By DAVID BAUD­ER | AS­SO­CI­AT­ED PRESS

NEW YORK (AP) — For two years, coro­n­avirus case counts and hos­pi­tal­iza­tions have been wide­ly used barom­e­ters of the pan­dem­ic’s march across the world.

But the omi­cron wave is mak­ing a mess of the usu­al sta­tis­tics, forc­ing news or­ga­ni­za­tions to re­think the way they re­port such fig­ures.

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

The num­ber of case counts soared over the hol­i­days, an ex­pect­ed de­vel­op­ment giv­en the emer­gence of a vari­ant more trans­mis­si­ble than its pre­de­ces­sors.

Yet these counts on­ly re­flect what is re­port­ed by health au­thor­i­ties. They do not in­clude most peo­ple who test them­selves at home, or are in­fect­ed with­out even know­ing about it. Hol­i­days and week­ends al­so lead to lags in re­port­ed cas­es.

If you could add all those num­bers up — and you can’t — case counts would like­ly be sub­stan­tial­ly high­er.

For that rea­son, The As­so­ci­at­ed Press has re­cent­ly told its ed­i­tors and re­porters to avoid em­pha­siz­ing case counts in sto­ries about the dis­ease. That means, for ex­am­ple, no more sto­ries fo­cused sole­ly on a par­tic­u­lar coun­try or state set­ting a one-day record for num­ber of cas­es, be­cause that claim has be­come un­re­li­able.

 Through­out the me­dia, there has been more cau­tion in use of of­fi­cial case counts.

An NBC News sto­ry on Mon­day about the sky­rock­et­ing num­ber of COVID cas­es re­lied on a one-week av­er­age of case counts. A Tues­day sto­ry sim­ply re­ferred to a “tidal wave” of cas­es.

Dur­ing its cov­er­age of a Sen­ate hear­ing with health ex­perts on Tues­day, the case counts CNN flashed on­screen were two-week av­er­ages. MSNBC used a va­ri­ety of mea­sure­ments, in­clud­ing a list­ing of the five states with high­est re­port­ed num­bers over the past three days.

On its web­site’s “Guide to the Pan­dem­ic,” The Wash­ing­ton Post used a sev­en-day av­er­age of cas­es and com­pared that num­ber to last Tues­day’s, show­ing a 56% in­crease. The New York Times used a dai­ly count in an on­line chart, yet al­so in­clud­ed a two-week trend in both cas­es and deaths.

An AP sto­ry Sat­ur­day by Jen­nifer Sin­co Kelle­her and Ter­ry Tang head­lined, “Omi­cron ex­plo­sion spurs na­tion­wide break­down of ser­vices” was full of sta­tis­tics from across the Unit­ed States on hos­pi­tal­iza­tion rates or em­ploy­ees call­ing out sick from work. The case count met­ric was not used.

“We def­i­nite­ly want­ed peo­ple to go a lit­tle deep­er and be more spe­cif­ic in re­port­ing,” said Josh Hoffn­er, the news ed­i­tor who helps over­see AP’s virus cov­er­age.

Many news or­ga­ni­za­tions are de­bat­ing how best to use sta­tis­tics now dur­ing the omi­cron surge, Wu said. But there are no easy an­swers.

“It’s how jour­nal­ism works,” Wu said. “We need the da­ta. We need to show re­ceipts to read­ers. But I try to do it care­ful­ly.”

Hos­pi­tal­iza­tion and death rates are con­sid­ered by some to be a more re­li­able pic­ture of COVID-19′s cur­rent im­pact on so­ci­ety. Yet even the use­ful­ness of those num­bers has been called in­to ques­tion in re­cent days. In many cas­es, hos­pi­tal­iza­tions are in­ci­den­tal: there are peo­ple be­ing ad­mit­ted for oth­er rea­sons and are sur­prised to find they test pos­i­tive for COVID, said Tanya Lewis, se­nior ed­i­tor for health and med­i­cine at Sci­en­tif­ic Amer­i­can.

De­spite the im­per­fec­tions, case counts should not be ig­nored, said Gary Schwitzer, a Uni­ver­si­ty of Min­neso­ta School of Pub­lic Health in­struc­tor and pub­lish­er of Health­News­Re­view.org, which mon­i­tors health cov­er­age in the me­dia.

The num­bers il­lus­trate trends, giv­ing a pic­ture of which ar­eas of the coun­try are be­ing hit par­tic­u­lar­ly hard or where the surge may have peaked, he said. They can pre­dict broad­er so­ci­etal im­pacts, like where hos­pi­tals are about to be slammed or where there will be work­er short­ages.

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

That’s a point em­pha­sized in AP’s in­ter­nal guid­ance, as well.

“They do have val­ue,” Hoffn­er said. “We don’t want peo­ple to elim­i­nate men­tion of case counts.”

There are some in pub­lic health and jour­nal­ism who be­lieve the cur­rent surge — painful as it is — may au­gur good news. It could be a sign that COVID-19 is head­ed to­ward be­com­ing an en­dem­ic dis­ease that peo­ple learn to live with, rather than be­ing a dis­rup­tive pan­dem­ic, wrote David Leon­hardt and Ash­ley Wu in The New York Times.

But if the past two years have taught any­thing, it’s about the dan­ger in pre­dic­tions, Lewis said.

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