Estimating COVID-19 Cases in Virginia with Hospitalization Rates

The availability of testing the the early months of the COVID-19 pandemic was extremely restricted for various reasons. This brings into question the true number of cases, assuming that had testing been more available more people would have been tested and tested positive.

This analysis proposes to use the hospitalization rate in the more recent months to predict what might have been the actual number of cases in the early months of the pandemic. We use data from the Virginia Department of Health .

The idea is that once testing reached a steady state, then the rate of hospitalization during that time could be used to predict the likely number of cases in the past. Many external factors could affect the number of positive tests, including availability, but one could presume that if a person is sick enough to be hospitalized, then they would be hospitalized, even in the early stages of the pandemic. One potential limiting factor in hospitalizations could have been hospital capacity, but there is no evidence that people who needed to be hospitalized were turned away due to lack of capacity.

\(~\)

Virginia Cases and Hospitalizations

The following charts show the 5-day moving average of cases, hospitalizations, and hospitalization rate for the state of Virginia. The hospitalization rate is the ratio of hospitalizations to positive cases reported each day. The source data was downloaded from the Virginia Department of Health website on May 04, 2021.

The hospitalization rate trends sharply down from mid April and seems to stabilize around mid July.

The average hospitalization rate in Virginia for the past 4 weeks is 0.053.

The average hospitalization rate in Virginia for the past 8 weeks is 0.051.

\(~\)

Predicting Cases with the Hospitalization Rate

Using the average hospitalization rate over the past 8 weeks to predict the number of cases is based on the idea that the rate of hospitalization of those infected with the virus is constant in the aggregate. This chart compares the reported number of cases with predicted cases.

The total number of reported cases is 662,696.

The total number of predicted cases is 557,707.

The difference between reported and predicted is -104,989 cases. This represents a -15.8 percent increase in the number of cases.

\(~\)

Impact on Trend Analysis

The finding that the number of cases predicted is greater than the reported cases is not surprising. But it does provide different perspective on the current state of the pandemic, in this case in Virginia. If the predicted number of cases is in fact closer to the actual number than those reported, then it appears that there are fewer new cases each day in Virginia now than there were in the beginning of the pandemic.

First 90 Days Last 90 Days
Reported Cases 54,506 149,357
Predicted Cases 107,818 136,291

\(~\)

Impact on Death Rate

The use of the predicted cases also reduces the likely death rate, since the number of deaths does not change. As of May 04, 2021, there were 10823 deaths from COVID-19 reported in Virginia.

Death Rate based on Reported Cases: 0.0163

Death Rate based on Predicted Cases: 0.0194

\(~\)

What About Just Northern Virginia?

When this approach is applied to the data for just Northern Virginia, the following chart shows the results.

The total number of reported cases is 181,557.

The total number of predicted cases is 172,181.1.

The difference between reported and predicted is -9,375.857 cases. This represents a -5.2 percent increase in the number of cases.

\(~\)

Data Fairways Home