A month ago, Governor Abbott announced that Texas would start “re-opening” on May 1. Restaurants opened to 25% capacity, those that could do so profitably. Many could not and remained closed. Retail stores also re-opened under similar restrictions. State parks re-opened with restrictions. On May 18, gyms and salons were allowed to re-open, again with significant restrictions. This past Friday, May 22 at 12:01am, bars were allowed to open with severe guidelines. Water parks will soon re-open. Schools can have in person summer school. Some summer camps can re-open. It’s not quite business as usual, but it’s moving in that direction, for those businesses still alive. Will this be a perfect storm or will it just be a drizzle?
The state is using various indicators, such as number of hospitalizations, to decide if we’re on track. We’ll get to that soon. First, let’s look at the national picture.
Below you see the number of cases reported per day since March 15. While we saw a steep increase in March and into April, with a peak of over 33,000 cases in one day, by late April we’ve begun to see a steady decline in daily reported cases. The red line shows the seven day average trend. So, that’s good, right?
It’s good if our rate of testing is increasing. One of the first rules of reviewing epi data is to look for a reason why. Are cases dropping because fewer people are getting sick? Are we testing less? Not really. Across the country, we’ve continued to increase the number of tests performed per day (or at least the number of test results received per day). So this is hopeful. Continuing targeted testing, along with tracing and isolation will be essential to preventing a second wave.
One of the challenges with epidemiological surveillance of these data is that unless a person knows they were exposed, they’re not likely to come in for testing until a few days after symptoms occur. Given the potential 14 day incubation period, that means they could have been exposed 2 weeks prior, though the average length of time from exposure to symptoms is five days. While labs are turning around much faster, there still can be anywhere from a few days to a week before results come in. So any reporting reflects exposure more than two weeks ago, at least.
It appears that for the past week, the number of daily cases has been declining. This may reflect a decline in transmission over the past two weeks, but that won’t really be clear until 1–2 weeks from now.
We also know that the spike in cases in mid-May is due to an outbreak in a meat packing plant. While the conditions in the plant lend itself to high transmission, the workers are not confined to the plant. They go home, they go shopping, they can expose others in the community. So while we may be able to trace exposures back to this source, transmission is still happening in the community.
We said that Texas is relying upon hospitalization data. The goal of the closures was to “flatten the curve” so that healthcare systems would not get overwhelmed. We needed to slow down the daily cases, in particular the severe cases needing hospitalization, so hospitals could adequately respond to every patient.
Hospitals have reported they have not been overwhelmed, but again, this is with all elective procedures stopped. The hospitalization curve appears to be flat, but it’s not really declining. We need to watch this very closely over the next two weeks.
Another indicator of severity is the Case Fatality Rate (CFR). Though the CFR is higher than it was in March, the number of deaths in Texas is beginning to decline. While we have been “open” for a few weeks, most of the individuals who’ve died were infected in April or before.
Bexar County and San Antonio
San Antonio Metro Health District has been playing with different ways of presenting the data. Date of report is problematic if reports aren’t received for 1–2 weeks. Date of onset is better, because it tells us when someone got sick. Only problem is that requires going back into the data to add information when it’s received, which means that data for the week prior will always be incomplete. This means that the rate will always appear to be falling. It also doesn’t account for asymptomatic individuals. Date of test will be useful, though again we need to look at the trends up to at least a week or more prior.
This graph show the data by date of onset, until two weeks ago. We can be confident that these data are up to date and accurate. Again, we look at the 7-day average. After a peak in April, we seem to have dropped. The first two weeks of May stayed fairly constant, but remember, despite re-opening, these cases were likely exposed before businesses opened. The real test will be to see these data in another 2 weeks, but it’s hopeful
Another positive note, despite more than 300 cases in the jail, for the past few days, no new cases have been identified among the inmates. This doesn’t mean the outbreak is over. We’ll need to continue constant surveillance.
Flatten the Curve
Is the curve really flattened? This shows the cases and deaths on a logarithmic scale. It looks reasonably flat, or at least on the right track.
Will it stay flat? That will required constant monitoring of potential cases, outreach to find high risk cases, contact tracing and isolation. As long as we have cases, we can’t fully relax.
State by State comparisons
The CFR is higher in the next most populous states (below by rank), but that may be due to testing of higher risk cases. This also shows in the positivity rate of testing. A lower positivity rate may be good, but also may mean we’re testing more people without symptoms. While it’s helpful to get an idea of infection across the population, we could end up with a sampling biased towards uninfected people. It’s important to continue targeted testing.
San Antonio and Texas are behind other states in testing, with less than 3,000 tests per 100,000 in the population.
I’ve talked about herd immunity before, and it still holds. Herd immunity based on natural infection is simply not a viable option when less than 1% are currently infected. That low infection rate may make it seem like COVID-19 isn’t a problem, except when you see we have nearly 100,000 deaths in the US in a 3-month period, and that’s with control measures.
Opening businesses is an essential element to health: people need an income and the US lacks extensive social services to make up that loss. We must continue to do so with caution. Our relatively stable numbers can quickly spiral out of control.
edit: If you’d like to see a beautiful display of data, see Dallas’s weekly report here.