Examine the statistical implications of mass COVID-19 testing. You will determine the anticipated PPV and NPV, you will analyze the possible sampling biases in the presented data, and you will identify possible correlations. Finally, you will examine the significance of these data implications on public policy.
Throughout the COVID-19 pandemic, there has been an urgency in many countries, to increase the testing of the general population to determine and contain the spread of the disease. The plan to control an epidemic is to identify and isolate all cases to stop the spread. This plan requires accurate identification of everyone that is infected and/or exposed to the disease. But no test is 100% accurate, so our project is to quantify the implications of testing accuracy during mass testing of COVID-19 and the real-life consequences of our findings on public policy.
1) Review the site below to understand Sensitivity, Specificity, Positive Predictive Value (PPV), and Negative Predictive Value (NPV). https://geekymedics.com/sensitivity-specificity-ppv-and-npv/
2) Review the accuracy of standard COVID-19 tests to determine their sensitivity and specificity.
The link below provides some information about the sensitivity and specificity of the rapid Covid tests. For our purposes, we will use the following:
Testing a population with symptoms:
Sensitivity: 92.0%
Specificity: 99.6%
Testing a population without symptoms:
Sensitivity: 80.0%
Specificity: 99.5%
Rapid, point‐of‐care antigen and molecular‐based tests for diagnosis of SARS‐CoV‐2 infection Opens in new window icon
3) Review the two links below to understand how to calculate false positives and false negatives given sensitivity, specificity, and prevalence.
False Positive CalculatorOpens in new window icon
False Negative Calculator Opens in new window icon
4) Review the following two instances in time:
– First consider March 15, 2000, when the U.S. decided to shut down. At the time, the CDC reported 383 cases in the U.S. (7-day moving average). For our analysis, please look at the anticipated results of testing 1,000 with symptoms and 1,000,000 without symptoms.
– Second, consider the peak of the outbreak on or about January 11, 2021, with 250,836 reported cases. Since this date is later in the outbreak, please consider testing 10,000 with symptoms and 10,000,000 without symptoms.
– Consider that the U.S. has 330,000,000 people during this time (for our purposes, we can consider the population of the U.S. constant a 330 million). This data will allow calculating the prevalence of the disease at these two points.
The following two resources may also be helpful in these calculations and also considerations of how to implement the results. Additional resources are also encouraged.
Pitfalls of Mass Testing for COVID-19 (https://www.brookings.edu/blog/usc-brookings-schaeffer-on-health-policy/2020/10/27/sars-cov-2-testing-what-tests-to-use-when-why-and-why-not/)
The Positives and Negatives of Mass Testing for Coronavirus (https://theconversation.com/the-positives-and-negatives-of-mass-testing-for-coronavirus-137792)
5) After reviewing the data and learning more about calculations:
• Determine the prevalence of the disease at these two dates. We will assume that the prevalence of those with symptoms is 20 times the general population for our purposes.
• Calculate the true positives, true negatives, false positives, and false negatives.
• Calculate the PPV and NPV for these two times (March and January) for the rapid Covid test using the sensitivity and specificity given above.
• Interpret the results of this statistical analysis and how these results are or are not helpful for public policy to quarantine and contract traces to control the disease.
• Examine how sampling bias (Section 9.4 in the text) could affect the results across large areas and smaller communities.
