My next-door neighbor died in his sleep awhile ago. The day before he was pruning his fruit trees. He had died in his bed. He was over 75 and had heart problems. We called the police. No investigation. No autopsy. His death was recorded in public records as a heart attack.
But when pandemics are surging, such as in New York and Italy, people die in their homes or on the way to the hospital. Hospitals were overwhelmed. We did not know much then about Covid-19. Sometimes deaths were not recorded as due to Covid-19.
There have been issues in the past about how the cause of death is estimated. For example, suicide, which many families don’t want recorded. Or AIDS, when the epidemic was very strong. In some rural areas, for example, the coroner is not a physician and maybe a funeral director or other occupation. For Covid, people who died tended to have serious pre-existing conditions, which can make selecting one specific cause of death difficult.
In the 1918 Spanish flu, so many people were dying (an estimated 635,000 in the U.S) it was hard to even get them buried. There were few funerals. But, there are estimates of how many people died. How was this done?
“Excess deaths” was used. You compare the typical death rates over a period of time for a city, state, or country to the death rate during the time that the pandemic occurred. For example, an average (or median) of 60,000 people die each 6 months. Compare this to how many total people died in that 6 months, say 75,000. That indicated 15,000 people died of the pandemic.
What is the best death rate to use?
I prefer the percent of deaths per population. For example, U.S. deaths to date is 186,000. The population is 328.2 million. The death rate is 0.56%. Using cases (infections) is less reliable, as they change over time. The population is fairly stable.
Unfortunately, we have never had enough testing in the U.S. to get an accurate number of cases. For the U.S., total cases are 6.15 million, which changes every day. This yields a 3% death rate with 186,000 dead. If the cases are actually higher (due to limited testing) the death rate would be lower. If the cases are lower, the death rate would be higher.