There will be a lot of guesswork regarding the true mortality rate until there's much more testing done — but I don't think the percentage of infections resulting in death is the most important number because it ignores the virulence of SARS-CoV-2. The more important number for me is the raw number of deaths (which still might be underreported, but is still easier to measure than total infections). Here's what we know based upon available data: there have been about 100,000 COVID-19-related deaths in the United States in about three months. Compared with a typical flu season, this is more than three times as many deaths in about half the time, making it at least six times more deadly than a typical seasonal flu. Considering that this is the case despite the many precautions taken to slow the spread of infection (e.g., face masks, physical distancing, disinfecting), the per capita mortality rate would likely have been much more severe had we maintained the status quo. I can't speculate on the CDC's methods or motivations, but even if their data were accurate it would still be incomplete regarding the rate of infection and the overall death toll.

The doctors l have seen say there have been more deaths from the virus because tests have not been done on everyone, especially in assisted living facilities.