Top news, reports and insights for today:
- Daily headline summaries for WHATDAY:
- New study of hospitalization trends in 4 states finds that stay-at-home orders cut the expected rate of COVID-19 hospitalizations by as much as half days after they were implemented. This study shows the advantages of looking at hospitalization data instead of cases and deaths (JAMA)
- New genetic analyses suggest that SARS-CoV-2 may have acquired the ability to infect human cells after a bat coronavirus and a pangolin coronavirus recombined to form the new pathogen (ScienceAdvances)
- New study of prescription data from all 50 states finds that compared to pre-pandemic, prescriptions for the discredited anti-malaria drug hydroxychloroquine/chloroquine rose 1,977% in mid-March. Overall there were 483,000 excess fills for these medications during a 10-week period compared to 2019 (JAMA)
- During a WHO news conference today, experts said there is no evidence to suggest the SARS-CoV-2 virus has lost any of its transmissibility or virulence, a belief that some have suggested as the rate of infections declines in many nations (Reuters)
- A look at rates of cases and deaths after 3 months
As May comes to a close, I found myself going back to what epidemiologists love most: rates. Looking at numbers of deaths and cases by place can be misleading given how widely divergent those places are in their population sizes, densities and distributions. So where are we in terms of states and regions when looking at rates? The top graph shows cumulative cases per 100,000 population as of May 31. This angle makes clear how much the transmission intensity has been distinctively amplified in the northeast. The overall incidence rate in that region, at 1,312 per 100,000 is 4 times higher than the West or South and almost three times higher than the Midwest. The top 7 states are all in the northeast. My own state of Maryland has a higher case incidence than Louisiana and nearly as high at 873 per 100,000 as Illinois, the state with the highest rate outside the northeast. New York and New Jersey had double the rate of Illinois. Nebraska, Iowa, Michigan and South Dakota all have rates higher than the overall U.S. case incidence of 541.
The bottom figure tells the story of COVID-19 mortality rates by state and region. The mortality story is even more extreme. The Northeast has a mortality rate of 88 per 100,000, 6-fold greater than the West and South and more than 4-fold greater than the midwest. The highest mortality rate has been in New Jersey (132 per 100,000), which is two and a half times greater than Illinois and Louisiana. Both New York and New Jersey have had 10 times higher mortality than California and more than 9 times higher than Washington State. For reasons that aren’t clear, Connecticut and Michigan have mortality rates higher than their case incidence would suggest. On the other hand, given their rank in cases, Utah, Nebraska, South Dakota, and Delaware have lower mortality rates than expected given their case incidence rates.
Bottom line: There is no single U.S. epidemic. States and regions vary enormously in the transmission intensity up till now. The northeast has overwhelmingly been hit hardest. The West has been spared more than many think. Some states have over- or under-performed at keeping COVID-19 patients alive.
- COVID-19 hospitalizations: a more reliable indicator or epidemic trends?
I recently came across a couple of articles mentioning a data initiative from the University of Minnesota’s Carlson School of Management to track the rate of change in the epidemic by monitoring coronavirus hospitalizations in a centralized and comprehensive way. I applaud this effort, launched in late March, because it has long seemed to me that we have way to many fancy models and not enough reliable basic data. Mostly, we track the ups and downs of this epidemic using reported lab-confirmed cases or deaths related to those cases. The problem (and it’s a really big problem) is that both these data streams are seriously distorted. Because the testing is limited, selective and prone to false results, our ability to track cases using that testing is severely limited. Are we capturing half the true cases? Relying on testing alone leaves us largely in the dark. The best available serology data tells us there are between 5 and 20 times more cases in the community than our testing data reveals. Mortality data is also highly problematic due to long time lags in the data, differences in how dates are recorded, and the fact that people who die at home or in nursing homes are almost never tested. That’s why I have increasingly been searching for good data on daily hospital admissions. People who are sick enough to go to the hospital and be admitted are quite possibly the best, most timely and accurate window into how the true transmission dynamics are playing out in communities. That is at least true for the subset of cases who experience serious symptoms. That’s why this new initiative, called the COVID-19 Hospitalization Tracking Project is so important and timely. You might notice that this is where the data came from that was featured in the third top story of today. The figure below is a screen grab from the website today showing recent trends in hospitalizations for COVID-19 per 100,000 adults for 5 states that have recently seen substantial new daily cases (see Saturday’s briefing). It shows that hospitalization rates are very high in Washington DC relative to the other states, almost double what they are in Maryland. It also shows that despite large reported new cases in all these states, hospitalization rates are generally falling, fastest in DC and slowest in Iowa.
Bottom line: These data tell a different story than the case data. If a big surge of new infections does occur over the next few weeks due to reopening or the ongoing demonstrations, we are likely to see it in these data first and most clearly.