Tuesday, April 14, 2020

Number Of Cases Of COVID-19 Surges Again In Northeast Health District And State

***Despite Big Increase, Some Evidence Growth Slowing***

The Georgia Department of Public Health with its noon Daily Status Report on Tuesday listed 36 new cases of confirmed COVID-19 in the 10-county Northeast Health District and one additional death–in Clarke County--attributable to the disease.

The increase of 36 cases was the largest increase since the 46 cases were added on Tuesday a week ago and the 36 were added on March 31, two weeks ago. The total number of cases in the region now stands at 355.

The added death in Clarke County was of a 76-year-old male without known underlying conditions.

Clarke County added 10 new confirmed COVID-19 cases in the 24-hour period ending at noon on Tuesday, and Oconee County added two.

Rolling Averages

The new data for the Northeast Health District parallel those for the state of Georgia.

The Department of Public Health added 908 confirmed cases of COVID-19 in the noon Tuesday Daily Status Report across the state. The new total is 14,223 cases.

The state added 37 new deaths attributable to the disease in the 24-hour period ending at noon on Tuesday, bringing the state total to 501.

The state figures show the same pattern as those locally, with numbers jumping on Tuesdays.

But the increases this Tuesday were less than those on the previous Tuesday.

A computing of rolling or moving averages across a seven-day period–reflecting the bias or pattern in the data–suggests that, despite the increase, the growth in cases and deaths may be leveling off or even showing a slight decline.


Chart 1 below provides data for the 10-county Northeast Health District and has been updated to reflect the Daily Status Report for noon on Tuesday.

The small map inserted into the chart from that Department of Public Health report reflects cases per 100,000 population and shows that Clarke, Oconee, Morgan and Oglethorpe counties have now “earned” the second level of classification because of the number of cases per population and Green County is classified at the third level.

The state report doesn’t show the actual data, but the Clarke ratio is 74.0 cases per 100,000 population, Oconee’s is 101.8, Morgan’s is 77.8, Oglethorpe’s is 118.0 and Green County’s is 163.7 cases per 100,000 population.

The ratio for the region is 68.1 cases per 100,000 population, and for the state it is 134.0 per 100,000 population.

Chart 1 (Click To Enlarge)

Chart 2 below provides the cumulative state-wide data for confirmed cases, deaths, tests, and hospitalizations, updated as of noon on Tuesday.

Chart 2 (Click To Enlarge)

Chart 3 shows the raw number of new cases of confirmed COVID-19 per day and the number of deaths attributable to COVID-19 per day, based on the noon Daily Status Reports for each day.

An examination of the chart shows the irregularity in the reports, which Tuesday reports showing particularly large numbers of cases.

April 7, with the largest number of reported cases, was a Tuesday, as was March 31, which also had a large number of cases.

Chart 3 (Click To Enlarge)

To compensate for the obvious bias or irregularity of the data, I computed moving or rolling averages across seven days.

In other words, the data point shown is not the actual data reported, but the average of data from that day and the previous six.

Those averages, labeled as “Forecast” in the chart below, smooth the lines.

They average also reflects the fact that while the data for Tuesday showed a high number of confirmed cases and deaths, the number was smaller than on the previous Tuesday, bringing the rolling or moving average down.

If that holds, the data suggest, the peaks in these statistic may be on the horizon.

The Institute for Health Metrics and Evaluation (IHME) at the University of Washington on Tuesday predicted that the peak number of deaths in Georgia will be on May 3, if current conditions hold.

Chart 4 (Click To Enlarge)


Rosemary Woodel said...

I don't understand it but it sounds hopeful.

Lee Becker said...

If you are referring to the rolling or moving average, think of it this way. Instead of reporting the daily figure, you report the average for the last seven days. You do that each day. I experimented with shorter periods, but it is clear that the data have a structure based on the day of the week, and that rolling average both makes more sense and smooths the line better.
Does that help?

Rosemary Woodel said...

Yes it does. Didn't you say a few days ago that numbers are higher on certain days for purely human reporting reasons?

Lee Becker said...

Yes, on Tuesdays numbers are higher.