Thread by @meganmesserly

I've gotten a lot of questions over the past couple of days about why the Southern Nevada Health District's data looks different than the data we report and I just wanted to do a brief little thread on it... 1/?

For starters, our data is statewide and, yes, Clark is a lot of the state, but that's one big difference worth getting out of the way first. 2/?

Let's look at the cases graphs. Our graph reflects the day that new cases were reported publicly, whether by a county or the state on behalf of a county. The Southern Nevada Health District tracks by symptom onset date. Why is that an important distinction? 3/?

Wait, first: SNHD's preference is to report the symptom onset date (obtained through investigation.) If they don't have that they'll use, in order: diagnosis date, lab collection date, the date the case was reported to SNHD or the date the case was entered into SNHD's system. 4/?

So if we look at SNHD's graph, it shows 73 cases from July 16, two days ago. That number is only low because we don't have a full picture of what the July 16 cases look like yet. Two days is too recent. 5/?

Say you got sick on Thursday. Maybe you took a day to decide if you should get tested. Maybe you couldn't get an appointment for a few days. Maybe it took time for the results to come back. Most people who started feeling sick on Thursday won't have results yet. 6/?

So the decline in that graph isn't a real decline. It's a decline that will always be built in by the virtue of the way it's reported. It's not a bad graph to look at, you just need to look at it in context of what it's showing. 7/?

SNHD also reports hospitalizations differently than the state does. This is because the health district reports the dates of admission. But there's also a lag with this data. Compare these reports from July 5 vs. July 18. 8/?

The July 5 report shows four hospital admissions on July 1, and a seven-day moving average on that day of of 8.0.

The July 18 report shows 22 hospital admissions on July 1, and a seven-day moving average on that day of 31.7. 9/?

All this to say that you just have to look at the data closely to see what you're really looking at and how complete the data it relies on is.

Anyways, #nvleg is doing things now (finally) so more COVID news later. 10/10