Amidst the bad news of COVID-19 increasing in the USA for the third time to record levels, there is some evidence that the case fatality rate (CFR = deaths/cases) has improved significantly from spring to summer.  Two large studies of patients in NYC and in England both show significant improvement in survival rates.  We have done a study of cases in Florida and find a similar trend in improvement. 

The figure above shows that cases (brown squares) in Florida first increased in March and then eased in June, only to surge again and burn out by early October.  Since then cases have risen again for the third time and deaths (blue diamonds) appear to have bottomed out and are increasing again.  During the first wave, the first 82,719 cases reported in FL on 6/17/20 were responsible for most of the 3702 cumulative deaths reported on 7/4/20.  This corresponded to an overall CFR of 4.5% (with deaths lagging cases by 2-3 weeks.  The second wave ended roughly on 10/20/20 with 676k more cases which lead to 13.8k more deaths that were mostly reported by 11/14/20.  This corresponds to an overall CFR of 2.0%, roughly a factor of 2 improvement from spring to summer.  The lag time between cases and deaths (as the reported date shown above) increased to 3-4 weeks as reporting lags increased in Florida (real lag time remains about 2-3 weeks). Notice that the scale on the right for deaths is 2% of the scale on the left for cases.  This is what we are predicting for the CFR for the third wave in FL.  A similar pattern is seen in the data for the USA as a whole where the CFR has improved from 6.7% to 1.5% currently. 

The strongest driver of CFR is age and the second wave was caused by many younger adults increasing their activity and getting infected.  We need to check how much a younger median age played into this observed improvement. 

The figure above shows that the improvement from wave 1 to wave 2 is remarkably similar for all age groups.  For example for adults between the ages of 65-74, 12.7% died in the first wave while only 5.5% died in the second wave.  Much of this improvement can be attributed to wider testing (identifying milder cases and cases earlier in their cycle), better hospital practices (e.g. proning rather than immediate ventilating), and better therapeutics (remdesivir and dexamethasone).  A newly approved monoclonal antibody, bamlanivimab, could improve this further. 

One note of caution is that FL and the USA are just approaching the low CFR levels seen in South Korea and other best-in-class countries all along suggesting that much of the improvement is due to the UK and USA finally getting their act together in terms of testing and treatment.  In the spring, USA and UK undertested and missed many asymptomatic and mildly symptomatic cases, and many hospitals were overwhelmed so only the most seriously ill were admitted who then died at a high rate.  Some of the improvements in survival rates could be reversed if hospitals become overwhelmed with equipment and staff shortages in this ferocious third wave.