State of the States – Week 2

Posted by

Today is just a weir day. I think it just takes a bit for it to sink in when you lose a pet. Did some clean up but I still find myself looking at the places Rocco used to be – almost expecting him to be there. So, this trip into the numbers is a distraction.

So I’m not entirely happy with my set of numbers but I said I’d work with them for a bit. I think as a trending tool, it probably isn’t bad but there are some flaws I might have to deal with.

To start, a look at the national picture. For the week that ended yesterday, new positive cases were down again. The drop was not as large as the week before but it still represents the fifth week in a row where new positive cases were lower. We are down almost 40% from the peak back in mid July so this is not a trivial change.

The number of newly reported deaths also dropped again which should be happening given the drop in cases. I don’t have hospitalization but those move in tandem. Cases rise first, then hospitalizations and finally deaths. They drop in the same order and a decline in deaths is further confirmation that we are on the downside of the second peak.

The number of tests dropped slightly but is still quite high. I am coming around to the idea that we are doing far too much testing and “finding” things that aren’t there. There were almost 4.9 million tests done last week. At that level, false positives become significant. Now this isn’t going to be exact because I’d have to estimate prevalence and take out the true positives but even assuming a false positive rate of 1%, it is not unreasonable that 50000 positives are actually false positives. That’s about 17% of the total number of positives.

In my work life, I did have experience with diagnostic tests and widespread testing is not always the best thing. Tests aren’t perfect and people running tests are far from perfect. At some point, widespread testing becomes counter productive because you can end up in a situation where a positive test is more likely to be a false positive than a true positive. Again, this is not implying that is the case here but comes from experience with the numbers behind diagnostic testing.

A brief recap if you didn’t see the last post – I’m basically looking at the current number of cases relative to the peak and population. I’m using a ten day average to smooth out daily trends in reporting by the various states. I arbitrarily set scales to give each variable a score of 1 to 10 and added the three together to get a final score for each state.

I kept the scales the same for this week so I could track progress. Compared to last week, 37 states saw improvement from last week, 3 are the same and 11 were worse. Forget that I’ve got DC in this mix so you get to 51. Most of the changes are small. The biggest movers in the right direction were Delaware, California, Texas and Nevada. The biggest movers in the wrong direction was South Dakota.

One last note – I have allowed a state to drop to 0 on one of the scales if the result fell below the minimum set last week. In a perfect world, a state with no ‘rona cases would then score 0 which makes sense to me.

Alright, lets do some states. Well, one last disclaimer, which is the same one I made last week, this is only looking at where they are now relative to where they were. It is not at all intended to be any kind of judgement on how the state handled the pandemic as a whole.

We start with New York at 2 points followed by New Jersey and Connecticut with 3 points. Next are Massachusetts and New Hampshire with 4 points. As mentioned last week (and with the exception of New Hampshire), these were the hardest hit areas back in April. They all still have cases but at low numbers relative to population and the April peak. I kind of wonder if we aren’t just at some floor of cases that we’ll never get below due to false positives. Given that these states have been flat for a couple of months, I think that may be happening.

With 5 points we have Arizona and DC. Other states improved more but Arizona was right up there. Highest than some of others on this part of the list in terms of current cases but way, way down from early this summer when it was in the news every day.

Vermont at 6 points and Rhode Island with 7 come next. Slightly different situations. Vermont just has a low overall number of cases while Rhode Island is down more from the peak.

Pennsylvania, New Mexico and Maine all have 8 points. Maine is another place with a low number of cases.

In this first group, all showed improvement from last week except Connecticut which stayed the same.

Moving down the list, we have Delaware, Louisiana and Colorado with 9 points. Should probably talk about Delaware because I’m not sure how real this improvement is. The average number of cases is down but they had an unusually large number that was included last week but is now more than 10 days away and doesn’t make the cut. Like they were clearing the books and dumped a bunch of old test results on a single day. We’ll have to see how things pan out going forward. Louisiana is still a little high on current cases but continues to trend in the right direction.

Closing out the next group we have Washington with 10 points and Oregon and Maryland with 11. All three are down about 50% from their peak and have low number of current cases relative to population.

So the areas that are best off now are the northeast, the pacific northwest and the southwest.

Starting the middle group at 12 points are Florida, Michigan and Utah. Florida was another big mover. Still a higher than average number of cases per day but way, way down from the peak. Moving in the same direction as Arizona and Louisiana.

