Covid-19 Data Issues in Saint Charles County

May 22 Update:
I reached out to the County Health Department about this discrepancy and received the following reply/explanation:
The County’s health and information systems departments did switch reporting systems recently and noticed an error in the presentation of reported numbers. This caused the issue you note, and this was corrected as soon as possible.
I am not sure if the switch refers to the update to the PowerBI dashboard or if it refers to the actual source they are collecting the data from. I have not nioticed any updates the the actual layout or content of the dashboard, so i assume it is the data source. If the change in source resulted in a greater than 5% negative change in reported cases, then there seems to be valid reason for concern with the validity of the numbers that have been previously reported.

Original post:
I have been tracking the spread of Covid-19 in St. Charles County since the day the first case was announced. For several weeks, the County only had a table of current data on the site, so I was manually tracking daily case totals in a spreadsheet in order to see the trends. Finally, the County utilized PowerBI to provide some graphs of the data in a dashboard format. However, the County had not created some of the graphs that I thought helped better tell the story, so I continue to input and update daily new cases and deaths into my spreadsheet, all based on the published County data.
Over the course of this pandemic, firm numbers regarding positive cases, and even deaths, have been hard to establish. For Saint Charles County, the tale is no different. Most days, I have to scroll back though the previous days and update a few up or down by 1 or 2 cases. Typically, most of the changes occur to dates within the past 14 days, which makes sense, as this is within the window of symptoms presenting and test results being obtained. However, over the past week, I have noticed that the numbers are changing on dates much further back in time. 


The reason I felt compelled to write this post today is do to what happened yesterday when I went to reconcile the newest data from the County. I waited a full day to make sure that the numbers did not get revised significantly, in case there had been an error. The numbers have since remained consistent, which is why this post now exists.

The Cumulative Case Total on March 19, 2020 was 761 cases. When I checked the data on March 20, 2020, the Cumulative Case Total was 719. Somewhere in a 12 hour period, the data was updated to reduce the total number of cases in Saint Charles County by 52 cases (6.23%). The vast majority of these updates occurred for case totals reported in April, with the largest reductions occurring on April 13th (-12 cases) and April 23rd (-11 cases).

I am certain that there is a lot of data being collected by the County from different sources, but I am not sure how you are not able to accurately track and count cases by such a wide margin of error. Given that each case is identified by a unique identifier when it is processed, I do not understand why these numbers continue to be so elusive to solidify. (see update at top of this post)  

St. Charles County continues to report the Cumulative Case Total and Cumulative Death Totals, as they should. While the county dashboard does provide a daily breakdown of the new cases, it does not do so for the daily death totals. I would like for this to be rectified, as there is no reason not to show that data. In fact, it is already being provided in a graph, but it is not readily accessible for inclusion in any sort of analysis. In fact, I would like to see the County release the data tables that are used to create the existing  graphs, placing them in a table on the website or through some other mechanism. This would allow the residents of our County the ability to track and perform their own analysis on the data without having to manually track and update the daily totals that appear to be getting updated constantly.

I would also like to see the county dashboard updated to include a graphical representation of the Active Case graph, which gives a much clearer picture of where our county is on the famous curve. Emphasizing the Cumulative Total is not helpful and only causes fear and confusion.

In this unprecedented time, I believe that is most beneficial to give residents access to as much base data as possible, without violating HIPAA and other protections. This allows each individual who is interested to analyze and think critically for themselves about what the situation may be and how they might best care for themselves and those around them.

Click here to view the most updated graphs of Covid-19 in St. Charles County that I have created.

No comments:

Post a Comment