On April 22, 2020, Governor Tom Wolf announced a plan to slowly reopen Pennsylvania. The state will open in three phases: red, yellow, and green. Here is a synopsis of each of the phases:
Currently, all Pennsylvania counties are in the red phase. According to the PA Governor’s website, a county can become a candidate to reopen and move to yellow if they, “on average for the past 14 days, had 50 or less new cases per 100,000 residents per day.”
UPDATE: 5/26/20 Governor Wolf clarified the guidance (right after I posted this of course!) The target for each county is less than 50 total new cases in the last 14 DAYS, per 100,000 residents. Check out the PA COVID-19 Dashboard to see where we stand according to this standard. Keep reading below to see what all the confusion was about.
Standard: 50 cases per day, per 100k residents?
I interpreted the state guidance to mean that each day, a county should have no more than, on average for the last two weeks, 50 new cases per 100k residents in the population. This essentially would equate to daily new cases of less than 50/100,000 or .05%. Using this formula for the standard, the Philadelphia area (as well as the entire state) would currently make the list to be considered for the yellow phase. In fact, in SE PA, only Berks county has even (briefly) surpassed this threshold throughout the course of this pandemic.
Here are graphs showing the daily average of new cases per 100k residents of the population for each of the counties in Southeastern PA:
However, this guidance may not be stringent enough to prevent further spread.
So first and foremost, I want to say that Montgomery County has done a terrific job in keeping the public informed with their transparency, and the various ways they have made data available to the public. However, I have a few major concerns with this chart.
Montco targets are calculated using a total of 50 cases per 100k residents in 14 day period
First, let’s go over how these targets are calculated. It appears they are taking the population divided by 100,000 and multiplying by 50 to determine the total number of acceptable cases in a 14 day period. This means the threshold they are using is a total of 50 cases per 100k residents in a 14 day period, NOT 50 new cases per 100k residents per day. They then take this two-week total, and divide it by 14 to determine the target number of new cases per day.
Calculating the targets using this formula creates goals that may not be attainable for several counties. The daily benchmark ends up being 1/14 of 50 per 100,000, or .00357%. In other words, in order to qualify for the yellow phase, daily new cases, on average over the course of a 14 day period, can’t total more than .00357% of the county’s population. If this is the case, Philadelphia’s growth rate would have to be less than .54% to get down to only 57 new cases a day. That is a doubling time of about 129 days! That is unlikely, if not impossible, without a vaccine! We have been under a stay-at-home order for a month, our growth rates are flat, and we are nowhere near these targets. I will analyze this in further detail in a subsequent post.
The use of cumulative totals makes it impossible to meet targets (literally!)
Montgomery County appears to be using the TOTAL NUMBER OF CASES to determine the current 14 day average. The Philadelphia Inquirer also used the total number of cases in their map. Here is a table showing the last 14 days totals as of 4/23/20, along with the 14 day average, for each county in Southeastern Pennsylvania. My calculations match closely with Montgomery County’s ‘Current – 14 Day’ metric.
As much as I love tables with data, I find it’s much easier to visualize through graphs… so here is the same information (minus the average) in a chart:
Notice how the total steadily rises over time? The total number of cases is, by definition, cumulative. The total will either stay the same or increase, but it will not decline. As a result, our ‘current’ status will never decrease for as long as we are using the total to determine our average.
Let me give you an example to demonstrate. Let’s say we have ZERO new cases from now until May 8th. Here’s how our hypothetical table would look:
And to visualize, here’s a graph showing the same information:
Even though we have zero growth – that is, no new cases – in this hypothetical, our 14 day averages skyrocket! The table below demonstrates our no growth scenario compared to our current totals as calculated by Montco. The columns in red show how much these metrics increase.
As you can see, it is NOT POSSIBLE to meet the targets set forth in Montgomery County’s table using this method.
So what measures should we use, if not total cases? Well for one, new cases will help us identify trends much better than cumulative totals. Look out for another post soon where I will explore this topic further!
Thanks for reading! If you found this information informative and/or useful, please follow me on social media and share!