Indications Social Distancing and Staying at Home Are Working
Comparing deaths and hospitalizations that have occurred with the projections suggests that some states are flattening the curve.
The overall projections for the US death toll from COVID-19 keep getting smaller. This is great news and does not mean that the original models were wrong. They were correct based on the information available at the time. What has changed is information about the spread of the virus and how effective the behavioral changes have been in managing the spread so that hospitals in many places have not been overwhelmed. As new data and information became available the models were updated to reflect these. The Institute for Health Metrics and Evaluation (IHME) models have been updated 6 times as of April 9, 2020.
As a citizen of Maryland, I have been monitoring the predictions for my state along with the actual data reported for Maryland for COVID-19. I have focused my attention on deaths and hospitalizations, because most other data are so incomplete or inaccurate that I don’t consider them reliable. What I have seen is encouraging and suggests that the steps Governor Hogan and the citizens of Maryland have taken are working to slow the spread of COVID-19.
I consider the official start of social distancing in Maryland to be when schools were closed on March 16, 2020. In addition to closing the schools on March 16, steps that Governor Hogan took to encourage social distancing included closing non-essential services on March 23 and issuing a stay-at-home order on March 30.
Several places have developed models to predict the outcomes based on effective social distancing and stay at home efforts. Two that I have used are the ones from (IHME) and Covid Act Now. What I like about the IHME models are that they include the uncertainty range. This means that they report both the most likely values and the upper and lower predictions from their models. Covid Act Now presents only one set of values.
A difficult aspect to understanding the output of the models is that the results change over time. Sometimes the models change, meaning that the equations for calculating the results are modified based on new information. Sometimes, the results are updated using new input data but calculated using the same equations as were used before. This means that on any given day, the graphs can look different if they have been updated since previously viewed.
One change that was readily obvious to me is that the peak of infection for Maryland keeps moving later in time. This is an indication that the interventions are working to reduce the speed that the virus is moving through the population and is exactly what would be expected if the interventions are working. This is evident by comparing the predictions from the Covid Act Now model for March 27 and April 7 (Figure 1). The peak moved from April 14 to April 20 in the limited action scenario and from April 28 to May 10 in the social distancing (referred to as lax stay at home in the April 7 version). These predicted dates of hospital overload reflect peaks of infection.Another piece of encouraging data is the reduction in estimated deaths from March 27 to April 7 even in the limited action scenario. The increase in deaths in the 3 months stay at home scenario likely reflect the increase in total percentage of Marylanders that are predicted to become infected in the results from April 7.
The IHME and Covid Act Now models are different, so the predicted results are different. The data from IHME are presented in an interactive format, so that each day’s data can be inspected (Figure 2). With the IHME model, the predicted values compared to the actual values are very encouraging. The predicted total COVID-19 death toll for Maryland for April 7 was 142 with a range of 100 to 210. The reported deaths as of April 7 are 103. Similarly, the predicted death totals for April 9 were 161 (range of 129–213); whereas the actual deaths were 138. This suggests that Maryland is on track to minimize the deaths from COVID-19, and the steps implemented have been effective.
I find the data encouraging, despite the wide variations in predicted outcomes from these two sources. Marylanders should continue following the stay-at-home order, practice social distancing when in public places with other people, and wash hands frequently. People in other states and countries should be similarly encouraged by how effective these measures appear to be and should strive to implement them or continue following them.
Also of interest
N. R. Gough, Social Distancing Key to Slowing COVID-19 Spread