The city kept its Wednesday. Read the dates anyway.

Five days ago this newspaper reported that Kansas City’s health department had switched on a public dashboard of its communicable disease counts, and that the department had promised to update it every Wednesday. Today was the first Wednesday that promise could be tested.

At 11:24 this morning, it had not been kept. The dashboard’s own “Available Data Range” still ended August 22, and the week the city said this update would add, Sunday August 23 through Saturday August 29, was nowhere in it.

At 1:32 this afternoon, it had been. The range ends August 29.

So the department did what it said it would do, on the day it said it would do it. Anyone who checked before lunch and concluded otherwise would have been wrong, this desk included. The version of this story that ran on the morning capture was written, and then it was thrown away.

What was promised, in the city’s own words

Release 3213, posted August 28: “The dashboard will be updated each Wednesday with case counts from the previous Sunday through Saturday.” Dr. Marvia Jones, the director of health, in the same release: “they want to know that they can trust what is presented to them… we can give them timely, accurate information.”

A named day and a named window is a checkable commitment, which is the only kind worth writing down. It was kept.

Now the part that will mislead you

Here is the summer’s headline number, enteric illness, the category an intestinal parasite like cyclospora falls into, as this newspaper captured it three times in five days:

Not one point of that movement is disease. No cases left Kansas City over the weekend and none arrived this afternoon. What moved was the window.

The dashboard’s date control fills itself in, and what it fills in is a rolling twelve weeks. On August 29 that meant May 31 through August 22. By this morning the start had rolled forward to June 7 while the data still stopped at August 22, so the window was a week shorter and every category fell with it. Animal bites went from 165 to 145. Respiratory went from 182 to 174. Sexually transmitted infections went from 1,645 to 1,526. A resident who had looked on Saturday and looked again this morning would have seen improvement across the board, and would have been reading a calendar rather than an epidemic.

Then this afternoon’s refresh pushed the far end out to August 29, and the numbers climbed back.

The percent-change column is the same trap in miniature, and it is worse, because it moves for two reasons at once. The prior-year comparison slides along with everything else. Enteric’s raw count fell from 382 to 374 while its percent change rose from 263.8 to 302.2. Both numbers were correct on the morning they were captured. Neither was about anyone getting sick.

How to read it without scaring yourself

The dashboard is not hiding any of this. It tells you exactly what it is doing, on a line most people will scroll straight past. Under the date control it prints This Period and Previous Period with their start and end dates spelled out, and beneath those, Available Data Range.

Read those three lines before you read a single count.

If you are comparing what you see today with what you saw last week, the only honest comparison is one where those dates match, and by design, next Wednesday, they will not. Screenshot the dates along with the number, or the number will mean nothing to you a week from now.

There is one more wrinkle worth knowing. At 11:24 this morning the auto-filled control read 6/7/2026 – 8/29/2026 while the panel underneath it reported data only through August 22. The control and the data disagreed by a week. By 1:32 they agreed again. When they disagree, believe the panel.

What is still an open question, and stays one

Three categories show year-over-year drops so large that this newspaper is not going to characterize them until the department explains them:

A fall of that size is not a plausible one-year swing in the real world. Something about how the 2025 side of the comparison was assembled is the likelier explanation, but “likelier” is not reporting. The department has not answered the question, and until it does, these stay questions and do not become findings.

Also still unread: the dashboard’s own Frequently Asked Questions page, which sits behind a button inside the report and almost certainly says whether these are report dates or onset dates, and whether small counts are suppressed. Until somebody reads it, this newspaper is not going to tell you the trailing edge is a lag. It may be the design working exactly as documented.

The record, as of this afternoon

Every figure above comes from three timestamped captures of the city’s own report, archived by this newspaper on August 29 and twice today, at 11:24 a.m. and 1:32 p.m. Anyone can pull the same file. It is a public Power BI report linked from the health department’s data page, free, no login, no request to file.

That matters more than the individual counts do. Three weeks ago, a column in this paper noted that the summer’s cyclospora numbers existed publicly only because a television station had thought to ask for them. “It should not require asking,” it ended. It no longer does.

Next Wednesday the dashboard should gain the week of August 30 through September 5. This desk will check, and will say either way.