Two people are dead in Michigan, and I want that sentence to stay at the top of everything I write this week. The Michigan Department of Health and Human Services announced this morning that two deaths have been identified as part of the cyclosporiasis outbreak affecting the state, that both individuals had significant underlying health conditions that may have been impacted by cyclosporiasis and dehydration. Those are the first deaths anywhere in this outbreak.

This has been a moving target all day. Friday night I published 21,452. This morning, with Michigan’s new figure and the deaths, I published 21,913. This afternoon, after going back through all fifty states one at a time, I published 22,181. Tonight it is 22,683. Michigan is at 11,234 with 193 hospitalizations — but the largest single move of the day did not come from Michigan, and it did not come from any state I was watching.

Missouri has 1,095 cases. The state Department of Health and Senior Services publishes a communicable disease report every Sunday, and the August 2 report put the state at 1,095, against fewer than 600 a week earlier. Boone County alone has 90. The three-week series is 216 on July 19, 593 on July 26, and 1,095 on August 2 — roughly doubling each week for three weeks running. Missouri normally logs about a dozen cases a year. This is something on the order of ninety times normal, in a state that is not part of the federal outbreak.

Now the part that ought to stop people. A CDC spokesperson told KCTV in Kansas City that cases in the remaining thirty-two states, Missouri included, are classified as unrelated to this outbreak and are being investigated as separate clusters driven by different foods. In the same news cycle, St. Louis County health officials said that people in their area who got sick reported eating at Taco Bell.

Set those two statements beside each other and hold them there. Missouri received the recalled Taylor Farms lettuce — it is on FDA’s own distribution list. Missouri has 1,095 cases. A Missouri county health department says its patients report the same restaurant chain that defines the nine-state outbreak. And the federal position is that none of it is related, driven instead by different foods that nobody has named, in clusters nobody has described.

Both statements can be true. It is entirely possible that St. Louis County residents ate at Taco Bell because a great many people eat at Taco Bell, and that their illnesses trace somewhere else. That is exactly the kind of thing an epidemiologist rules in or out with a case-control study, and it is the kind of thing a health department publishes when it has done one. What is not acceptable is asserting the negative without showing the work. Somebody at CDC has decided that 1,095 Missourians are not part of this outbreak. I would like to know what that determination rests on, how many of those 1,095 were interviewed, how many reported Taco Bell, and what the different food is.

The nine-state outbreak is not a description of who got sick. It is a description of who got interviewed in a particular way. The entry ticket has always been documented Taco Bell exposure captured through the federal case definition, which is why Michigan’s 11,234 produces a federal count of 1,947 and why Alabama’s two restaurant clusters and New Hampshire’s Cheshire cases sit outside it. Missouri now looks like the largest example yet of a state on the outside with a county pointing straight at the inside.

The two deaths belong in this same argument, because there is an unanswered question about which set of numbers, they sit in. CDC’s investigation page says it is aware of two cyclosporiasis-related deaths in people with underlying health conditions from Michigan as part of this outbreak, and that it will update its website to include laboratory confirmed cases, hospitalizations, and deaths. On that page, this outbreak means the nine-state cluster of 1,947 people. Michigan says the deaths are part of the cyclosporiasis outbreak affecting Michigan, which is 11,234 people. Those denominators differ by a factor of about six, and no agency has said which one applies. The FDA advisory is still stamped July 24 and still reads Deaths: 0.

The corrected arithmetic: 22,683 across fifty states, against the 18,207 the federal government describes when its two published figures are added together. The two counts are 4,476 apart, which is 24.6 percent of the federal base — the gap has crossed a quarter for the first time. The nine-state outbreak, at 1,947 people, is 8.58 percent of the illness this country is reporting.

One more thing about today that I think matters more than the total. The count rose 1,231 since Friday, and only 461 of that was a state filing a genuinely new daily report. Iowa at 226 rather than 57, Wisconsin at 160rather than 61, and Missouri at 1,095 rather than 593. I found those by going and reading what the states publish instead of what a national roundup said they published three weeks ago. Nobody got sick today because I looked. They were already sick, and the number simply had not traveled.

Indiana stays at 1,176 here. A national outlet reported this afternoon that Indiana has recorded nearly 1,349, and it may well be right — the state posts a new dashboard figure every weekday by one o’clock. But no Indiana newsroom carries that number, and WANE, 21Alive and Indiana Public Radio have written up every previous state figure as it landed. 

Where each state stands tonight, with the source for every number.

Thirteen states in that table publish no count of their own. Eleven more still trace to a national roundup last updated on July 17. CDC says 45 states have cases.

Twenty-eight days remain in the season CDC defines as running from May 1 through August 31. Two families in Michigan are already on the other side of this. If a state can double to 1,095 cases while the federal government classifies every one of them as something else, then the number worth arguing about is not the total at the bottom of my table.