In addition to deaths, more attention has shifted to measuring coronavirus-related hospitalizations. Frustratingly, these data tend not to go back as far as numbers on confirmed cases or deaths, but in most states there are hospitalization figures going back over two months. The hospital data are measured two ways, the first being a cumulative measure, similar to the way confirmed cases and deaths are measured.

The number can only go up as more hospitalizations are added to the total. From that number, the daily number of hospitalizations can be plotted; however that number is very noisy because the numbers are submitted at the state level in a variety of ways and do not seem to reflect the true numbers per day.

In other words, the hospitalization numbers seem to come in in clumps. They can be reported as weekly totals or weekly averages, as well. But a weakness of the cumulative data is that they do not tell us much about the burden on hospitals and health care workers. The total number of coronavirus hospitalizations increased dramatically, from zero to nearly 60,000 in a month nationally, and stayed high for weeks afterward. The chart below shows that the decrease in hospitalization has been fairly steady, and overall there is far less strain on the health care system than there was in mid-April.

The northeastern U.S. was hit hardest, but most states are either seeing declining or flat trends in hospitalizations, with a few notable exceptions such as North Carolina, Texas, and Arizona. But in those states the number of hospitalizations is still relatively low, a fraction of the totals that New York and New Jersey were seeing in April. Claims that Alabama, Georgia, and Florida are emerging “hotspots” are not supported by the hospitalization numbers despite media reports to the contrary.