For two years, the number of coronavirus cases and hospitalizations has been widely used as a barometer of the spread of the global epidemic.

But the omicron wave is messing up the usual statistics, forcing news organizations to rethink how they report them.

“It’s just a data disaster,” said Katherine Wu, a staff writer covering COVID-19 for The Atlantic.

Cases spiked during the holiday season, an expected development given the emergence of a more contagious variant than its predecessor.

However, these figures only reflect what health authorities have reported. They do not include most people who self-test at home or who are unknowingly infected. Holidays and weekends can also cause lag in reporting cases.

If you could add up all those numbers — which you can’t — the number of cases could go up considerably.

For this reason, the Associated Press recently told its editors and reporters to avoid highlighting case numbers in its coverage of the disease. This means, for example, that there are no more single-day record numbers of cases that focus solely on a particular country or state, because that claim has become unreliable.

Looking at the media, more caution is being used when using official case numbers.

On Monday, NBC News’ report on the surge in COVID cases relied on the weekly average of cases. A story on Tuesday briefly referred to a “wave” of cases.

In reporting from Tuesday’s Senate hearing with health experts, the number of cases CNN flashed on the screen was a two-week average. MSNBC used a variety of measurements, including listing the five states with the most reports over the past three days.

In its “Pandemic Guidance” on its website, The Washington Post used a seven-day average of cases and compared that number to last Tuesday’s data, which showed a 56 percent increase. The New York Times used daily counts in the online chart, but also included two-week trends in cases and deaths.

An Associated Press report by Jennifer Sinco Kelleher and Terry Tang on Saturday said the “Omicron explosion caused services to collapse across the country” filled with statistics from across the United States on hospitalization rates or employee sick leave. The case count metric is not used.

“We definitely want people to report more deeply, to report more concretely,” said Josh Hoffner, a news editor who helped oversee the AP’s virus coverage.

During the omicron surge, Wu said, many news organizations were discussing how best to use statistics. But there is no easy answer.

“That’s how journalism works,” Wu said. “We need the data. We need to show the reader the receipt. But I’ll try to do that as carefully as possible.”

Some argue that hospitalizations and mortality are more reliable descriptions of the current impact of COVID-19 on society. In recent days, however, even the usefulness of these numbers has been called into question. In many cases, hospitalizations are accidental: Some people are admitted for other reasons and are surprised to learn they have tested positive for the coronavirus, said Tanya Lewis, senior editor of health and medicine at Scientific American.

Despite the flaws, the number of cases should not be ignored, said Gary Schwitzer, a lecturer at the University of Minnesota School of Public Health and publisher of HealthNewsReview.org, which monitors health coverage in the media.

The figures, he said, illustrate trends and paint a picture of which parts of the country have been hit particularly hard, or where the surge may have peaked. They can predict wider societal impacts, such as where hospitals are about to be incarcerated or where there will be worker shortages.

“These stories may not be told adequately if only hospitalizations and deaths are emphasized,” Schwitzer said.

This is also a point highlighted in the AP’s internal guidelines.

“They do have value,” Hoffner said. “We don’t want people to stop mentioning the number of cases.”

Some in public health and the press believe that the current surge — as painful as it is — could portend good news. The New York Times’ David Leonhardt and Ashley Wu wrote that this may indicate that rather than a devastating epidemic, COVID-19 is becoming an endemic disease that people learn to live with.

But if the past two years have taught anything, it’s the dangers of forecasting, Lewis said.

“We’ve been surprised time and time again,” she said. “We don’t know anything about the course of the pandemic. We still need to be humble and open to where things are going.”