What Statistics say about the ‘United’ States of COVID
The Feds left it up to the states. Some missed the window.
Note: This piece highlights key findings from a broader statistical analysis. For the full analysis & methodology, please see the original article: Statistical Answers to Your Covid-19 Questions.
In the absence of a national protocol around COVID-19, individual states were forced to enact their own restrictions and devise their own plans for lifting them. Predictions on impact vary from state to state, depending on their government’s response.
COVID-19 is a systemic risk for America. It has not, and will not, affect us all equally.
Restrictions put in place this spring and early summer provided relief for overwhelmed ICUs and allowed the COVID-19 mortality rate to decline. Some states, such as Illinois and New Mexico, began lifting restrictions contingent upon continuous review of new infections rates, with laudable success to date. In stark contrast, reopening plans published in late April by states like Texas and Florida lead with charts highlighting the elderly and nursing home residents as those most at risk; downward trends in ‘flu-like illness,’ in the case of Florida’s plan, are emphasized before introducing phased reopen plans. This approach, despite reminding citizens to wash their hands and social distance, serves to reinforce the idea to young adults, teenagers, and children that they are not at risk.
A Look at Statistics
While the fruits of relaxing restrictions are currently being born, it is apparent that a large proportion of cases and resulting preventable deaths observed during this first wave, unequivocally, were the consequence of failing to swiftly implement critical public health safeguards such as stay at home (SAH) orders and the shuttering of non-essential businesses. Such safeguards are known as non-pharmaceutical interventions (NPIs) and they are currently our best defense against the spread of COVID-19. Many epidemiological studies have predicted the growth of new cases and deaths at varying levels of granularity across the United States (LANL, Stanford, Imperial College of London, University of Washington); likewise, many articles have been published around the efficacy of NPIs in reducing COVID-19’s transmission rate. Our contribution is to approach both of these topics from a background in probabilistic statistical theory, applying a different lens than the tools typically used to analyze this type of data.
Key Findings
- Human behavior significantly influences rates of new cases and consequently, preventable deaths due to community spread of COVID-19.
- Forty-five percent of states are now at high risk of a second surge in new Coronavirus cases, or are already experiencing a second surge.
- States at a heightened risk of a second surge failed to enact SAH orders quickly (if at all), mandated shorter shelter in place periods, and lifted restrictions sooner than states currently at a low risk of a second surge.
- Key predictors of both new cases and new deaths are the quick application of NPIs; enacting SAH orders soon after the first case appeared and a longer duration of SAH period significantly reduced both new cases and new deaths.
- New deaths lag new cases, but by how long varies state-to-state. Two distributions formed around a lag of 13 and 30 days, respectively.
- Timely response at the national level is critical. While some states have acted responsibly, others have not and when those residents travel out of state, they put citizens in states with low case rates at risk. Failing to immediately re-instate public health interventions across states identified as most at risk will allow new cases to grow exponentially, quickly overwhelming local and regional health care infrastructure.
- States at low risk of a second surge observe the overwhelming majority (94%) of new cases over the past 14 days at or less than 10 per 100,000 residents. States with middling risk count 68% of cases over the past two weeks at less than 10 per 100,000, and just 42% of cases in the high risk group are at or below 10 per 100,000 over the same period.
Discussion
Given the precarious state of the United States’ health care ‘system,’ now more than ever is the time to proceed with caution to ensure that cases do not increase exponentially. In the absence of a national mandate, states high at risk for experiencing a second surge of cases should re-instate lifted restrictions immediately; masking, social distancing, and mandatory 14-day quarantine after out-of-state travel and for visitors included. Given that standalone temperature controls and contact tracing seem to be ineffective at containing COVID-19, test rates need to dramatically increase. Accurate, responsible, timely collection and reporting of COVID-19 related cases and deaths is paramount: we cannot combat what we can’t track.
We can’t properly prepare or respond without being informed; planning to react is planning to fail.
