Image Credit: Toby Morris @xtotl — featured on The Spinoff
Yesterday, the U.S. experienced our largest daily COVID-19 death toll with over 800 deaths. President Trump rightly implemented another 30 days of social distancing in the face of the grim reality of the virus. Further, the White House Task Force now says we may have up to 240,000 deaths, even if we implement and perfectly follow social distancing guidelines. These projections and other policy decisions are made based on data. So I decided to take a stab at some questions of my own.
To fulfill my love for both politics and science, I’ve been diving through the massive trove of COVID-19 data and analyzing with some Census data. Let me take you through some of the few questions I posed:
Question 1: How’s the U.S. doing compared to Italy and South Korea in regard to COVID-19 testing?
Question 2: Is social distancing actually working?
Question 3: Are Democratic and Republican governors responding differently to COVID-19?
Let me preface this by saying there are a lot of caveats in the data I’m about to show. This was a very quick and dirty analysis and I welcome any discussion or thoughts anyone would like to share.
I utilized the Institute of Health Metrics and Evaluation (IHME) data sets which are open and free to the public. They are awesome. I highly recommend taking some time to check them out. Another website I used for this analyses was the Worldometer real-time COVID-19 tracking website (also fantastic), and some data sets from Statistica. All of the data analyzed were through March 31, 2020.
Alright, let’s go.
Question 1: How does the COVID-19 testing capacity of the U.S. compare to Italy and South Korea?
The US has tested the most people than any other country in the world – 1,019,794. BUT, when we measure how many tests we have completed per capita, both South Korea and Italy have ran more COVID-19 tests than us. In fact, they have performed over two times as many tests as us, at rates of 2.25 and 2.75 to the U.S., respectively.
In order to continue to understand where the next New York State caseload (>70,000!) is going to be in the US, we have to keep increasing our testing capacity. For some reason, there are a lot of issues related to testing here in the states. There’s testing kit shortage, we lack appropriate PPE, should we test asymptomatic cases, etc. Hopefully, the new 5 minute Abbott Labs COVID-19 test will really help us out.
Question 2: Are social distancing measures having an effect on the COVID-19 curve?
I sorted “social distancing measures” by School Closures, Non-Essential Work Closures, or Stay at Home orders. The majority of states have instituted at least two-thirds of these measures. 18 have yet to impose more than one.
If we break down the numbers real quick: As of 3/31/2020 and according to IHME data, 45 states + DC have imposed school closures, 29 states + DC have shut down non-essential businesses, and 27 + DC have ordered stay-at-home measures. So the obvious question becomes: Is there any effect on the curve between states that have or have not implemented social distancing measures?
Figure 3 illustrates that states with social distancing measures (regardless of which one) have a projected COVID-19 death curve to peak 5-7 days earlier than states that have not implemented the respective social measures!
We can infer a couple of things from these data, and I want to remind us to think about this on a state-by-state basis: 1) Social distancing measures reduce the time it takes to reach the apex of the curve, but 2) may not necessarily reduce the curves overall height (increased death rate at the peak).
Essentially, some states may have more/less deaths than others regardless of social distancing. But states that chose to enforce these measures compared to others that do not might prevent more deaths had they done nothing at all. Additionally, they may recover from the outbreak sooner than later. This would allow their state to “reopen” earlier as well.
(For any stats buffs, these data are significant by 2-way ANOVA (p<0.005) based on column factor (states that have vs. have not implemented social measures)).
Question 3: Are Democratic and Republican governors responding differently to COVID-19?
Let’s see how these data correlate in states between Democratic and Republican governors. We know some states have, while others have not implemented social distancing measures across the country. When we break this down by their Governor’s political party affiliation, what do we see?
In the states that have imposed such measures, it looks like both parties did so around the same time (Figure 4). Democratic Governors maybe did so a day or two before Republican Governors. I just became aware of some research that studied this. Not surprisingly, the date these measures were enforced may have had a modest effect on preventing community spread and transmission. Time will tell as we learn more about the outbreak.
‘Ello Governor
What really stood out in the data, is that there were many more states run by Democratic Governors that have issued social distancing measures than the proportion of states run by Republican Governors (Figure 5). Further, when we dive deeper into the potential ramifications of this, we see that the projected COVID-19 death peak for states lead by Republican Governors is nearly 3 days later than states with Democratic Governors (Figure 6).
What does this mean? First off, this does not mean that states with Republican Governors are going to have more COVID-19 deaths. What it could mean, is that a delayed curve could result in deaths that may have been prevented had social distancing measures been in place (Similar to Figure 3). However, this too needs to be taken with a grain of salt. For instance, the peak could be delayed simply because the first positive cases may have arrived later than states with Democratic Governors.
I find this reason to be fleeting though too. First off, nearly every state and Washington, D.C. in the country (Democratic or Republican, 46 in total – see Figure 5) closed down schools at roughly the same time (Figure 4, ~March 17th) to slow down potential community transmission. With the botched testing rollout, and the more likely scenario that community transmission was occurring right beneath our noses, I think it is more reasonable to imagine that the delay in the projected death peak is due to the lack of control mitigations. My hypothesis, is that these projections are taking into consideration the current caseload in these states that are not practicing the necessary social distancing measures that would flatten and minimize the curve all together to most effectively reduce the total number of deaths — Dr. Fauci and Dr. Birx of the White House Task Force actually discussed this at their latest briefing.
Final Thoughts
This post is not intended to be partisan. I’m simply analyzing the data freely available at my disposal. I think these data are very relevant in making an impact on COVID-19 cases and death rates. The data allow us to make all different kinds of comparisons when we apply U.S. Census data to it as well. We can compare the responses between states that voted Democratic or Republican in the 2016 Presidential election (very interesting), or see how states with increased risk of obesity or heart disease are being impacted by COVID-19 on a per capita basis. The data presented today is just one sliver of it all. It will be interesting to see when this is over how states varied in their social distancing orders, and how that affects caseload and mortality.
As such, my final recommendations are below:
#1: We still need to get our testing capacity up. Sure, we’ve tested the most people absolute-wise (Figure 1), but for our standards in the U.S.A. we need to do more (See Figure 2). More testing means more information, and better informed policy decisions.
#2. Social distancing measures work (Figures 3-6). Other states that have not begun implementing these measures should do so immediately. Stated best by Seattle Mayor Jenny Durkan on CNN (3/31/2020): “Just because you haven’t seen the virus yet, it’s coming. If you take steps now to prepare, you can help all of America. Every day you wait you risk lives.”
#3. Yes, we need to prioritize flattening the curve, but we must try to minimize the curve all together! Remember, the longer the curve stretches, the later we have to wait to reopen the country! Although reducing the curve’s apex is crucial in order ease the burden on our healthcare infrastructure, we have to chip away at total area of the bell curve too. We are probably going to experience the outbreak across the country in waves. New York and New Jersey are predicted to reach their peak in just over one week. After that, it could move to the DMV, Louisiana, or Michigan. Two weeks ago Michigan had 100 cases. Now they have over 7000. Let’s not get caught flat-footed.
#4. There are no “sides” during a pandemic. Just one team (See above quote by Mayor Durkan). It’s not about Democrats and Republicans. Time to ELE baby (Everybody Love Everybody).
#5. Let’s take time to appreciate and thank all of the first responders, doctors, nurses (shoutout to ICU nurses Kristi and Biri), military, essential workers, and everyone still working in the grocery stores, transportation, the Post Office, and the public. Y’all are the real heroes.