Some COVID-19 predictions

It has been six days since my last blog when I expressed that I hope the pandemic will not increase separate-ness in the world. Since then, I have been watching the news of Italy, New York, Spain, and the harrowing stories of health care workers on the front lines talking about running out of PPE, ventilators, and surges on emergency rooms. I am unable to reconcile those stories with the hopeful picture painted by most politicians and most of the media so I decided to just start tracking trends myself using publicly available data. Here is some simple modeling of potential impact using today's data: 

Metrics as of March 25th, 2020, 8:42 PM PT

• Tests given in the US = 472,820
• Positive tests in the US = 64,180
• Current deaths in the US = 900
• Positive hit rate in US = 13.57%
• Death rate in US = 0.19%
• Population of US = 327,000,000
• Estimated positive cases using current hit rate = 44,386,574
• Estimated positive cases using 25% of current hit rate = 11,096,644
• Estimated deaths - US - current death rate = 84,489
• Estimated deaths - US - current global death rate (4.5%) = 2,005,144
• Estimated deaths - US - system collapse (10%) = 4,438,657

If the current death rate holds, which it won't, and we estimate with current hit rate, then we will have 85K deaths, a terrible number that is hard to comprehend and 7 times the annual number of flu deaths. Since we know that we are self-selecting tests to the most sick and assume 25% of that, you still have 21K deaths on the low end. I believe that estimate is much too conservative, however, since our death metrics will trail 1-2 weeks behind the positive test case metrics. 

If we use the global death rate with the current hit rate in the US, we get 2 million deaths. If we estimate system collapse death rate of 10% (what is happening in Italy), then you get 4.4 million deaths. If we apply the same 25% ratio due to self-selecting only the sickest to get tests, then we still have deaths from 500,000 to 1,000,000 people in the US alone. 

I sure hope that our measures of self-isolation will stop the exponential rise of cases and deaths and we are able to bring these estimates much closer to the thousands, not millions. I will try to update this regularly with the latest testing data.