With 13 points we have California, Wyoming, Nebraska and South Carolina. In terms of the raw number of cases, California is still on top but there are also more people there than anywhere else which is why I do this relative to population. Based on that, the current case load is about average and they are down about 40% from the peak. Some of the change is because a day with a big number of cases fell off the average but the general trends are good. Wyoming has a small number of cases but is moving farther away from their peak so they show up on the list of states with big changes. South Carolina is similar to the other previously mentioned southern states with cases dropping quite a bit from the peak.

To close out this mid group we have Texas, Nevada and Ohio with 14 points. Like the other southern/sun belt states, the story in Texas and Nevada is that the number of new cases continues to drop even if the current count of cases remains above average. But I’d expect these two to continue dropping. Ohio didn’t move much from last week.

Now we move into the next group and state with Idaho and West Virginia at 15 points. Idaho had a real spike at the end of July and is dropping from that. West Virginia didn’t seem to have much of a spike so they are still close to their peak but the case number low which helps. Idaho had a big improvement from last week as well.

With 16 points, we have Alaska and Virginia. This group is certainly more random than the group that was basically the Northeast. A couple of states that peaked in early August and having moved too far off that peak.

Next we have Wisconsin and Illinois with 17 points. Illinois had a little second wave that didn’t overtake the first but got close and they haven’t moved that far off of it. Wisconsin peaked at the end of July but is dropping slower than other places. Both showed small improvements from last week.

Closing out the group, we have a cluster of five states at 18 points. Mississippi and Oklahoma were part of that second Sun Belt wave and peaked at the end of July but still have a high number of cases relative to the population. Tennessee and North Carolina are kind of part of the second wave but kind of not. Tennessee is further from the peak but North Carolina is better in terms of cases per population. The oddball in the group is Minnesota which is like other places in the Midwest seeing kind of a silent second peak. Like Illinois, they didn’t quite catch the first peak but they are very close to it and could end up passing it.

In that group, Minnesota and North Carolina are two places where things are slightly worse this week than last week. Both due to increasing case counts.

We start the last group with another five state cluster at 19 points. Again, as I said last week, differences of a couple of points aren’t meaningful. It is just how they fell relative to the scales. But ranking is fun so I’m doing it.

We have Alabama, Arkansas and Georgia which are the Sun Belt laggards. They both peaked at the end of July but haven’t dropped as fast or as much as other places in the region. Missouri kind of fits in that Tennessee/North Carolina group although Missouri and Georgia are slightly improved from last week while Alabama and Arkansas are slightly worse. And our Midwest representative in this group is Indiana. Currently it sits right at the peak. The only thing that saves it from a worse score is that the case count relative to population is about average. It is potentially an artifact because there was a big case dump on the 24th that will wash out in time. Still the trend there is flat and not decreasing.

Moving along we have Kansas, Kentucky and Montana at 20 points. All three have higher scores than they did last week. Kentucky and Montana have established new peaks and Kansas is just a couple of days away from its peak. Looking at graphs of all three places, they are not seeing a substantial increase in cases. All three were part of the stealth increase in July when everyone was focusing on Florida, Arizona and Texas. Just that they’ve been flat now for over a month. All three were low in terms of cases per population so this may run for a bit.

Next, we have Iowa with 24 points. Iowa had basically been flat for a month but there were two days with a huge dump of new cases this past week. I don’t know if there was some trigger for more testing but it looks more artificial than an increasing trend. It does mean that Iowa is currently sitting at peak and will likely continue to be close to peak until that dump gets out of the average.

At 26 points, we have South Dakota which was the biggest mover in the wrong direction. It also looks like Iowa. The last three days have seen large number of new cases which has driven the average up to put them at peak. The case count can rise exponentially but you don’t go from 63 to 623 in one day. Not with a state that doesn’t have a lot of people. Like Iowa, we’ll know more over the next couple of days but the average is going to remain high until we get past this bump.

At the tail end, we have North Dakota and Hawaii with 28 points. Both states are also at peak. With North Dakota, this looks like a real spike because we’ve got several points in a row above the 10 day average and things appear to be in an exponential growth phase. Which means the average probably goes higher. Relative to the population, North Dakota also has the highest number of cases. With fewer people, it may reach burn out quicker.

Hawaii was mentioned last week. The growth in new cases has flattened but they are still at their peak. Getting out of the exponential growth part is good but history with other places suggests they’ll be at this level for a bit.

Kind of odd that North Dakota and Hawaii end up together. They look to be the really problematic places but we’ll have to watch a couple of the others to see.

I’ll just close with another reminder that we are only looking at cases here. It is just a measure of where we stand now and not how serious things might be in the various places. Still, the picture looks a lot better now than it did a month ago.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.