As rapid tests are becoming more widely available, delivering results in minutes in doctor's offices, nursing homes, schools and even the White House, officials warn of a significant undercount, blurring the spread.
Officials say that antigen tests, which are faster than polymerase chain reaction (PCR) tests but less able to detect low levels of the coronavirus, are an important tool for limiting the spread. But they caution that with inconsistent public reporting, the case undercount may worsen.
"We want to be sure that we're not now saying, 'there's no disease,' when there is lots of disease. All that's happened is that the science with which we identify it's evolved," said Dr Janet Hamilton, the executive director of the Council of State and Territorial Epidemiologists, the group that helps the Centres for Disease Control and Prevention (CDC) define cases of the coronavirus.
Experts say that more rapid testing will lead to more undercounting of cases, even in states that attempt to count them publicly. What's going on?
Some states do not count cases based on antigen tests.
Despite CDC guidance to report cases based on PCR and antigen testing, Washington DC and seven states don't publicly share case counts for those with antigen positive tests, including California, New Jersey and Texas.
Another six states keep these tallies separate from their total case counts, and most of these report them less frequently.
The differences among states are due in part to each state's comfort level with the rapid tests, which aren't "confirmatory" like PCR tests because they can miss low levels of the virus.
Yet most states treat antigen positive cases or "probable" cases the same as "confirmed" cases, by following up with interviews and contact tracing.
And a growing number of states, including New Mexico, Oregon and Utah, include individuals with positive antigen tests in their total confirmed case counts.
Individuals who test positive from rapid tests are a small but growing share of all publicly reported cases.
For example, in Florida, positive cases from antigen tests make up about 4 per cent of all cases reported since March, the majority of which are based on PCR. But on a daily basis antigen tests contribute a more visible share, making up as much as a third of reported cases in recent days.
These counts are expected to increase as rapid tests purchased by the federal government reach schools and nursing homes across the country this fall.
And cases based on antigen tests can be even more significant in counties where these rapid tests have been more abundant.
Non-traditional testing centres can leave states in the dark.
Whether states publicly report antigen positive cases or just track them internally, several public health officials say their counts are incomplete because they don't know where rapid testing is taking place in their jurisdictions.
And unlike laboratories that typically perform more complicated PCR tests, several of the "point-of-care" centres doing antigen testing, such as nursing homes, urgent care centres and schools, either don't realise they need to report lab data, or may rely on slower, less efficient methods such as phone calls and faxes.
"We don't know for sure what we don't know," said Dr Edward Lifshitz, medical director of the Communicable Disease Service at the New Jersey Department of Health, which doesn't publicly report antigen positive cases because the figure is incomplete.
Dr Lifshitz added that sharing data on antigen positive cases might make some parts of the state, where point-of-care testing centres are reporting correctly, appear worse off than other areas that have missing data.
"It will look like, 'boy, this part of the state is having an increase in cases,' when it's really not," he said. "It's just that they are reporting what they are supposed to be reporting."
To avoid missing antigen test results, some health departments have had to do their own outreach.
In Houston, where infections surged over the summer, health officials searched for rapid testing centres online, made phone calls and posted a "dear provider" letter on their website.
In New Jersey, officials have been collecting testing facility names from antigen test manufacturers and the Centres for Medicare and Medicaid Services, which distributed rapid tests to nursing homes across the state.
But even then, testing centres can still be missed.
Since late September, Alabama has reported three spikes of older cases because point-of-care facilities had neglected to report to public health officials their antigen test results. An urgent care group was responsible for the largest increase, which was reported on Oct 23, after the addition of older cases nearly tripled the daily count.
Alabama health officials say that as they learn of new testing centres, they are training them to share test data in an electronic format that they can easily work with.
Health officials across the country in states like California, Georgia, New Jersey, Oklahoma and Texas, are doing the same, but training can be labour and time intensive, and several testing centres don't have the technical support they need.
"This is a pipeline that is not well established," said Dr Kirstin Short, bureau chief of epidemiology at the Houston Health Department.