On weekends, some of the people in labs, health departments, hospitals, and medical examiner’s offices who do the work of translating individual illnesses and deaths into data points get to go home. On Sundays and Mondays, when weekend COVID-19 data are reported, we see drops in most of the metrics we compile from states, then higher numbers during the rest of the week. Major US holidays act like super-size weekends: For most metrics, we see big drops followed by equally big spikes—neither of which are likely to be accurate measures of what’s actually happening across the country.
Since Christmas, reported cases, tests, and deaths have all declined sharply. Cases and deaths are once again rising, but given that the New Year’s holiday weekend will also cause data disruptions, we aren’t expecting a return to normal reporting until closer to the middle of January.
Of our four top-line metrics, only hospitalization counts remain relatively stable through holiday data disruptions. There’s no responsible way to interpret the other major metrics until holiday backlogs have come and gone, so for this final update of the year, we’re focusing on hospitalizations, which show only mild and transient holiday reporting artifacts.