaccine or the placebo, leaving out those who may have been asymptomatically infected. Moderna trials also primarily looked at preventing symptomatic disease.